diff options
Diffstat (limited to 'dahdi-linux-2.10.1-allo.patch')
-rw-r--r-- | dahdi-linux-2.10.1-allo.patch | 7300 |
1 files changed, 7300 insertions, 0 deletions
diff --git a/dahdi-linux-2.10.1-allo.patch b/dahdi-linux-2.10.1-allo.patch new file mode 100644 index 000000000000..99a1ed2d21f0 --- /dev/null +++ b/dahdi-linux-2.10.1-allo.patch @@ -0,0 +1,7300 @@ +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/Kbuild dahdi-linux-2.10.0.1-allo/drivers/dahdi/Kbuild +--- dahdi-linux-2.10.0.1/drivers/dahdi/Kbuild 2014-09-22 20:40:19.000000000 +0200 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/Kbuild 2015-02-10 15:12:17.829888792 +0100 +@@ -14,6 +14,8 @@ + obj-$(DAHDI_BUILD_ALL)$(CONFIG_DAHDI_WCTE12XP) += wcte12xp/ + obj-$(DAHDI_BUILD_ALL)$(CONFIG_DAHDI_WCTE13XP) += wcte13xp.o + ++obj-$(DAHDI_BUILD_ALL)$(CONFIG_DAHDI_ALLO4XXP) += allo4xxp/ ++ + wcte13xp-objs := wcte13xp-base.o wcxb_spi.o wcxb.o wcxb_flash.o + CFLAGS_wcte13xp-base.o += -I$(src)/oct612x -I$(src)/oct612x/include -I$(src)/oct612x/octdeviceapi -I$(src)/oct612x/octdeviceapi/oct6100api + ifeq ($(HOTPLUG_FIRMWARE),yes) +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/Kconfig dahdi-linux-2.10.0.1-allo/drivers/dahdi/Kconfig +--- dahdi-linux-2.10.0.1/drivers/dahdi/Kconfig 2014-09-22 20:40:19.000000000 +0200 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/Kconfig 2015-02-10 15:01:15.886129051 +0100 +@@ -291,4 +291,14 @@ + + If unsure, say Y. + ++config DAHDI_ALLO4XXP ++ tristate "Allo TE410P Quad-T1/E1 PCI" ++ depends on DAHDI && PCI ++ default DAHDI ++ ---help--- ++ To compile this driver as a module, choose M here: the ++ module will be called wcte11xp. ++ ++ If unsure, say Y. ++ + source "drivers/dahdi/xpp/Kconfig" +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/Kbuild dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/Kbuild +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/Kbuild 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/Kbuild 2015-02-10 15:01:15.886129051 +0100 +@@ -0,0 +1,31 @@ ++obj-$(DAHDI_BUILD_ALL)$(CONFIG_DAHDI_ALLO4XXP) += allo4xxp.o ++ ++FIRM_DIR := ../firmware ++ ++EXTRA_CFLAGS += -I$(src)/.. -I$(src)/../oct612x/ $(shell $(src)/../oct612x/octasic-helper cflags $(src)/../oct612x) -Wno-undef ++ ++# The OCT612X source files are from a vendor drop and we do not want to edit ++# them to make this warning go away. Therefore, turn off the ++# unused-but-set-variable warning for this driver. ++ ++EXTRA_CFLAGS += $(call cc-option, -Wno-unused-but-set-variable) ++ ++ifeq ($(HOTPLUG_FIRMWARE),yes) ++ EXTRA_CFLAGS+=-DHOTPLUG_FIRMWARE ++endif ++ ++allo4xxp-objs := base.o vpm450m.o ++ ++ifneq ($(HOTPLUG_FIRMWARE),yes) ++allo4xxp-objs += $(FIRM_DIR)/dahdi-fw-oct6114-064.o $(FIRM_DIR)/dahdi-fw-oct6114-128.o $(FIRM_DIR)/dahdi-fw-oct6114-256.o ++$(warning WARNING: You are compiling firmware into allo4xxp.ko which is not available under the terms of the GPL. It may be a violation of the GPL to distribute the resulting image since it combines both GPL and non-GPL work. You should consult a lawyer of your own before distributing such an image.) ++endif ++ ++$(obj)/$(FIRM_DIR)/dahdi-fw-oct6114-064.o: $(obj)/base.o ++ $(MAKE) -C $(obj)/$(FIRM_DIR) dahdi-fw-oct6114-064.o ++ ++$(obj)/$(FIRM_DIR)/dahdi-fw-oct6114-128.o: $(obj)/base.o ++ $(MAKE) -C $(obj)/$(FIRM_DIR) dahdi-fw-oct6114-128.o ++ ++$(obj)/$(FIRM_DIR)/dahdi-fw-oct6114-256.o: $(obj)/base.o ++ $(MAKE) -C $(obj)/$(FIRM_DIR) dahdi-fw-oct6114-256.o +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/allo4xxp.h dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/allo4xxp.h +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/allo4xxp.h 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/allo4xxp.h 2015-02-10 15:01:15.886129051 +0100 +@@ -0,0 +1,144 @@ ++/* ++ * Wildcard T400P FXS Interface Driver for DAHDI Telephony interface ++ * ++ * Written by Mark Spencer <markster@linux-support.net> ++ * ++ * Copyright (C) 2001-2010, Digium, Inc. ++ * ++ * All rights reserved. ++ * ++ */ ++ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++ ++#include <linux/ioctl.h> ++ ++#define FRMR_TTR_BASE 0x10 ++#define FRMR_RTR_BASE 0x0c ++#define FRMR_TSEO 0xa0 ++#define FRMR_TSBS1 0xa1 ++#define FRMR_CCR1 0x09 ++#define FRMR_CCR1_ITF 0x08 ++#define FRMR_CCR1_EITS 0x10 ++#define FRMR_CCR2 0x0a ++#define FRMR_CCR2_RCRC 0x04 ++#define FRMR_CCR2_RADD 0x10 ++#define FRMR_MODE 0x03 ++#define FRMR_MODE_NO_ADDR_CMP 0x80 ++#define FRMR_MODE_SS7 0x20 ++#define FRMR_MODE_HRAC 0x08 ++#define FRMR_IMR0 0x14 ++#define FRMR_IMR0_RME 0x80 ++#define FRMR_IMR0_RPF 0x01 ++#define FRMR_IMR1 0x15 ++#define FRMR_IMR1_ALLS 0x20 ++#define FRMR_IMR1_XDU 0x10 ++#define FRMR_IMR1_XPR 0x01 ++#define FRMR_XC0 0x22 ++#define FRMR_XC1 0x23 ++#define FRMR_RC0 0x24 ++#define FRMR_RC1 0x25 ++#define FRMR_SIC1 0x3e ++#define FRMR_SIC2 0x3f ++#define FRMR_SIC3 0x40 ++#define FRMR_CMR1 0x44 ++#define FRMR_CMR2 0x45 ++/* OctalFALC Only */ ++#define FRMR_CMR4 0x41 ++#define FRMR_CMR5 0x42 ++#define FRMR_CMR6 0x43 ++#define FRMR_GPC2 0x8a ++/* End Octal */ ++#define FRMR_GCR 0x46 ++#define FRMR_ISR0 0x68 ++#define FRMR_ISR0_RME 0x80 ++#define FRMR_ISR0_RPF 0x01 ++#define FRMR_ISR1 0x69 ++#define FRMR_ISR1_ALLS 0x20 ++#define FRMR_ISR1_XDU 0x10 ++#define FRMR_ISR1_XPR 0x01 ++#define FRMR_ISR2 0x6a ++#define FRMR_ISR3 0x6b ++#define FRMR_ISR4 0x6c ++#define FRMR_GIS 0x6e ++#define FRMR_GIS_ISR0 0x01 ++#define FRMR_GIS_ISR1 0x02 ++#define FRMR_GIS_ISR2 0x04 ++#define FRMR_GIS_ISR3 0x08 ++#define FRMR_GIS_ISR4 0x10 ++#define FRMR_CIS 0x6f ++#define FRMR_CIS_GIS1 0x01 ++#define FRMR_CIS_GIS2 0x02 ++#define FRMR_CIS_GIS3 0x04 ++#define FRMR_CIS_GIS4 0x08 ++ ++/* CIS - Octal falc bits */ ++#define FRMR_CIS_GIS5 0x10 ++#define FRMR_CIS_GIS6 0x20 ++#define FRMR_CIS_GIS7 0x40 ++#define FRMR_CIS_GIS8 0x80 ++ ++#define FRMR_CMDR 0x02 ++#define FRMR_CMDR_SRES 0x01 ++#define FRMR_CMDR_XRES 0x10 ++#define FRMR_CMDR_RMC 0x80 ++#define FRMR_CMDR_XTF 0x04 ++#define FRMR_CMDR_XHF 0x08 ++#define FRMR_CMDR_XME 0x02 ++#define FRMR_RSIS 0x65 ++#define FRMR_RSIS_VFR 0x80 ++#define FRMR_RSIS_RDO 0x40 ++#define FRMR_RSIS_CRC16 0x20 ++#define FRMR_RSIS_RAB 0x10 ++#define FRMR_RBCL 0x66 ++#define FRMR_RBCL_MAX_SIZE 0x1f ++#define FRMR_RBCH 0x67 ++#define FRMR_RXFIFO 0x00 ++#define FRMR_SIS 0x64 ++#define FRMR_SIS_XFW 0x40 ++#define FRMR_TXFIFO 0x00 ++ ++#define FRS0 0x4c ++#define FRS0_LOS (1<<7) ++#define FRS0_LFA (1<<5) ++#define FRS0_LMFA (1<<1) ++ ++#define FRS1 0x4d ++#define FRS1_XLS (1<<1) ++#define FRS1_XLO (1<<0) ++ ++#define NUM_REGS 0xa9 ++#define NUM_PCI 12 ++ ++struct t4_regs { ++ unsigned int pci[NUM_PCI]; ++ unsigned char regs[NUM_REGS]; ++}; ++ ++struct t4_reg { ++ unsigned int reg; ++ unsigned int val; ++}; ++ ++#define T4_CHECK_VPM 0 ++#define T4_LOADING_FW 1 ++#define T4_STOP_DMA 2 ++#define T4_CHECK_TIMING 3 ++#define T4_CHANGE_LATENCY 4 ++#define T4_IGNORE_LATENCY 5 ++ ++#define WCT4_GET_REGS _IOW(DAHDI_CODE, 60, struct t4_regs) ++#define WCT4_GET_REG _IOW(DAHDI_CODE, 61, struct t4_reg) ++#define WCT4_SET_REG _IOW(DAHDI_CODE, 62, struct t4_reg) ++ +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/base.c dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/base.c +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/base.c 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/base.c 2015-02-10 15:01:15.889462151 +0100 +@@ -0,0 +1,5677 @@ ++/* ++ * TE410P Quad-T1/E1 PCI Driver version 0.1, 12/16/02 ++ * ++ * Written by Mark Spencer <markster@digium.com> ++ * Based on previous works, designs, and archetectures conceived and ++ * written by Jim Dixon <jim@lambdatel.com>. ++ * Further modified, optimized, and maintained by ++ * Matthew Fredrickson <creslin@digium.com> and ++ * Russ Meyerriecks <rmeyerriecks@digium.com> ++ * ++ * Copyright (C) 2001 Jim Dixon / Zapata Telephony. ++ * Copyright (C) 2001-2012, Digium, Inc. ++ * ++ * All rights reserved. ++ * ++ */ ++ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++#include <linux/kernel.h> ++#include <linux/errno.h> ++#include <linux/module.h> ++#include <linux/pci.h> ++#include <linux/init.h> ++#include <linux/sched.h> ++#include <linux/interrupt.h> ++#include <linux/spinlock.h> ++#include <asm/io.h> ++#include <linux/version.h> ++#include <linux/delay.h> ++#include <linux/moduleparam.h> ++#include <linux/crc32.h> ++ ++#include <stdbool.h> ++#include <dahdi/kernel.h> ++ ++#include "allo4xxp.h" ++#include "vpm450m.h" ++ ++/* Work queues are a way to better distribute load on SMP systems */ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20)) ++/* ++ * Work queues can significantly improve performance and scalability ++ * on multi-processor machines, but requires bypassing some kernel ++ * API's, so it's not guaranteed to be compatible with all kernels. ++ */ ++/* #define ENABLE_WORKQUEUES */ ++#endif ++ ++/* Support first generation cards? */ ++#define SUPPORT_GEN1 ++ ++/* Define to get more attention-grabbing but slightly more I/O using ++ alarm status */ ++#undef FANCY_ALARM ++ ++/* Define to support Digium Voice Processing Module expansion card */ ++#define VPM_SUPPORT ++ ++#define DEBUG_MAIN (1 << 0) ++#define DEBUG_DTMF (1 << 1) ++#define DEBUG_REGS (1 << 2) ++#define DEBUG_TSI (1 << 3) ++#define DEBUG_ECHOCAN (1 << 4) ++#define DEBUG_RBS (1 << 5) ++#define DEBUG_FRAMER (1 << 6) ++ ++/* Maximum latency to be used with Gen 5 */ ++#define GEN5_MAX_LATENCY 127 ++ ++#ifdef ENABLE_WORKQUEUES ++#include <linux/cpu.h> ++ ++/* XXX UGLY!!!! XXX We have to access the direct structures of the workqueue which ++ are only defined within workqueue.c because they don't give us a routine to allow us ++ to nail a work to a particular thread of the CPU. Nailing to threads gives us substantially ++ higher scalability in multi-CPU environments though! */ ++ ++/* ++ * The per-CPU workqueue (if single thread, we always use cpu 0's). ++ * ++ * The sequence counters are for flush_scheduled_work(). It wants to wait ++ * until until all currently-scheduled works are completed, but it doesn't ++ * want to be livelocked by new, incoming ones. So it waits until ++ * remove_sequence is >= the insert_sequence which pertained when ++ * flush_scheduled_work() was called. ++ */ ++ ++struct cpu_workqueue_struct { ++ ++ spinlock_t lock; ++ ++ long remove_sequence; /* Least-recently added (next to run) */ ++ long insert_sequence; /* Next to add */ ++ ++ struct list_head worklist; ++ wait_queue_head_t more_work; ++ wait_queue_head_t work_done; ++ ++ struct workqueue_struct *wq; ++ task_t *thread; ++ ++ int run_depth; /* Detect run_workqueue() recursion depth */ ++} ____cacheline_aligned; ++ ++/* ++ * The externally visible workqueue abstraction is an array of ++ * per-CPU workqueues: ++ */ ++struct workqueue_struct { ++ /* TODO: Find out exactly where the API changed */ ++ struct cpu_workqueue_struct *cpu_wq; ++ const char *name; ++ struct list_head list; /* Empty if single thread */ ++}; ++ ++/* Preempt must be disabled. */ ++static void __t4_queue_work(struct cpu_workqueue_struct *cwq, ++ struct work_struct *work) ++{ ++ unsigned long flags; ++ ++ spin_lock_irqsave(&cwq->lock, flags); ++ work->wq_data = cwq; ++ list_add_tail(&work->entry, &cwq->worklist); ++ cwq->insert_sequence++; ++ wake_up(&cwq->more_work); ++ spin_unlock_irqrestore(&cwq->lock, flags); ++} ++ ++/* ++ * Queue work on a workqueue. Return non-zero if it was successfully ++ * added. ++ * ++ * We queue the work to the CPU it was submitted, but there is no ++ * guarantee that it will be processed by that CPU. ++ */ ++static inline int t4_queue_work(struct workqueue_struct *wq, struct work_struct *work, int cpu) ++{ ++ int ret = 0; ++ get_cpu(); ++ if (!test_and_set_bit(0, &work->pending)) { ++ BUG_ON(!list_empty(&work->entry)); ++ __t4_queue_work(wq->cpu_wq + cpu, work); ++ ret = 1; ++ } ++ put_cpu(); ++ return ret; ++} ++ ++#endif ++ ++/* ++ * Define CONFIG_FORCE_EXTENDED_RESET to allow the qfalc framer extra time ++ * to reset itself upon hardware initialization. This exits for rare ++ * cases for customers who are seeing the qfalc returning unexpected ++ * information at initialization ++ */ ++/* #define CONFIG_FORCE_EXTENDED_RESET */ ++/* #define CONFIG_NOEXTENDED_RESET */ ++ ++/* ++ * Uncomment the following definition in order to disable Active-State Power ++ * Management on the PCIe bridge for PCIe cards. This has been known to work ++ * around issues where the BIOS enables it on the cards even though the ++ * platform does not support it. ++ * ++ */ ++/* #define CONFIG_WCT4XXP_DISABLE_ASPM */ ++ ++#if defined(CONFIG_FORCE_EXTENDED_RESET) && defined(CONFIG_NOEXTENDED_RESET) ++#error "You cannot define both CONFIG_FORCE_EXTENDED_RESET and " \ ++ "CONFIG_NOEXTENDED_RESET." ++#endif ++ ++int debug = 0; ++static int timingcable = 0; ++static int t1e1override = -1; /* deprecated */ ++static char *default_linemode = "auto"; ++static int j1mode = 0; ++static int sigmode = FRMR_MODE_NO_ADDR_CMP; ++static int alarmdebounce = 2500; /* LOF/LFA def to 2.5s AT&T TR54016*/ ++static int losalarmdebounce = 2500;/* LOS def to 2.5s AT&T TR54016*/ ++static int aisalarmdebounce = 2500;/* AIS(blue) def to 2.5s AT&T TR54016*/ ++static int yelalarmdebounce = 500;/* RAI(yellow) def to 0.5s AT&T devguide */ ++static int max_latency = GEN5_MAX_LATENCY; /* Used to set a maximum latency (if you don't wish it to hard cap it at a certain value) in milliseconds */ ++#ifdef VPM_SUPPORT ++static int vpmsupport = 1; ++/* If set to auto, vpmdtmfsupport is enabled for VPM400M and disabled for VPM450M */ ++static int vpmdtmfsupport = -1; /* -1=auto, 0=disabled, 1=enabled*/ ++#endif /* VPM_SUPPORT */ ++ ++/* Enabling bursting can more efficiently utilize PCI bus bandwidth, but ++ can also cause PCI bus starvation, especially in combination with other ++ aggressive cards. Please note that burst mode has no effect on CPU ++ utilization / max number of calls / etc. */ ++static int noburst; ++/* For 56kbps links, set this module parameter to 0x7f */ ++static int hardhdlcmode = 0xff; ++ ++static int latency = 1; ++ ++static int ms_per_irq = 1; ++static int ignore_rotary; ++ ++#ifdef FANCY_ALARM ++static int altab[] = { ++0, 0, 0, 1, 2, 3, 4, 6, 8, 9, 11, 13, 16, 18, 20, 22, 24, 25, 27, 28, 29, 30, 31, 31, 32, 31, 31, 30, 29, 28, 27, 25, 23, 22, 20, 18, 16, 13, 11, 9, 8, 6, 4, 3, 2, 1, 0, 0, ++}; ++#endif ++ ++#define MAX_SPANS 16 ++ ++#define FLAG_STARTED (1 << 0) ++#define FLAG_NMF (1 << 1) ++#define FLAG_SENDINGYELLOW (1 << 2) ++ ++#define FLAG_2NDGEN (1 << 3) ++#define FLAG_2PORT (1 << 4) ++#define FLAG_VPM2GEN (1 << 5) ++#define FLAG_OCTOPT (1 << 6) ++#define FLAG_3RDGEN (1 << 7) ++#define FLAG_BURST (1 << 8) ++#define FLAG_EXPRESS (1 << 9) ++#define FLAG_5THGEN (1 << 10) ++#define FLAG_8PORT (1 << 11) ++#define FLAG_1PORT (1 << 12) ++ ++#define CANARY 0xc0de ++ ++/* names of available HWEC modules */ ++#ifdef VPM_SUPPORT ++#define T4_VPM_PRESENT (1 << 28) ++static const char *vpmoct032_name = "VPMOCT032"; //cem:1port change ++static const char *vpmoct064_name = "VPMOCT064"; ++static const char *vpmoct128_name = "VPMOCT128"; ++static const char *vpmoct256_name = "VPMOCT256"; ++#endif ++ ++struct devtype { ++ char *desc; ++ unsigned int flags; ++}; ++/*ALLO:*/ ++static struct devtype allo1280p2 = { "Allocard 2aCP8e (2nd Gen)", FLAG_8PORT | FLAG_5THGEN | FLAG_BURST | FLAG_2NDGEN | FLAG_3RDGEN | FLAG_EXPRESS }; ++static struct devtype allo1240p2 = { "Allocard 2aCP4e (2nd Gen)", FLAG_5THGEN | FLAG_BURST | FLAG_2NDGEN | FLAG_3RDGEN | FLAG_EXPRESS }; ++static struct devtype allo1241p2 = { "Allocard 2aCP4 (2nd Gen)", FLAG_5THGEN | FLAG_BURST | FLAG_2NDGEN | FLAG_3RDGEN }; ++static struct devtype allo1220p2 = { "Allocard 2aCP2e (2nd Gen)", FLAG_2PORT | FLAG_5THGEN | FLAG_BURST | FLAG_2NDGEN | FLAG_3RDGEN | FLAG_EXPRESS }; ++static struct devtype allo1210p2 = { "Allocard 2aCP1e (2nd Gen)", FLAG_5THGEN | FLAG_BURST | FLAG_2NDGEN | FLAG_3RDGEN | FLAG_1PORT }; ++ ++ ++struct t4; ++ ++enum linemode {T1, E1, J1}; ++ ++struct spi_state { ++ int wrreg; ++ int rdreg; ++}; ++ ++struct t4_span { ++ struct t4 *owner; ++ u32 *writechunk; /* Double-word aligned write memory */ ++ u32 *readchunk; /* Double-word aligned read memory */ ++ enum linemode linemode; ++ int sync; ++ int alarmtimer; ++ int notclear; ++ unsigned long alarm_time; ++ unsigned long losalarm_time; ++ unsigned long aisalarm_time; ++ unsigned long yelalarm_time; ++ unsigned long alarmcheck_time; ++ int spanflags; ++ int syncpos; ++ ++#ifdef SUPPORT_GEN1 ++ int e1check; /* E1 check */ ++#endif ++ struct dahdi_span span; ++ unsigned char txsigs[16]; /* Transmit sigs */ ++ int loopupcnt; ++ int loopdowncnt; ++#ifdef SUPPORT_GEN1 ++ unsigned char ec_chunk1[31][DAHDI_CHUNKSIZE]; /* first EC chunk buffer */ ++ unsigned char ec_chunk2[31][DAHDI_CHUNKSIZE]; /* second EC chunk buffer */ ++#endif ++ /* HDLC controller fields */ ++ struct dahdi_chan *sigchan; ++ unsigned char sigmode; ++ int sigactive; ++ int frames_out; ++ int frames_in; ++ ++#ifdef VPM_SUPPORT ++ unsigned long dtmfactive; ++ unsigned long dtmfmask; ++ unsigned long dtmfmutemask; ++#endif ++#ifdef ENABLE_WORKQUEUES ++ struct work_struct swork; ++#endif ++ struct dahdi_chan *chans[32]; /* Individual channels */ ++ struct dahdi_echocan_state *ec[32]; /* Echocan state for each channel */ ++}; ++ ++struct t4 { ++ /* This structure exists one per card */ ++ struct pci_dev *dev; /* Pointer to PCI device */ ++ unsigned int intcount; ++ int num; /* Which card we are */ ++ int syncsrc; /* active sync source */ ++ struct dahdi_device *ddev; ++ struct t4_span *tspans[8]; /* Individual spans */ ++ int numspans; /* Number of spans on the card */ ++ int blinktimer; ++#ifdef FANCY_ALARM ++ int alarmpos; ++#endif ++ int irq; /* IRQ used by device */ ++ int order; /* Order */ ++ const struct devtype *devtype; ++ unsigned int reset_required:1; /* If reset needed in serial_setup */ ++ unsigned int falc31:1; /* are we falc v3.1 (atomic not necessary) */ ++ unsigned int t1e1:8; /* T1 / E1 select pins */ ++ int ledreg; /* LED Register */ ++ int ledreg2; /* LED Register2 */ ++ unsigned int gpio; ++ unsigned int gpioctl; ++ int e1recover; /* E1 recovery timer */ ++ spinlock_t reglock; /* lock register access */ ++ int spansstarted; /* number of spans started */ ++ u32 *writechunk; /* Double-word aligned write memory */ ++ u32 *readchunk; /* Double-word aligned read memory */ ++#ifdef ENABLE_WORKQUEUES ++ atomic_t worklist; ++ struct workqueue_struct *workq; ++#endif ++ int last0; /* for detecting double-missed IRQ */ ++ ++ /* DMA related fields */ ++ unsigned int dmactrl; ++ dma_addr_t readdma; ++ dma_addr_t writedma; ++ void __iomem *membase; /* Base address of card */ ++ ++ /* Flags for our bottom half */ ++ unsigned long checkflag; ++ struct tasklet_struct t4_tlet; ++ /* Latency related additions */ ++ unsigned char rxident; ++ unsigned char lastindex; ++ int numbufs; ++ int needed_latency; ++ ++#ifdef VPM_SUPPORT ++ struct vpm450m *vpm; ++#endif ++ struct spi_state st; ++}; ++ ++static inline bool is_pcie(const struct t4 *wc) ++{ ++ return (wc->devtype->flags & FLAG_EXPRESS) > 0; ++} ++ ++static inline bool has_e1_span(const struct t4 *wc) ++{ ++ return (wc->t1e1 > 0); ++} ++ ++static inline bool is_octal(const struct t4 *wc) ++{ ++ return (wc->devtype->flags & FLAG_8PORT) > 0; ++} ++ ++static inline int T4_BASE_SIZE(struct t4 *wc) ++{ ++ if (is_octal(wc)) ++ return DAHDI_MAX_CHUNKSIZE * 32 * 8; ++ else ++ return DAHDI_MAX_CHUNKSIZE * 32 * 4; ++} ++ ++/** ++ * ports_on_framer - The number of ports on the framers. ++ * @wc: Board to check. ++ * ++ * The framer ports could be different the the number of ports on the card ++ * since the dual spans have four ports internally but two ports extenally. ++ * ++ */ ++static inline unsigned int ports_on_framer(const struct t4 *wc) ++{ ++ return (is_octal(wc)) ? 8 : 4; ++} ++ ++#ifdef VPM_SUPPORT ++static void t4_vpm_init(struct t4 *wc); ++ ++static void echocan_free(struct dahdi_chan *chan, struct dahdi_echocan_state *ec); ++ ++static const struct dahdi_echocan_features vpm_ec_features = { ++ .NLP_automatic = 1, ++ .CED_tx_detect = 1, ++ .CED_rx_detect = 1, ++}; ++ ++static const struct dahdi_echocan_ops vpm_ec_ops = { ++ .echocan_free = echocan_free, ++}; ++#endif ++ ++static void __set_clear(struct t4 *wc, int span); ++static int _t4_startup(struct file *file, struct dahdi_span *span); ++static int t4_startup(struct file *file, struct dahdi_span *span); ++static int t4_shutdown(struct dahdi_span *span); ++static int t4_rbsbits(struct dahdi_chan *chan, int bits); ++static int t4_maint(struct dahdi_span *span, int cmd); ++static int t4_clear_maint(struct dahdi_span *span); ++static int t4_reset_counters(struct dahdi_span *span); ++#ifdef SUPPORT_GEN1 ++static int t4_reset_dma(struct t4 *wc); ++#endif ++static void t4_hdlc_hard_xmit(struct dahdi_chan *chan); ++static int t4_ioctl(struct dahdi_chan *chan, unsigned int cmd, unsigned long data); ++static void t4_tsi_assign(struct t4 *wc, int fromspan, int fromchan, int tospan, int tochan); ++static void t4_tsi_unassign(struct t4 *wc, int tospan, int tochan); ++static void __t4_set_rclk_src(struct t4 *wc, int span); ++static void __t4_set_sclk_src(struct t4 *wc, int mode, int master, int slave); ++static void t4_check_alarms(struct t4 *wc, int span); ++static void t4_check_sigbits(struct t4 *wc, int span); ++ ++#define WC_RDADDR 0 ++#define WC_WRADDR 1 ++#define WC_COUNT 2 ++#define WC_DMACTRL 3 ++#define WC_INTR 4 ++/* #define WC_GPIO 5 */ ++#define WC_VERSION 6 ++#define WC_LEDS 7 ++#define WC_GPIOCTL 8 ++#define WC_GPIO 9 ++#define WC_LADDR 10 ++#define WC_LDATA 11 ++#define WC_LEDS2 12 ++ ++#define WC_SET_AUTH (1 << 20) ++#define WC_GET_AUTH (1 << 12) ++ ++#define WC_LFRMR_CS (1 << 10) /* Framer's ChipSelect signal */ ++#define WC_LCS (1 << 11) ++#define WC_LCS2 (1 << 12) ++#define WC_LALE (1 << 13) ++#define WC_LFRMR_CS2 (1 << 14) /* Framer's ChipSelect signal 2 */ ++#define WC_LREAD (1 << 15) ++#define WC_LWRITE (1 << 16) ++ ++#define WC_ACTIVATE (1 << 12) ++ ++#define WC_OFF (0) ++#define WC_RED (1) ++#define WC_GREEN (2) ++#define WC_YELLOW (3) ++ ++#define WC_RECOVER 0 ++#define WC_SELF 1 ++ ++#define LIM0_T 0x36 /* Line interface mode 0 register */ ++#define LIM0_LL (1 << 1) /* Local Loop */ ++#define LIM1_T 0x37 /* Line interface mode 1 register */ ++#define LIM1_RL (1 << 1) /* Remote Loop */ ++ ++#define FMR0 0x1C /* Framer Mode Register 0 */ ++#define FMR0_SIM (1 << 0) /* Alarm Simulation */ ++#define FMR1_T 0x1D /* Framer Mode Register 1 */ ++#define FMR1_ECM (1 << 2) /* Error Counter 1sec Interrupt Enable */ ++#define FMR5 0x21 /* Framer Mode Register 5 */ ++#define FMR5_XLU (1 << 4) /* Transmit loopup code */ ++#define FMR5_XLD (1 << 5) /* Transmit loopdown code */ ++#define FMR5_EIBR (1 << 6) /* Internal Bit Robbing Access */ ++#define DEC_T 0x60 /* Diable Error Counter */ ++#define IERR_T 0x1B /* Single Bit Defect Insertion Register */ ++#define IBV (1 << 0) /* Bipolar violation */ ++#define IPE (1 << 1) /* PRBS defect */ ++#define ICASE (1 << 2) /* CAS defect */ ++#define ICRCE (1 << 3) /* CRC defect */ ++#define IMFE (1 << 4) /* Multiframe defect */ ++#define IFASE (1 << 5) /* FAS defect */ ++#define ISR3_SEC (1 << 6) /* Internal one-second interrupt bit mask */ ++#define ISR3_ES (1 << 7) /* Errored Second interrupt bit mask */ ++#define ESM 0x47 /* Errored Second mask register */ ++ ++#define FMR2_T 0x1E /* Framer Mode Register 2 */ ++#define FMR2_PLB (1 << 2) /* Framer Mode Register 2 */ ++ ++#define FECL_T 0x50 /* Framing Error Counter Lower Byte */ ++#define FECH_T 0x51 /* Framing Error Counter Higher Byte */ ++#define CVCL_T 0x52 /* Code Violation Counter Lower Byte */ ++#define CVCH_T 0x53 /* Code Violation Counter Higher Byte */ ++#define CEC1L_T 0x54 /* CRC Error Counter 1 Lower Byte */ ++#define CEC1H_T 0x55 /* CRC Error Counter 1 Higher Byte */ ++#define EBCL_T 0x56 /* E-Bit Error Counter Lower Byte */ ++#define EBCH_T 0x57 /* E-Bit Error Counter Higher Byte */ ++#define BECL_T 0x58 /* Bit Error Counter Lower Byte */ ++#define BECH_T 0x59 /* Bit Error Counter Higher Byte */ ++#define COEC_T 0x5A /* COFA Event Counter */ ++#define PRBSSTA_T 0xDA /* PRBS Status Register */ ++ ++#define LCR1_T 0x3B /* Loop Code Register 1 */ ++#define EPRM (1 << 7) /* Enable PRBS rx */ ++#define XPRBS (1 << 6) /* Enable PRBS tx */ ++#define FLLB (1 << 1) /* Framed line loop/Invert */ ++#define LLBP (1 << 0) /* Line Loopback Pattern */ ++#define TPC0_T 0xA8 /* Test Pattern Control Register */ ++#define FRA (1 << 6) /* Framed/Unframed Selection */ ++#define PRBS23 (3 << 4) /* Pattern selection (23 poly) */ ++#define PRM (1 << 2) /* Non framed mode */ ++#define FRS1_T 0x4D /* Framer Receive Status Reg 1 */ ++#define LLBDD (1 << 4) ++#define LLBAD (1 << 3) ++ ++#define MAX_T4_CARDS 64 ++ ++static void t4_isr_bh(unsigned long data); ++ ++static struct t4 *cards[MAX_T4_CARDS]; ++ ++struct t8_firm_header { ++ u8 header[6]; ++ __le32 chksum; ++ u8 pad[18]; ++ __le32 version; ++} __packed; ++ ++#define MAX_TDM_CHAN 32 ++#define MAX_DTMF_DET 16 ++ ++#define HDLC_IMR0_MASK (FRMR_IMR0_RME | FRMR_IMR0_RPF) ++#define HDLC_IMR1_MASK (FRMR_IMR1_XDU | FRMR_IMR1_XPR) ++ ++static inline unsigned int __t4_pci_in(struct t4 *wc, const unsigned int addr) ++{ ++ unsigned int res = readl(wc->membase + (addr * sizeof(u32))); ++ return res; ++} ++ ++static inline void __t4_pci_out(struct t4 *wc, const unsigned int addr, const unsigned int value) ++{ ++#ifdef DEBUG ++ unsigned int tmp; ++#endif ++ writel(value, wc->membase + (addr * sizeof(u32))); ++#ifdef DEBUG ++ tmp = __t4_pci_in(wc, WC_VERSION); ++ if ((tmp & 0xffff0000) != 0xc01a0000) ++ dev_notice(&wc->dev->dev, ++ "Version Synchronization Error!\n"); ++#else ++ __t4_pci_in(wc, WC_VERSION); ++#endif ++} ++ ++static inline void __t4_gpio_set(struct t4 *wc, unsigned bits, unsigned int val) ++{ ++ unsigned int newgpio; ++ newgpio = wc->gpio & (~bits); ++ newgpio |= val; ++ if (newgpio != wc->gpio) { ++ wc->gpio = newgpio; ++ __t4_pci_out(wc, WC_GPIO, wc->gpio); ++ } ++} ++ ++static inline void __t4_gpio_setdir(struct t4 *wc, unsigned int bits, unsigned int val) ++{ ++ unsigned int newgpioctl; ++ newgpioctl = wc->gpioctl & (~bits); ++ newgpioctl |= val; ++ if (newgpioctl != wc->gpioctl) { ++ wc->gpioctl = newgpioctl; ++ __t4_pci_out(wc, WC_GPIOCTL, wc->gpioctl); ++ } ++} ++ ++static inline void t4_gpio_setdir(struct t4 *wc, unsigned int bits, unsigned int val) ++{ ++ unsigned long flags; ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_gpio_setdir(wc, bits, val); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static inline void t4_gpio_set(struct t4 *wc, unsigned int bits, unsigned int val) ++{ ++ unsigned long flags; ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_gpio_set(wc, bits, val); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static inline void t4_pci_out(struct t4 *wc, const unsigned int addr, const unsigned int value) ++{ ++ unsigned long flags; ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_pci_out(wc, addr, value); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static inline void __t4_set_led(struct t4 *wc, int span, int color) ++{ ++ if (span <= 3) { ++ int oldreg = wc->ledreg; ++ ++ wc->ledreg &= ~(0x3 << (span << 1)); ++ wc->ledreg |= (color << (span << 1)); ++ if (oldreg != wc->ledreg) ++ __t4_pci_out(wc, WC_LEDS, wc->ledreg); ++ } else { ++ int oldreg = wc->ledreg2; ++ ++ span &= 3; ++ wc->ledreg2 &= ~(0x3 << (span << 1)); ++ wc->ledreg2 |= (color << (span << 1)); ++ if (oldreg != wc->ledreg2) ++ __t4_pci_out(wc, WC_LEDS2, wc->ledreg2); ++ } ++} ++ ++static inline void t4_activate(struct t4 *wc) ++{ ++ wc->ledreg |= WC_ACTIVATE; ++ t4_pci_out(wc, WC_LEDS, wc->ledreg); ++} ++ ++static inline unsigned int t4_pci_in(struct t4 *wc, const unsigned int addr) ++{ ++ unsigned int ret; ++ unsigned long flags; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ret = __t4_pci_in(wc, addr); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return ret; ++} ++ ++static unsigned int __t4_framer_in(const struct t4 *wc, int unit, ++ const unsigned int addr) ++{ ++ unsigned int ret; ++ register u32 val; ++ void __iomem *const wc_laddr = wc->membase + (WC_LADDR*sizeof(u32)); ++ void __iomem *const wc_version = wc->membase + (WC_VERSION*sizeof(u32)); ++ void __iomem *const wc_ldata = wc->membase + (WC_LDATA*sizeof(u32)); ++ int haddr = (((unit & 4) ? 0 : WC_LFRMR_CS2)); ++ unit &= 0x3; ++ ++ val = ((unit & 0x3) << 8) | (addr & 0xff) | haddr; ++ writel(val, wc_laddr); ++ readl(wc_version); ++ writel(val | WC_LFRMR_CS | WC_LREAD, wc_laddr); ++ readl(wc_version); ++ ret = readb(wc_ldata); ++ writel(val, wc_laddr); ++ readl(wc_version); ++ return ret; ++} ++ ++static unsigned int ++t4_framer_in(struct t4 *wc, int unit, const unsigned int addr) ++{ ++ unsigned long flags; ++ unsigned int ret; ++ spin_lock_irqsave(&wc->reglock, flags); ++ ret = __t4_framer_in(wc, unit, addr); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return ret; ++} ++ ++static void __t4_framer_out(const struct t4 *wc, int unit, const u8 addr, ++ const unsigned int value) ++{ ++ register u32 val; ++ void __iomem *const wc_laddr = wc->membase + (WC_LADDR*sizeof(u32)); ++ void __iomem *const wc_version = wc->membase + (WC_VERSION*sizeof(u32)); ++ void __iomem *const wc_ldata = wc->membase + (WC_LDATA*sizeof(u32)); ++ int haddr = (((unit & 4) ? 0 : WC_LFRMR_CS2)); ++ ++ val = ((unit & 0x3) << 8) | (addr & 0xff) | haddr; ++ writel(val, wc_laddr); ++ readl(wc_version); ++ writel(value, wc_ldata); ++ readl(wc_version); ++ writel(val | WC_LFRMR_CS | WC_LWRITE, wc_laddr); ++ readl(wc_version); ++ writel(val, wc_laddr); ++ readl(wc_version); ++} ++ ++static void t4_framer_out(struct t4 *wc, int unit, ++ const unsigned int addr, ++ const unsigned int value) ++{ ++ unsigned long flags; ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_framer_out(wc, unit, addr, value); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++#ifdef VPM_SUPPORT ++ ++static inline void __t4_raw_oct_out(struct t4 *wc, const unsigned int addr, const unsigned int value) ++{ ++ int octopt = wc->tspans[0]->spanflags & FLAG_OCTOPT; ++ if (!octopt) ++ __t4_gpio_set(wc, 0xff, (addr >> 8)); ++ __t4_pci_out(wc, WC_LDATA, 0x10000 | (addr & 0xffff)); ++ if (!octopt) ++ __t4_pci_out(wc, WC_LADDR, (WC_LWRITE)); ++ __t4_pci_out(wc, WC_LADDR, (WC_LWRITE | WC_LALE)); ++ if (!octopt) ++ __t4_gpio_set(wc, 0xff, (value >> 8)); ++ __t4_pci_out(wc, WC_LDATA, (value & 0xffff)); ++ __t4_pci_out(wc, WC_LADDR, (WC_LWRITE | WC_LALE | WC_LCS)); ++ __t4_pci_out(wc, WC_LADDR, (0)); ++} ++ ++static inline unsigned int __t4_raw_oct_in(struct t4 *wc, const unsigned int addr) ++{ ++ unsigned int ret; ++ int octopt = wc->tspans[0]->spanflags & FLAG_OCTOPT; ++ if (!octopt) ++ __t4_gpio_set(wc, 0xff, (addr >> 8)); ++ __t4_pci_out(wc, WC_LDATA, 0x10000 | (addr & 0xffff)); ++ if (!octopt) ++ __t4_pci_out(wc, WC_LADDR, (WC_LWRITE)); ++ __t4_pci_out(wc, WC_LADDR, (WC_LWRITE | WC_LALE)); ++ __t4_pci_out(wc, WC_LADDR, (WC_LALE)); ++ if (!octopt) { ++ __t4_gpio_setdir(wc, 0xff, 0x00); ++ __t4_gpio_set(wc, 0xff, 0x00); ++ } ++ __t4_pci_out(wc, WC_LADDR, (WC_LREAD | WC_LALE | WC_LCS)); ++ if (octopt) { ++ ret = __t4_pci_in(wc, WC_LDATA) & 0xffff; ++ } else { ++ ret = __t4_pci_in(wc, WC_LDATA) & 0xff; ++ ret |= (__t4_pci_in(wc, WC_GPIO) & 0xff) << 8; ++ } ++ __t4_pci_out(wc, WC_LADDR, (0)); ++ if (!octopt) ++ __t4_gpio_setdir(wc, 0xff, 0xff); ++ return ret & 0xffff; ++} ++ ++static inline unsigned int __t4_oct_in(struct t4 *wc, unsigned int addr) ++{ ++#ifdef PEDANTIC_OCTASIC_CHECKING ++ int count = 1000; ++#endif ++ __t4_raw_oct_out(wc, 0x0008, (addr >> 20)); ++ __t4_raw_oct_out(wc, 0x000a, (addr >> 4) & ((1 << 16) - 1)); ++ __t4_raw_oct_out(wc, 0x0000, (((addr >> 1) & 0x7) << 9) | (1 << 8) | (1)); ++#ifdef PEDANTIC_OCTASIC_CHECKING ++ while((__t4_raw_oct_in(wc, 0x0000) & (1 << 8)) && --count); ++ if (count != 1000) ++ dev_notice(&wc->dev->dev, "Yah, read can be slow...\n"); ++ if (!count) ++ dev_notice(&wc->dev->dev, "Read timed out!\n"); ++#endif ++ return __t4_raw_oct_in(wc, 0x0004); ++} ++ ++static inline unsigned int t4_oct_in(struct t4 *wc, const unsigned int addr) ++{ ++ unsigned long flags; ++ unsigned int ret; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ret = __t4_oct_in(wc, addr); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return ret; ++} ++ ++static inline void __t4_oct_out(struct t4 *wc, unsigned int addr, unsigned int value) ++{ ++#ifdef PEDANTIC_OCTASIC_CHECKING ++ int count = 1000; ++#endif ++ __t4_raw_oct_out(wc, 0x0008, (addr >> 20)); ++ __t4_raw_oct_out(wc, 0x000a, (addr >> 4) & ((1 << 16) - 1)); ++ __t4_raw_oct_out(wc, 0x0004, value); ++ __t4_raw_oct_out(wc, 0x0000, (((addr >> 1) & 0x7) << 9) | (1 << 8) | (3 << 12) | 1); ++#ifdef PEDANTIC_OCTASIC_CHECKING ++ while((__t4_raw_oct_in(wc, 0x0000) & (1 << 8)) && --count); ++ if (count != 1000) ++ dev_notice(&wc->dev->dev, "Yah, write can be slow\n"); ++ if (!count) ++ dev_notice(&wc->dev->dev, "Write timed out!\n"); ++#endif ++} ++ ++static inline void t4_oct_out(struct t4 *wc, const unsigned int addr, const unsigned int value) ++{ ++ unsigned long flags; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_oct_out(wc, addr, value); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static void t4_check_vpm(struct t4 *wc) ++{ ++ int channel, tone, start, span; ++ ++ if (vpm450m_checkirq(wc->vpm)) { ++ while(vpm450m_getdtmf(wc->vpm, &channel, &tone, &start)) { ++ span = channel & 0x3; ++ channel >>= 2; ++ if (!has_e1_span(wc)) ++ channel -= 5; ++ else ++ channel -= 1; ++ if (unlikely(debug)) ++ dev_info(&wc->dev->dev, "Got tone %s of '%c' " ++ "on channel %d of span %d\n", ++ (start ? "START" : "STOP"), ++ tone, channel, span + 1); ++ if (test_bit(channel, &wc->tspans[span]->dtmfmask) && (tone != 'u')) { ++ if (start) { ++ /* The octasic is supposed to mute us, but... Yah, you ++ guessed it. */ ++ if (test_bit(channel, &wc->tspans[span]->dtmfmutemask)) { ++ unsigned long flags; ++ struct dahdi_chan *chan = wc->tspans[span]->span.chans[channel]; ++ int y; ++ spin_lock_irqsave(&chan->lock, flags); ++ for (y=0;y<chan->numbufs;y++) { ++ if ((chan->inreadbuf > -1) && (chan->readidx[y])) ++ memset(chan->readbuf[chan->inreadbuf], DAHDI_XLAW(0, chan), chan->readidx[y]); ++ } ++ spin_unlock_irqrestore(&chan->lock, flags); ++ } ++ set_bit(channel, &wc->tspans[span]->dtmfactive); ++ dahdi_qevent_lock(wc->tspans[span]->span.chans[channel], (DAHDI_EVENT_DTMFDOWN | tone)); ++ } else { ++ clear_bit(channel, &wc->tspans[span]->dtmfactive); ++ dahdi_qevent_lock(wc->tspans[span]->span.chans[channel], (DAHDI_EVENT_DTMFUP | tone)); ++ } ++ } ++ } ++ } ++} ++ ++#endif /* VPM_SUPPORT */ ++ ++static void hdlc_stop(struct t4 *wc, unsigned int span) ++{ ++ struct t4_span *t = wc->tspans[span]; ++ unsigned char imr0, imr1, mode; ++ unsigned long flags; ++ int i = 0; ++ ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Stopping HDLC controller on span " ++ "%d\n", span+1); ++ ++ /* Clear receive and transmit timeslots */ ++ for (i = 0; i < 4; i++) { ++ t4_framer_out(wc, span, FRMR_RTR_BASE + i, 0x00); ++ t4_framer_out(wc, span, FRMR_TTR_BASE + i, 0x00); ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ imr0 = __t4_framer_in(wc, span, FRMR_IMR0); ++ imr1 = __t4_framer_in(wc, span, FRMR_IMR1); ++ ++ /* Disable HDLC interrupts */ ++ imr0 |= HDLC_IMR0_MASK; ++ __t4_framer_out(wc, span, FRMR_IMR0, imr0); ++ ++ imr1 |= HDLC_IMR1_MASK; ++ __t4_framer_out(wc, span, FRMR_IMR1, imr1); ++ ++ mode = __t4_framer_in(wc, span, FRMR_MODE); ++ mode &= ~FRMR_MODE_HRAC; ++ __t4_framer_out(wc, span, FRMR_MODE, mode); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ t->sigactive = 0; ++} ++ ++static inline void __t4_framer_cmd(struct t4 *wc, unsigned int span, int cmd) ++{ ++ __t4_framer_out(wc, span, FRMR_CMDR, cmd); ++} ++ ++static inline void t4_framer_cmd_wait(struct t4 *wc, unsigned int span, int cmd) ++{ ++ int sis; ++ int loops = 0; ++ ++ /* XXX could be time consuming XXX */ ++ for (;;) { ++ sis = t4_framer_in(wc, span, FRMR_SIS); ++ if (!(sis & 0x04)) ++ break; ++ if (!loops++ && (debug & DEBUG_FRAMER)) { ++ dev_notice(&wc->dev->dev, "!!!SIS Waiting before cmd " ++ "%02x\n", cmd); ++ } ++ } ++ if (loops && (debug & DEBUG_FRAMER)) ++ dev_notice(&wc->dev->dev, "!!!SIS waited %d loops\n", loops); ++ ++ t4_framer_out(wc, span, FRMR_CMDR, cmd); ++} ++ ++static int hdlc_start(struct t4 *wc, unsigned int span, struct dahdi_chan *chan, unsigned char mode) ++{ ++ struct t4_span *t = wc->tspans[span]; ++ unsigned char imr0, imr1; ++ int offset = chan->chanpos; ++ unsigned long flags; ++ ++ if (debug & DEBUG_FRAMER) ++ dev_info(&wc->dev->dev, "Starting HDLC controller for channel " ++ "%d span %d\n", offset, span+1); ++ ++ if (mode != FRMR_MODE_NO_ADDR_CMP) ++ return -1; ++ ++ mode |= FRMR_MODE_HRAC; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ /* Make sure we're in the right mode */ ++ __t4_framer_out(wc, span, FRMR_MODE, mode); ++ __t4_framer_out(wc, span, FRMR_TSEO, 0x00); ++ __t4_framer_out(wc, span, FRMR_TSBS1, hardhdlcmode); ++ ++ /* Set the interframe gaps, etc */ ++ __t4_framer_out(wc, span, FRMR_CCR1, FRMR_CCR1_ITF|FRMR_CCR1_EITS); ++ ++ __t4_framer_out(wc, span, FRMR_CCR2, FRMR_CCR2_RCRC); ++ ++ /* Set up the time slot that we want to tx/rx on */ ++ __t4_framer_out(wc, span, FRMR_TTR_BASE + (offset / 8), (0x80 >> (offset % 8))); ++ __t4_framer_out(wc, span, FRMR_RTR_BASE + (offset / 8), (0x80 >> (offset % 8))); ++ ++ imr0 = __t4_framer_in(wc, span, FRMR_IMR0); ++ imr1 = __t4_framer_in(wc, span, FRMR_IMR1); ++ ++ /* Enable our interrupts again */ ++ imr0 &= ~HDLC_IMR0_MASK; ++ __t4_framer_out(wc, span, FRMR_IMR0, imr0); ++ ++ imr1 &= ~HDLC_IMR1_MASK; ++ __t4_framer_out(wc, span, FRMR_IMR1, imr1); ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ /* Reset the signaling controller */ ++ t4_framer_cmd_wait(wc, span, FRMR_CMDR_SRES); ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ t->sigchan = chan; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ t->sigactive = 0; ++ ++ return 0; ++} ++ ++static void __set_clear(struct t4 *wc, int span) ++{ ++ int i,j; ++ int oldnotclear; ++ unsigned short val=0; ++ struct t4_span *ts = wc->tspans[span]; ++ ++ oldnotclear = ts->notclear; ++ if (E1 != ts->linemode) { ++ for (i=0;i<24;i++) { ++ j = (i/8); ++ if (ts->span.chans[i]->flags & DAHDI_FLAG_CLEAR) { ++ val |= 1 << (7 - (i % 8)); ++ ts->notclear &= ~(1 << i); ++ } else ++ ts->notclear |= (1 << i); ++ if ((i % 8)==7) { ++ if (debug) ++ dev_notice(&wc->dev->dev, "Putting %d " ++ "in register %02x on span %d" ++ "\n", val, 0x2f + j, span + 1); ++ __t4_framer_out(wc, span, 0x2f + j, val); ++ val = 0; ++ } ++ } ++ } else { ++ for (i=0;i<31;i++) { ++ if (ts->span.chans[i]->flags & DAHDI_FLAG_CLEAR) ++ ts->notclear &= ~(1 << i); ++ else ++ ts->notclear |= (1 << i); ++ } ++ } ++ if (ts->notclear != oldnotclear) { ++ unsigned char reg; ++ reg = __t4_framer_in(wc, span, FRMR_IMR0); ++ if (ts->notclear) ++ reg &= ~0x08; ++ else ++ reg |= 0x08; ++ __t4_framer_out(wc, span, FRMR_IMR0, reg); ++ } ++} ++ ++static int t4_dacs(struct dahdi_chan *dst, struct dahdi_chan *src) ++{ ++ struct t4 *wc; ++ struct t4_span *ts; ++ wc = dst->pvt; ++ ts = wc->tspans[dst->span->offset]; ++ if (src && (src->pvt != dst->pvt)) { ++ if (ts->spanflags & FLAG_2NDGEN) ++ t4_tsi_unassign(wc, dst->span->offset, dst->chanpos); ++ wc = src->pvt; ++ if (ts->spanflags & FLAG_2NDGEN) ++ t4_tsi_unassign(wc, src->span->offset, src->chanpos); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Unassigning %d/%d by " ++ "default and...\n", src->span->offset, ++ src->chanpos); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Unassigning %d/%d by " ++ "default\n", dst->span->offset, dst->chanpos); ++ return -1; ++ } ++ if (src) { ++ t4_tsi_assign(wc, src->span->offset, src->chanpos, dst->span->offset, dst->chanpos); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Assigning channel %d/%d -> " ++ "%d/%d!\n", src->span->offset, src->chanpos, ++ dst->span->offset, dst->chanpos); ++ } else { ++ t4_tsi_unassign(wc, dst->span->offset, dst->chanpos); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Unassigning channel %d/%d!" ++ "\n", dst->span->offset, dst->chanpos); ++ } ++ return 0; ++} ++ ++#ifdef VPM_SUPPORT ++ ++void oct_set_reg(void *data, unsigned int reg, unsigned int val) ++{ ++ struct t4 *wc = data; ++ t4_oct_out(wc, reg, val); ++} ++ ++unsigned int oct_get_reg(void *data, unsigned int reg) ++{ ++ struct t4 *wc = data; ++ unsigned int ret; ++ ret = t4_oct_in(wc, reg); ++ return ret; ++} ++ ++static const char *__t4_echocan_name(struct t4 *wc) ++{ ++ if (wc->vpm) { ++ if (wc->numspans == 2) ++ return vpmoct064_name; ++ else if (wc->numspans == 4) ++ return vpmoct128_name; ++ else if (wc->numspans == 1) //cem:1port change ++ return vpmoct032_name; //cem:1port change ++ else if (wc->numspans == 8) ++ return vpmoct256_name; ++ } ++ return NULL; ++} ++ ++static const char *t4_echocan_name(const struct dahdi_chan *chan) ++{ ++ struct t4 *wc = chan->pvt; ++ return __t4_echocan_name(wc); ++} ++ ++static int t4_echocan_create(struct dahdi_chan *chan, ++ struct dahdi_echocanparams *ecp, ++ struct dahdi_echocanparam *p, ++ struct dahdi_echocan_state **ec) ++{ ++ struct t4 *wc = chan->pvt; ++ struct t4_span *tspan = container_of(chan->span, struct t4_span, span); ++ int channel; ++ const bool alaw = (chan->span->deflaw == 2); ++ ++ if (!vpmsupport || !wc->vpm) ++ return -ENODEV; ++ ++ if (ecp->param_count > 0) { ++ dev_warn(&wc->dev->dev, "%s echo canceller does not support " ++ "parameters; failing request\n", ++ chan->ec_factory->get_name(chan)); ++ return -EINVAL; ++ } ++ ++ *ec = tspan->ec[chan->chanpos - 1]; ++ (*ec)->ops = &vpm_ec_ops; ++ (*ec)->features = vpm_ec_features; ++ ++ channel = has_e1_span(wc) ? chan->chanpos : chan->chanpos + 4; ++ ++ if (is_octal(wc)) ++ channel = channel << 3; ++ else ++ channel = channel << 2; ++ channel |= chan->span->offset; ++ if (debug & DEBUG_ECHOCAN) { ++ dev_notice(&wc->dev->dev, ++ "echocan: Card is %d, Channel is %d, Span is %d, offset is %d length %d\n", ++ wc->num, chan->chanpos, chan->span->offset, ++ channel, ecp->tap_length); ++ } ++ vpm450m_set_alaw_companding(wc->vpm, channel, alaw); ++ vpm450m_setec(wc->vpm, channel, ecp->tap_length); ++ return 0; ++} ++ ++static void echocan_free(struct dahdi_chan *chan, struct dahdi_echocan_state *ec) ++{ ++ struct t4 *wc = chan->pvt; ++ int channel; ++ ++ if (!wc->vpm) ++ return; ++ ++ memset(ec, 0, sizeof(*ec)); ++ channel = has_e1_span(wc) ? chan->chanpos : chan->chanpos + 4; ++ ++ if (is_octal(wc)) ++ channel = channel << 3; ++ else ++ channel = channel << 2; ++ channel |= chan->span->offset; ++ if (debug & DEBUG_ECHOCAN) { ++ dev_notice(&wc->dev->dev, ++ "echocan: Card is %d, Channel is %d, Span is %d, offset is %d length 0\n", ++ wc->num, chan->chanpos, chan->span->offset, channel); ++ } ++ vpm450m_setec(wc->vpm, channel, 0); ++} ++#endif ++ ++static int t4_ioctl(struct dahdi_chan *chan, unsigned int cmd, unsigned long data) ++{ ++ struct t4_regs regs; ++ struct t4_reg reg; ++ int x; ++ struct t4 *wc = chan->pvt; ++#ifdef VPM_SUPPORT ++ int j; ++ int channel; ++ struct t4_span *ts = wc->tspans[chan->span->offset]; ++#endif ++ ++ switch(cmd) { ++ case WCT4_SET_REG: ++ if (copy_from_user(®, (struct t4_reg __user *)data, ++ sizeof(reg))) ++ return -EFAULT; ++ t4_pci_out(wc, reg.reg, reg.val); ++ break; ++ case WCT4_GET_REG: ++ if (copy_from_user(®, (struct t4_reg __user *)data, ++ sizeof(reg))) ++ return -EFAULT; ++ reg.val = t4_pci_in(wc, reg.reg); ++ if (copy_to_user((struct t4_reg __user *)data, ++ ®, sizeof(reg))) ++ return -EFAULT; ++ break; ++ case WCT4_GET_REGS: ++ for (x=0;x<NUM_PCI;x++) ++ regs.pci[x] = t4_pci_in(wc, x); ++ for (x=0;x<NUM_REGS;x++) ++ regs.regs[x] = t4_framer_in(wc, chan->span->offset, x); ++ if (copy_to_user((void __user *) data, ++ ®s, sizeof(regs))) ++ return -EFAULT; ++ break; ++#ifdef VPM_SUPPORT ++ case DAHDI_TONEDETECT: ++ if (get_user(j, (__user int *) data)) ++ return -EFAULT; ++ if (!wc->vpm) ++ return -ENOSYS; ++ if (j && (vpmdtmfsupport == 0)) ++ return -ENOSYS; ++ if (j & DAHDI_TONEDETECT_ON) ++ set_bit(chan->chanpos - 1, &ts->dtmfmask); ++ else ++ clear_bit(chan->chanpos - 1, &ts->dtmfmask); ++ if (j & DAHDI_TONEDETECT_MUTE) ++ set_bit(chan->chanpos - 1, &ts->dtmfmutemask); ++ else ++ clear_bit(chan->chanpos - 1, &ts->dtmfmutemask); ++ ++ channel = has_e1_span(wc) ? chan->chanpos : chan->chanpos + 4; ++ if (is_octal(wc)) ++ channel = channel << 3; ++ else ++ channel = channel << 2; ++ channel |= chan->span->offset; ++ vpm450m_setdtmf(wc->vpm, channel, j & DAHDI_TONEDETECT_ON, ++ j & DAHDI_TONEDETECT_MUTE); ++ return 0; ++#endif ++ default: ++ return -ENOTTY; ++ } ++ return 0; ++} ++ ++static void inline t4_hdlc_xmit_fifo(struct t4 *wc, unsigned int span, struct t4_span *ts) ++{ ++ int res, i; ++ unsigned int size = 32; ++ unsigned char buf[32]; ++ ++ res = dahdi_hdlc_getbuf(ts->sigchan, buf, &size); ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Got buffer sized %d and res %d " ++ "for %d\n", size, res, span); ++ if (size > 0) { ++ ts->sigactive = 1; ++ ++ if (debug & DEBUG_FRAMER) { ++ dev_notice(&wc->dev->dev, "TX("); ++ for (i = 0; i < size; i++) ++ dev_notice(&wc->dev->dev, "%s%02x", ++ (i ? " " : ""), buf[i]); ++ dev_notice(&wc->dev->dev, ")\n"); ++ } ++ ++ for (i = 0; i < size; i++) ++ t4_framer_out(wc, span, FRMR_TXFIFO, buf[i]); ++ ++ if (res) /* End of message */ { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, ++ "transmiting XHF|XME\n"); ++ t4_framer_cmd_wait(wc, span, FRMR_CMDR_XHF | FRMR_CMDR_XME); ++ ++ts->frames_out; ++ if ((debug & DEBUG_FRAMER) && !(ts->frames_out & 0x0f)) ++ dev_notice(&wc->dev->dev, "Transmitted %d " ++ "frames on span %d\n", ts->frames_out, ++ span); ++ } else { /* Still more to transmit */ ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "transmiting XHF\n"); ++ t4_framer_cmd_wait(wc, span, FRMR_CMDR_XHF); ++ } ++ } ++ else if (res < 0) ++ ts->sigactive = 0; ++} ++ ++static void t4_hdlc_hard_xmit(struct dahdi_chan *chan) ++{ ++ struct t4 *wc = chan->pvt; ++ int span = chan->span->offset; ++ struct t4_span *ts = wc->tspans[span]; ++ unsigned long flags; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ if (!ts->sigchan) { ++ dev_notice(&wc->dev->dev, "t4_hdlc_hard_xmit: Invalid (NULL) " ++ "signalling channel\n"); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return; ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "t4_hdlc_hard_xmit on channel %s " ++ "(sigchan %s), sigactive=%d\n", chan->name, ++ ts->sigchan->name, ts->sigactive); ++ ++ if ((ts->sigchan == chan) && !ts->sigactive) ++ t4_hdlc_xmit_fifo(wc, span, ts); ++} ++ ++/** ++ * t4_set_framer_bits - Atomically set bits in a framer register. ++ */ ++static void t4_set_framer_bits(struct t4 *wc, unsigned int spanno, ++ unsigned int const addr, u16 bits) ++{ ++ unsigned long flags; ++ unsigned int reg; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ reg = __t4_framer_in(wc, spanno, addr); ++ __t4_framer_out(wc, spanno, addr, (reg | bits)); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static int t4_maint(struct dahdi_span *span, int cmd) ++{ ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ unsigned int reg; ++ unsigned long flags; ++ ++ if (E1 == ts->linemode) { ++ switch(cmd) { ++ case DAHDI_MAINT_NONE: ++ dev_info(&wc->dev->dev, "Clearing all maint modes\n"); ++ t4_clear_maint(span); ++ break; ++ case DAHDI_MAINT_LOCALLOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on local loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, LIM0_T, LIM0_LL); ++ break; ++ case DAHDI_MAINT_NETWORKLINELOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on network line loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, LIM1_T, LIM1_RL); ++ break; ++ case DAHDI_MAINT_NETWORKPAYLOADLOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on network payload loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, FMR2_T, FMR2_PLB); ++ break; ++ case DAHDI_MAINT_LOOPUP: ++ case DAHDI_MAINT_LOOPDOWN: ++ dev_info(&wc->dev->dev, ++ "Loopup & loopdown not supported in E1 mode\n"); ++ return -ENOSYS; ++ case DAHDI_MAINT_FAS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IFASE); ++ break; ++ case DAHDI_MAINT_MULTI_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IMFE); ++ break; ++ case DAHDI_MAINT_CRC_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, ICRCE); ++ break; ++ case DAHDI_MAINT_CAS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, ICASE); ++ break; ++ case DAHDI_MAINT_PRBS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IPE); ++ break; ++ case DAHDI_MAINT_BIPOLAR_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IBV); ++ break; ++ case DAHDI_RESET_COUNTERS: ++ t4_reset_counters(span); ++ break; ++ case DAHDI_MAINT_ALARM_SIM: ++ dev_info(&wc->dev->dev, "Invoking alarm state"); ++ t4_set_framer_bits(wc, span->offset, FMR0, FMR0_SIM); ++ break; ++ default: ++ dev_info(&wc->dev->dev, ++ "Unknown E1 maint command: %d\n", cmd); ++ return -ENOSYS; ++ } ++ } else { ++ switch(cmd) { ++ case DAHDI_MAINT_NONE: ++ dev_info(&wc->dev->dev, "Clearing all maint modes\n"); ++ t4_clear_maint(span); ++ break; ++ case DAHDI_MAINT_LOCALLOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on local loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, LIM0_T, LIM0_LL); ++ break; ++ case DAHDI_MAINT_NETWORKLINELOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on network line loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, LIM1_T, LIM1_RL); ++ break; ++ case DAHDI_MAINT_NETWORKPAYLOADLOOP: ++ dev_info(&wc->dev->dev, ++ "Turning on network payload loopback\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, FMR2_T, FMR2_PLB); ++ break; ++ case DAHDI_MAINT_LOOPUP: ++ dev_info(&wc->dev->dev, "Transmitting loopup code\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, FMR5, FMR5_XLU); ++ ts->span.maintstat = DAHDI_MAINT_REMOTELOOP; ++ break; ++ case DAHDI_MAINT_LOOPDOWN: ++ dev_info(&wc->dev->dev, "Transmitting loopdown code\n"); ++ t4_clear_maint(span); ++ t4_set_framer_bits(wc, span->offset, FMR5, FMR5_XLD); ++ break; ++ case DAHDI_MAINT_FAS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IFASE); ++ break; ++ case DAHDI_MAINT_MULTI_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IMFE); ++ break; ++ case DAHDI_MAINT_CRC_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, ICRCE); ++ break; ++ case DAHDI_MAINT_CAS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, ICASE); ++ break; ++ case DAHDI_MAINT_PRBS_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IPE); ++ break; ++ case DAHDI_MAINT_BIPOLAR_DEFECT: ++ t4_framer_out(wc, span->offset, IERR_T, IBV); ++ break; ++ case DAHDI_MAINT_PRBS: ++ dev_info(&wc->dev->dev, "PRBS not supported\n"); ++ return -ENOSYS; ++ case DAHDI_RESET_COUNTERS: ++ t4_reset_counters(span); ++ break; ++ case DAHDI_MAINT_ALARM_SIM: ++ spin_lock_irqsave(&wc->reglock, flags); ++ reg = __t4_framer_in(wc, span->offset, FMR0); ++ ++ /* ++ * The alarm simulation state machine requires us to ++ * bring this bit up and down for at least 1 clock cycle ++ */ ++ __t4_framer_out(wc, span->offset, ++ FMR0, (reg | FMR0_SIM)); ++ udelay(1); ++ __t4_framer_out(wc, span->offset, ++ FMR0, (reg & ~FMR0_SIM)); ++ udelay(1); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ reg = t4_framer_in(wc, span->offset, 0x4e); ++ if (debug & DEBUG_MAIN) { ++ dev_info(&wc->dev->dev, ++ "FRS2(alarm state): %d\n", ++ ((reg & 0xe0) >> 5)); ++ } ++ break; ++ default: ++ dev_info(&wc->dev->dev, "Unknown T1 maint command:%d\n", ++ cmd); ++ break; ++ } ++ } ++ return 0; ++} ++ ++static int t4_clear_maint(struct dahdi_span *span) ++{ ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ unsigned int reg; ++ unsigned long flags; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ /* Clear local loop */ ++ reg = __t4_framer_in(wc, span->offset, LIM0_T); ++ __t4_framer_out(wc, span->offset, LIM0_T, (reg & ~LIM0_LL)); ++ ++ /* Clear Remote Loop */ ++ reg = __t4_framer_in(wc, span->offset, LIM1_T); ++ __t4_framer_out(wc, span->offset, LIM1_T, (reg & ~LIM1_RL)); ++ ++ /* Clear Remote Payload Loop */ ++ reg = __t4_framer_in(wc, span->offset, FMR2_T); ++ __t4_framer_out(wc, span->offset, FMR2_T, (reg & ~FMR2_PLB)); ++ ++ /* Clear PRBS */ ++ reg = __t4_framer_in(wc, span->offset, LCR1_T); ++ __t4_framer_out(wc, span->offset, LCR1_T, (reg & ~(XPRBS | EPRM))); ++ ++ /* Clear loopup/loopdown signals on the line */ ++ reg = __t4_framer_in(wc, span->offset, FMR5); ++ __t4_framer_out(wc, span->offset, FMR5, (reg & ~(FMR5_XLU | FMR5_XLD))); ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ span->mainttimer = 0; ++ ++ return 0; ++} ++ ++static int t4_reset_counters(struct dahdi_span *span) ++{ ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ memset(&ts->span.count, 0, sizeof(ts->span.count)); ++ return 0; ++} ++ ++static int t4_rbsbits(struct dahdi_chan *chan, int bits) ++{ ++ u_char m,c; ++ int k,n,b; ++ struct t4 *wc = chan->pvt; ++ struct t4_span *ts = wc->tspans[chan->span->offset]; ++ unsigned long flags; ++ ++ if (debug & DEBUG_RBS) ++ dev_notice(&wc->dev->dev, "Setting bits to %d on channel %s\n", ++ bits, chan->name); ++ spin_lock_irqsave(&wc->reglock, flags); ++ k = chan->span->offset; ++ if (E1 == ts->linemode) { ++ if (chan->chanpos == 16) { ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return 0; ++ } ++ n = chan->chanpos - 1; ++ if (chan->chanpos > 15) n--; ++ b = (n % 15); ++ c = ts->txsigs[b]; ++ m = (n / 15) << 2; /* nibble selector */ ++ c &= (0xf << m); /* keep the other nibble */ ++ c |= (bits & 0xf) << (4 - m); /* put our new nibble here */ ++ ts->txsigs[b] = c; ++ /* output them to the chip */ ++ __t4_framer_out(wc,k,0x71 + b,c); ++ } else if (ts->span.lineconfig & DAHDI_CONFIG_D4) { ++ n = chan->chanpos - 1; ++ b = (n/4); ++ c = ts->txsigs[b]; ++ m = ((3 - (n % 4)) << 1); /* nibble selector */ ++ c &= ~(0x3 << m); /* keep the other nibble */ ++ c |= ((bits >> 2) & 0x3) << m; /* put our new nibble here */ ++ ts->txsigs[b] = c; ++ /* output them to the chip */ ++ __t4_framer_out(wc,k,0x70 + b,c); ++ __t4_framer_out(wc,k,0x70 + b + 6,c); ++ } else if (ts->span.lineconfig & DAHDI_CONFIG_ESF) { ++ n = chan->chanpos - 1; ++ b = (n/2); ++ c = ts->txsigs[b]; ++ m = ((n % 2) << 2); /* nibble selector */ ++ c &= (0xf << m); /* keep the other nibble */ ++ c |= (bits & 0xf) << (4 - m); /* put our new nibble here */ ++ ts->txsigs[b] = c; ++ /* output them to the chip */ ++ __t4_framer_out(wc,k,0x70 + b,c); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ if (debug & DEBUG_RBS) ++ dev_notice(&wc->dev->dev, "Finished setting RBS bits\n"); ++ return 0; ++} ++ ++static int t4_shutdown(struct dahdi_span *span) ++{ ++ int tspan; ++ int wasrunning; ++ unsigned long flags; ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ ++ tspan = span->offset + 1; ++ if (tspan < 0) { ++ dev_notice(&wc->dev->dev, "T%dXXP: Span '%d' isn't us?\n", ++ wc->numspans, span->spanno); ++ return -1; ++ } ++ ++ if (debug & DEBUG_MAIN) ++ dev_notice(&wc->dev->dev, "Shutting down span %d (%s)\n", ++ span->spanno, span->name); ++ ++ /* Stop HDLC controller if runned */ ++ if (ts->sigchan) ++ hdlc_stop(wc, span->offset); ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ wasrunning = span->flags & DAHDI_FLAG_RUNNING; ++ ++ span->flags &= ~DAHDI_FLAG_RUNNING; ++ __t4_set_led(wc, span->offset, WC_OFF); ++ if (((wc->numspans == 8) && ++ (!(wc->tspans[0]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[1]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[2]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[3]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[4]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[5]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[6]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[7]->span.flags & DAHDI_FLAG_RUNNING))) ++ || ++ ((wc->numspans == 4) && ++ (!(wc->tspans[0]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[1]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[2]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[3]->span.flags & DAHDI_FLAG_RUNNING))) ++ || ++ ((wc->numspans == 1) && //cem:1port change ++ (!(wc->tspans[0]->span.flags & DAHDI_FLAG_RUNNING))) //cem:1port change ++ || ++ ((wc->numspans == 2) && ++ (!(wc->tspans[0]->span.flags & DAHDI_FLAG_RUNNING)) && ++ (!(wc->tspans[1]->span.flags & DAHDI_FLAG_RUNNING)))) { ++ /* No longer in use, disable interrupts */ ++ dev_info(&wc->dev->dev, "ALLO%dXXP: Disabling interrupts since " ++ "there are no active spans\n", wc->numspans); ++ set_bit(T4_STOP_DMA, &wc->checkflag); ++ } else ++ set_bit(T4_CHECK_TIMING, &wc->checkflag); ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ /* Wait for interrupt routine to shut itself down */ ++ msleep(10); ++ if (wasrunning) ++ wc->spansstarted--; ++ ++ if (debug & DEBUG_MAIN) ++ dev_notice(&wc->dev->dev, "Span %d (%s) shutdown\n", ++ span->spanno, span->name); ++ return 0; ++} ++ ++static void t4_chan_set_sigcap(struct dahdi_span *span, int x) ++{ ++ struct t4_span *wc = container_of(span, struct t4_span, span); ++ struct dahdi_chan *chan = wc->chans[x]; ++ chan->sigcap = DAHDI_SIG_CLEAR; ++ /* E&M variant supported depends on span type */ ++ if (E1 == wc->linemode) { ++ /* E1 sigcap setup */ ++ if (span->lineconfig & DAHDI_CONFIG_CCS) { ++ /* CCS setup */ ++ chan->sigcap |= DAHDI_SIG_MTP2 | DAHDI_SIG_SF | ++ DAHDI_SIG_HARDHDLC; ++ return; ++ } ++ /* clear out sig and sigcap for channel 16 on E1 CAS ++ * lines, otherwise, set it correctly */ ++ if (x == 15) { ++ /* CAS signaling channel setup */ ++ wc->chans[15]->sigcap = 0; ++ wc->chans[15]->sig = 0; ++ return; ++ } ++ /* normal CAS setup */ ++ chan->sigcap |= DAHDI_SIG_EM_E1 | DAHDI_SIG_FXSLS | ++ DAHDI_SIG_FXSGS | DAHDI_SIG_FXSKS | DAHDI_SIG_SF | ++ DAHDI_SIG_FXOLS | DAHDI_SIG_FXOGS | DAHDI_SIG_FXOKS | ++ DAHDI_SIG_CAS | DAHDI_SIG_DACS_RBS; ++ } else { ++ /* T1 sigcap setup */ ++ chan->sigcap |= DAHDI_SIG_EM | DAHDI_SIG_FXSLS | ++ DAHDI_SIG_FXSGS | DAHDI_SIG_FXSKS | DAHDI_SIG_MTP2 | ++ DAHDI_SIG_SF | DAHDI_SIG_FXOLS | DAHDI_SIG_FXOGS | ++ DAHDI_SIG_FXOKS | DAHDI_SIG_CAS | DAHDI_SIG_DACS_RBS | ++ DAHDI_SIG_HARDHDLC; ++ } ++} ++ ++static int ++_t4_spanconfig(struct file *file, struct dahdi_span *span, ++ struct dahdi_lineconfig *lc) ++{ ++ int i; ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ ++ if (debug) ++ dev_info(&wc->dev->dev, "About to enter spanconfig!\n"); ++ if (debug & DEBUG_MAIN) ++ dev_notice(&wc->dev->dev, "ALLO%dXXP: Configuring span %d\n", ++ wc->numspans, span->spanno); ++ ++ if (lc->sync < 0) ++ lc->sync = 0; ++ if (lc->sync > wc->numspans) { ++ dev_warn(&wc->dev->dev, "WARNING: Cannot set priority on span %d to %d. Please set to a number between 1 and %d\n", ++ span->spanno, lc->sync, wc->numspans); ++ lc->sync = 0; ++ } ++ ++ /* remove this span number from the current sync sources, if there */ ++ for(i = 0; i < wc->numspans; i++) { ++ if (wc->tspans[i]->sync == span->spanno) ++ wc->tspans[i]->sync = 0; ++ } ++ wc->tspans[span->offset]->syncpos = lc->sync; ++ /* if a sync src, put it in proper place */ ++ if (lc->sync) ++ wc->tspans[lc->sync - 1]->sync = span->spanno; ++ ++ set_bit(T4_CHECK_TIMING, &wc->checkflag); ++ ++ /* Make sure this is clear in case of multiple startup and shutdown ++ * iterations */ ++ clear_bit(T4_STOP_DMA, &wc->checkflag); ++ ++ /* make sure that sigcaps gets updated if necessary */ ++ for (i = 0; i < span->channels; i++) ++ t4_chan_set_sigcap(span, i); ++ ++ /* If we're already running, then go ahead and apply the changes */ ++ if (span->flags & DAHDI_FLAG_RUNNING) ++ return _t4_startup(file, span); ++ ++ if (debug) ++ dev_info(&wc->dev->dev, "Done with spanconfig!\n"); ++ return 0; ++} ++ ++static int ++t4_spanconfig(struct file *file, struct dahdi_span *span, ++ struct dahdi_lineconfig *lc) ++{ ++ int ret; ++ struct dahdi_device *const ddev = span->parent; ++ struct dahdi_span *s; ++ ++ ret = _t4_spanconfig(file, span, lc); ++ ++ /* Make sure all the spans have a basic configuration in case they are ++ * not all specified in the configuration files. */ ++ lc->sync = 0; ++ list_for_each_entry(s, &ddev->spans, device_node) { ++ WARN_ON(!s->channels); ++ if (!s->channels) ++ continue; ++ if (!s->chans[0]->sigcap) ++ _t4_spanconfig(file, s, lc); ++ } ++ return ret; ++} ++ ++static int ++t4_chanconfig(struct file *file, struct dahdi_chan *chan, int sigtype) ++{ ++ int alreadyrunning; ++ unsigned long flags; ++ struct t4 *wc = chan->pvt; ++ struct t4_span *ts = wc->tspans[chan->span->offset]; ++ ++ alreadyrunning = ts->span.flags & DAHDI_FLAG_RUNNING; ++ if (debug & DEBUG_MAIN) { ++ if (alreadyrunning) ++ dev_notice(&wc->dev->dev, "ALLO%dXXP: Reconfigured " ++ "channel %d (%s) sigtype %d\n", wc->numspans, ++ chan->channo, chan->name, sigtype); ++ else ++ dev_notice(&wc->dev->dev, "ALLO%dXXP: Configured channel" ++ " %d (%s) sigtype %d\n", wc->numspans, ++ chan->channo, chan->name, sigtype); ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ if (alreadyrunning) ++ __set_clear(wc, chan->span->offset); ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ /* (re)configure signalling channel */ ++ if ((sigtype == DAHDI_SIG_HARDHDLC) || (ts->sigchan == chan)) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "%sonfiguring hardware HDLC " ++ "on %s\n", ++ ((sigtype == DAHDI_SIG_HARDHDLC) ? "C" : "Unc"), ++ chan->name); ++ if (alreadyrunning) { ++ if (ts->sigchan) ++ hdlc_stop(wc, ts->sigchan->span->offset); ++ if (sigtype == DAHDI_SIG_HARDHDLC) { ++ if (hdlc_start(wc, chan->span->offset, chan, ts->sigmode)) { ++ dev_notice(&wc->dev->dev, "Error " ++ "initializing signalling " ++ "controller\n"); ++ return -1; ++ } ++ } else { ++ spin_lock_irqsave(&wc->reglock, flags); ++ ts->sigchan = NULL; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ } ++ ++ } ++ else { ++ spin_lock_irqsave(&wc->reglock, flags); ++ ts->sigchan = (sigtype == DAHDI_SIG_HARDHDLC) ? chan : NULL; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ts->sigactive = 0; ++ } ++ } ++ return 0; ++} ++ ++static int t4_open(struct dahdi_chan *chan) ++{ ++ return 0; ++} ++ ++static int t4_close(struct dahdi_chan *chan) ++{ ++ return 0; ++} ++ ++static int set_span_devicetype(struct t4 *wc) ++{ ++#ifdef VPM_SUPPORT ++ const char *vpmstring = __t4_echocan_name(wc); ++ ++ if (vpmstring) { ++ wc->ddev->devicetype = kasprintf(GFP_KERNEL, "%s (%s)", ++ wc->devtype->desc, vpmstring); ++ } else { ++ wc->ddev->devicetype = kasprintf(GFP_KERNEL, wc->devtype->desc); ++ } ++#else ++ wc->ddev->devicetype = kasprintf(GFP_KERNEL, wc->devtype->desc); ++#endif ++ ++ if (!wc->ddev->devicetype) ++ return -ENOMEM; ++ return 0; ++} ++ ++/* The number of cards we have seen with each ++ possible 'order' switch setting. ++*/ ++static unsigned int order_index[16]; ++ ++static void setup_chunks(struct t4 *wc, int which) ++{ ++ struct t4_span *ts; ++ int offset = 1; ++ int x, y; ++ int gen2; ++ int basesize = T4_BASE_SIZE(wc) >> 2; ++ ++ if (!has_e1_span(wc)) ++ offset += 4; ++ ++ gen2 = (wc->tspans[0]->spanflags & FLAG_2NDGEN); ++ ++ for (x = 0; x < wc->numspans; x++) { ++ ts = wc->tspans[x]; ++ ts->writechunk = (void *)(wc->writechunk + (x * 32 * 2) + (which * (basesize))); ++ ts->readchunk = (void *)(wc->readchunk + (x * 32 * 2) + (which * (basesize))); ++ for (y=0;y<wc->tspans[x]->span.channels;y++) { ++ struct dahdi_chan *mychans = ts->chans[y]; ++ if (gen2) { ++ mychans->writechunk = (void *)(wc->writechunk + ((x * 32 + y + offset) * 2) + (which * (basesize))); ++ mychans->readchunk = (void *)(wc->readchunk + ((x * 32 + y + offset) * 2) + (which * (basesize))); ++ } ++ } ++ } ++} ++ ++static int __t4_hardware_init_1(struct t4 *wc, unsigned int cardflags, ++ bool first_time); ++static int __t4_hardware_init_2(struct t4 *wc, bool first_time); ++ ++static int t4_hardware_stop(struct t4 *wc); ++ ++static void t4_framer_reset(struct t4 *wc) ++{ ++ const bool first_time = false; ++ bool have_vpm = wc->vpm != NULL; ++ if (have_vpm) { ++ release_vpm450m(wc->vpm); ++ wc->vpm = NULL; ++ } ++ t4_hardware_stop(wc); ++ __t4_set_sclk_src(wc, WC_SELF, 0, 0); ++ __t4_hardware_init_1(wc, wc->devtype->flags, first_time); ++ __t4_hardware_init_2(wc, first_time); ++ if (have_vpm) { ++ t4_vpm_init(wc); ++ wc->dmactrl |= (wc->vpm) ? T4_VPM_PRESENT : 0; ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } ++ setup_chunks(wc, 0); ++ wc->lastindex = 0; ++} ++ ++/** ++ * t4_serial_setup - Setup serial parameters and system interface. ++ * @wc: The card to configure. ++ * ++ */ ++static void t4_serial_setup(struct t4 *wc) ++{ ++ unsigned long flags; ++ unsigned int unit; ++ bool reset_required = false; ++ ++ if (debug) { ++ dev_info(&wc->dev->dev, ++ "ALLO%dXXP: Setting up global serial parameters\n", ++ wc->numspans);//cEM: ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ reset_required = wc->reset_required > 0; ++ wc->reset_required = 0; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ if (reset_required) ++ t4_framer_reset(wc); ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ /* GPC1: Multiplex mode enabled, FSC is output, active low, RCLK from ++ * channel 0 */ ++ __t4_framer_out(wc, 0, 0x85, 0xe0); ++ if (is_octal(wc)) ++ __t4_framer_out(wc, 0, FRMR_GPC2, 0x00); ++ ++ /* IPC: Interrupt push/pull active low */ ++ __t4_framer_out(wc, 0, 0x08, 0x01); ++ ++ if (is_octal(wc)) { ++ /* Global clocks (16.384 Mhz CLK) */ ++ __t4_framer_out(wc, 0, 0x92, 0x00); /* GCM1 */ ++ __t4_framer_out(wc, 0, 0x93, 0x18); ++ __t4_framer_out(wc, 0, 0x94, 0xfb); ++ __t4_framer_out(wc, 0, 0x95, 0x0b); ++ __t4_framer_out(wc, 0, 0x96, 0x01); ++ __t4_framer_out(wc, 0, 0x97, 0x0b); ++ __t4_framer_out(wc, 0, 0x98, 0xdb); ++ __t4_framer_out(wc, 0, 0x99, 0xdf); ++ } else { ++ /* Global clocks (8.192 Mhz CLK) */ ++ __t4_framer_out(wc, 0, 0x92, 0x00); ++ __t4_framer_out(wc, 0, 0x93, 0x18); ++ __t4_framer_out(wc, 0, 0x94, 0xfb); ++ __t4_framer_out(wc, 0, 0x95, 0x0b); ++ __t4_framer_out(wc, 0, 0x96, 0x00); ++ __t4_framer_out(wc, 0, 0x97, 0x0b); ++ __t4_framer_out(wc, 0, 0x98, 0xdb); ++ __t4_framer_out(wc, 0, 0x99, 0xdf); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ for (unit = 0; unit < ports_on_framer(wc); ++unit) { ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ /* Configure interrupts */ ++ /* GCR: Interrupt on Activation/Deactivation of each */ ++ __t4_framer_out(wc, unit, FRMR_GCR, 0x00); ++ ++ /* Configure system interface */ ++ if (is_octal(wc)) { ++ /* SIC1: 16.384 Mhz clock/bus, double buffer receive / ++ * transmit, byte interleaved */ ++ __t4_framer_out(wc, unit, FRMR_SIC1, 0xc2 | 0x08); ++ } else { ++ /* SIC1: 8.192 Mhz clock/bus, double buffer receive / ++ * transmit, byte interleaved */ ++ __t4_framer_out(wc, unit, FRMR_SIC1, 0xc2); ++ } ++ /* SIC2: No FFS, no center receive eliastic buffer, phase */ ++ __t4_framer_out(wc, unit, FRMR_SIC2, 0x20 | (unit << 1)); ++ /* SIC3: Edges for capture */ ++ if (is_octal(wc)) { ++ __t4_framer_out(wc, unit, FRMR_SIC3, 0x04 | (1 << 4)); ++ } else { ++ __t4_framer_out(wc, unit, FRMR_SIC3, 0x04); ++ } ++ /* CMR2: We provide sync and clock for tx and rx. */ ++ __t4_framer_out(wc, unit, FRMR_CMR2, 0x00); ++ ++ if (is_octal(wc)) { ++ /* Set RCLK to 16 MHz */ ++ __t4_framer_out(wc, unit, FRMR_CMR4, 0x5); ++ ++ if (!has_e1_span(wc)) { /* T1/J1 mode */ ++ __t4_framer_out(wc, unit, FRMR_XC0, 0x07); ++ __t4_framer_out(wc, unit, FRMR_XC1, 0x04); ++ if (wc->tspans[unit]->linemode == J1) ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x87); ++ else ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x07); ++ __t4_framer_out(wc, unit, FRMR_RC1, 0x04); ++ } else { /* E1 mode */ ++ __t4_framer_out(wc, unit, FRMR_XC0, 0x00); ++ __t4_framer_out(wc, unit, FRMR_XC1, 0x04); ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x00); ++ __t4_framer_out(wc, unit, FRMR_RC1, 0x04); ++ } ++ ++ } else { ++ if (!has_e1_span(wc)) { /* T1/J1 mode */ ++ __t4_framer_out(wc, unit, FRMR_XC0, 0x03); ++ __t4_framer_out(wc, unit, FRMR_XC1, 0x84); ++ if (J1 == wc->tspans[unit]->linemode) ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x83); ++ else ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x03); ++ __t4_framer_out(wc, unit, FRMR_RC1, 0x84); ++ } else { /* E1 mode */ ++ __t4_framer_out(wc, unit, FRMR_XC0, 0x00); ++ __t4_framer_out(wc, unit, FRMR_XC1, 0x04); ++ __t4_framer_out(wc, unit, FRMR_RC0, 0x04); ++ __t4_framer_out(wc, unit, FRMR_RC1, 0x04); ++ } ++ } ++ ++ /* Configure ports */ ++ ++ /* PC1: SPYR/SPYX input on RPA/XPA */ ++ __t4_framer_out(wc, unit, 0x80, 0x00); ++ ++ /* PC2: RMFB/XSIG output/input on RPB/XPB */ ++ /* PC3: Some unused stuff */ ++ /* PC4: Some more unused stuff */ ++ if (is_octal(wc)) { ++ __t4_framer_out(wc, unit, 0x81, 0xBB); ++ __t4_framer_out(wc, unit, 0x82, 0xf5); ++ __t4_framer_out(wc, unit, 0x83, 0x35); ++ } else if (wc->falc31) { ++ __t4_framer_out(wc, unit, 0x81, 0xBB); ++ __t4_framer_out(wc, unit, 0x82, 0xBB); ++ __t4_framer_out(wc, unit, 0x83, 0xBB); ++ } else { ++ __t4_framer_out(wc, unit, 0x81, 0x22); ++ __t4_framer_out(wc, unit, 0x82, 0x65); ++ __t4_framer_out(wc, unit, 0x83, 0x35); ++ } ++ ++ /* PC5: XMFS active low, SCLKR is input, RCLK is output */ ++ __t4_framer_out(wc, unit, 0x84, 0x01); ++ ++ if (debug & DEBUG_MAIN) { ++ dev_notice(&wc->dev->dev, ++ "Successfully initialized serial bus " ++ "for unit %d\n", unit); ++ } ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ } ++} ++ ++/** ++ * t4_span_assigned - Called when the span is assigned by DAHDI. ++ * @span: Span that has been assigned. ++ * ++ * When this function is called, the span has a valid spanno and all the ++ * channels on the span have valid channel numbers assigned. ++ * ++ * This function is necessary because a device may be registered, and ++ * then user space may then later decide to assign span numbers and the ++ * channel numbers. ++ * ++ */ ++static void t4_span_assigned(struct dahdi_span *span) ++{ ++ struct t4_span *tspan = container_of(span, struct t4_span, span); ++ struct t4 *wc = tspan->owner; ++ struct dahdi_span *pos; ++ unsigned int unassigned_spans = 0; ++ unsigned long flags; ++ ++ /* We use this to make sure all the spans are assigned before ++ * running the serial setup. */ ++ list_for_each_entry(pos, &wc->ddev->spans, device_node) { ++ if (!test_bit(DAHDI_FLAGBIT_REGISTERED, &pos->flags)) ++ ++unassigned_spans; ++ } ++ ++ if (0 == unassigned_spans) { ++ t4_serial_setup(wc); ++ ++ set_bit(T4_CHECK_TIMING, &wc->checkflag); ++ spin_lock_irqsave(&wc->reglock, flags); ++ __t4_set_sclk_src(wc, WC_SELF, 0, 0); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ } ++} ++ ++static void free_wc(struct t4 *wc) ++{ ++ unsigned int x, y; ++ ++ for (x = 0; x < ARRAY_SIZE(wc->tspans); x++) { ++ if (!wc->tspans[x]) ++ continue; ++ for (y = 0; y < ARRAY_SIZE(wc->tspans[x]->chans); y++) { ++ kfree(wc->tspans[x]->chans[y]); ++ kfree(wc->tspans[x]->ec[y]); ++ } ++ kfree(wc->tspans[x]); ++ } ++ ++ kfree(wc->ddev->devicetype); ++ kfree(wc->ddev->location); ++ kfree(wc->ddev->hardware_id); ++ dahdi_free_device(wc->ddev); ++ kfree(wc); ++} ++ ++/** ++ * t4_alloc_channels - Allocate the channels on a span. ++ * @wc: The board we're allocating for. ++ * @ts: The span we're allocating for. ++ * @linemode: Which mode (T1/E1/J1) to use for this span. ++ * ++ * This function must only be called before the span is assigned it's ++ * possible for user processes to have an open reference to the ++ * channels. ++ * ++ */ ++static int t4_alloc_channels(struct t4 *wc, struct t4_span *ts, ++ enum linemode linemode) ++{ ++ int i; ++ ++ if (test_bit(DAHDI_FLAGBIT_REGISTERED, &ts->span.flags)) { ++ dev_dbg(&wc->dev->dev, ++ "Cannot allocate channels on a span that is already " ++ "assigned.\n"); ++ return -EINVAL; ++ } ++ ++ /* Cleanup any previously allocated channels. */ ++ for (i = 0; i < ARRAY_SIZE(ts->chans); ++i) { ++ kfree(ts->chans[i]); ++ kfree(ts->ec[i]); ++ ts->chans[i] = NULL; ++ ts->ec[i] = NULL; ++ } ++ ++ ts->linemode = linemode; ++ for (i = 0; i < ((E1 == ts->linemode) ? 31 : 24); i++) { ++ struct dahdi_chan *chan; ++ struct dahdi_echocan_state *ec; ++ ++ chan = kzalloc(sizeof(*chan), GFP_KERNEL); ++ if (!chan) { ++ free_wc(wc); ++ return -ENOMEM; ++ } ++ ts->chans[i] = chan; ++ ++ ec = kzalloc(sizeof(*ec), GFP_KERNEL); ++ if (!ec) { ++ free_wc(wc); ++ return -ENOMEM; ++ } ++ ts->ec[i] = ec; ++ } ++ ++ return 0; ++} ++ ++static void t4_init_one_span(struct t4 *wc, struct t4_span *ts) ++{ ++ unsigned long flags; ++ unsigned int reg; ++ int i; ++ if(wc->numspans != 1) { ++ snprintf(ts->span.name, sizeof(ts->span.name) - 1, ++ "ALLO%d/%d/%d", wc->numspans, wc->num, ts->span.offset + 1); ++ snprintf(ts->span.desc, sizeof(ts->span.desc) - 1, ++ "T%dXXP (PCI) Card %d Span %d", wc->numspans, wc->num, ++ ts->span.offset + 1); ++ } ++ else { ++ snprintf(ts->span.name, sizeof(ts->span.name) - 1, ++ "ALLO4/%d/%d", wc->num, ts->span.offset + 1); //cem:1port change ++ snprintf(ts->span.desc, sizeof(ts->span.desc) - 1,"T4XXP (PCI) Card %d Span %d", ++ wc->num, ts->span.offset + 1); ++ } ++ switch (ts->linemode) { ++ case T1: ++ ts->span.spantype = SPANTYPE_DIGITAL_T1; ++ break; ++ case E1: ++ ts->span.spantype = SPANTYPE_DIGITAL_E1; ++ break; ++ case J1: ++ ts->span.spantype = SPANTYPE_DIGITAL_J1; ++ break; ++ } ++ ++ /* HDLC Specific init */ ++ ts->sigchan = NULL; ++ ts->sigmode = sigmode; ++ ts->sigactive = 0; ++ ++ if (E1 != ts->linemode) { ++ ts->span.channels = 24; ++ ts->span.deflaw = DAHDI_LAW_MULAW; ++ ts->span.linecompat = DAHDI_CONFIG_AMI | ++ DAHDI_CONFIG_B8ZS | DAHDI_CONFIG_D4 | ++ DAHDI_CONFIG_ESF; ++ } else { ++ ts->span.channels = 31; ++ ts->span.deflaw = DAHDI_LAW_ALAW; ++ ts->span.linecompat = DAHDI_CONFIG_AMI | ++ DAHDI_CONFIG_HDB3 | DAHDI_CONFIG_CCS | ++ DAHDI_CONFIG_CRC4; ++ } ++ ts->span.chans = ts->chans; ++ ts->span.flags = DAHDI_FLAG_RBS; ++ ++ for (i = 0; i < ts->span.channels; i++) { ++ struct dahdi_chan *const chan = ts->chans[i]; ++ chan->pvt = wc; ++ snprintf(chan->name, sizeof(chan->name) - 1, ++ "%s/%d", ts->span.name, i + 1); ++ t4_chan_set_sigcap(&ts->span, i); ++ chan->chanpos = i + 1; ++ } ++ ++ /* Enable 1sec timer interrupt */ ++ spin_lock_irqsave(&wc->reglock, flags); ++ reg = __t4_framer_in(wc, ts->span.offset, FMR1_T); ++ __t4_framer_out(wc, ts->span.offset, FMR1_T, (reg | FMR1_ECM)); ++ ++ /* Enable Errored Second interrupt */ ++ __t4_framer_out(wc, ts->span.offset, ESM, 0); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ t4_reset_counters(&ts->span); ++} ++ ++/** ++ * t4_set_linemode - Allows user space to change the linemode before spans are assigned. ++ * @span: span on which to change the linemode. ++ * @linemode: A value from enumerated spantypes ++ * ++ * This callback is used to override the E1/T1 mode jumper settings and set ++ * the linemode on for each span. Called when the "spantype" attribute ++ * is written in sysfs under the dahdi_device. ++ * ++ */ ++static int t4_set_linemode(struct dahdi_span *span, enum spantypes linemode) ++{ ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ int res = 0; ++ enum linemode mode; ++ const char *old_name; ++ static DEFINE_MUTEX(linemode_lock); ++ unsigned long flags; ++ ++ dev_dbg(&wc->dev->dev, "Setting '%s' to '%s'\n", span->name, ++ dahdi_spantype2str(linemode)); ++ ++ if (span->spantype == linemode) ++ return 0; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ wc->reset_required = 1; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ /* Do not allow the t1e1 member to be changed by multiple threads. */ ++ mutex_lock(&linemode_lock); ++ old_name = dahdi_spantype2str(span->spantype); ++ switch (linemode) { ++ case SPANTYPE_DIGITAL_T1: ++ dev_info(&wc->dev->dev, ++ "Changing from %s to T1 line mode.\n", old_name); ++ mode = T1; ++ wc->t1e1 &= ~(1 << span->offset); ++ break; ++ case SPANTYPE_DIGITAL_E1: ++ dev_info(&wc->dev->dev, ++ "Changing from %s to E1 line mode.\n", old_name); ++ mode = E1; ++ wc->t1e1 |= (1 << span->offset); ++ break; ++ case SPANTYPE_DIGITAL_J1: ++ dev_info(&wc->dev->dev, ++ "Changing from %s to J1 line mode.\n", old_name); ++ mode = J1; ++ wc->t1e1 &= ~(1 << span->offset); ++ break; ++ default: ++ dev_err(&wc->dev->dev, ++ "Got invalid linemode %d from dahdi\n", linemode); ++ res = -EINVAL; ++ } ++ ++ if (!res) { ++ t4_alloc_channels(wc, ts, mode); ++ t4_init_one_span(wc, ts); ++ dahdi_init_span(span); ++ } ++ ++ mutex_unlock(&linemode_lock); ++ return res; ++} ++ ++static const struct dahdi_span_ops t4_gen1_span_ops = { ++ .owner = THIS_MODULE, ++ .spanconfig = t4_spanconfig, ++ .chanconfig = t4_chanconfig, ++ .startup = t4_startup, ++ .shutdown = t4_shutdown, ++ .rbsbits = t4_rbsbits, ++ .maint = t4_maint, ++ .open = t4_open, ++ .close = t4_close, ++ .ioctl = t4_ioctl, ++ .hdlc_hard_xmit = t4_hdlc_hard_xmit, ++ .assigned = t4_span_assigned, ++ .set_spantype = t4_set_linemode, ++}; ++ ++static const struct dahdi_span_ops t4_gen2_span_ops = { ++ .owner = THIS_MODULE, ++ .spanconfig = t4_spanconfig, ++ .chanconfig = t4_chanconfig, ++ .startup = t4_startup, ++ .shutdown = t4_shutdown, ++ .rbsbits = t4_rbsbits, ++ .maint = t4_maint, ++ .open = t4_open, ++ .close = t4_close, ++ .ioctl = t4_ioctl, ++ .hdlc_hard_xmit = t4_hdlc_hard_xmit, ++ .dacs = t4_dacs, ++ .assigned = t4_span_assigned, ++ .set_spantype = t4_set_linemode, ++#ifdef VPM_SUPPORT ++ .echocan_create = t4_echocan_create, ++ .echocan_name = t4_echocan_name, ++#endif ++}; ++ ++/** ++ * init_spans - Do first initialization on all the spans ++ * @wc: Card to initialize the spans on. ++ * ++ * This function is called *before* the dahdi_device is first registered ++ * with the system. What happens in t4_init_one_span can happen between ++ * when the device is registered and when the spans are assigned via ++ * sysfs (or automatically). ++ * ++ */ ++static void init_spans(struct t4 *wc) ++{ ++ int x, y; ++ int gen2; ++ struct t4_span *ts; ++ unsigned int reg; ++ unsigned long flags; ++ ++ gen2 = (wc->tspans[0]->spanflags & FLAG_2NDGEN); ++ for (x = 0; x < wc->numspans; x++) { ++ ts = wc->tspans[x]; ++ if(wc->numspans != 1) { ++ sprintf(ts->span.name, "ALLO%d/%d/%d", wc->numspans, wc->num, x + 1); ++ snprintf(ts->span.desc, sizeof(ts->span.desc) - 1, ++ "T%dXXP (PCI) Card %d Span %d", wc->numspans, wc->num, x+1); ++ } ++ else { ++ sprintf(ts->span.name, "ALLO4/%d/%d", wc->num, x + 1); //cem:1port change ++ snprintf(ts->span.desc, sizeof(ts->span.desc) - 1, ++ "T4XXP (PCI) Card %d Span %d", wc->num, x+1); ++ } ++ switch (ts->linemode) { ++ case T1: ++ ts->span.spantype = SPANTYPE_DIGITAL_T1; ++ break; ++ case E1: ++ ts->span.spantype = SPANTYPE_DIGITAL_E1; ++ break; ++ case J1: ++ ts->span.spantype = SPANTYPE_DIGITAL_J1; ++ break; ++ } ++ ++ /* HDLC Specific init */ ++ ts->sigchan = NULL; ++ ts->sigmode = sigmode; ++ ts->sigactive = 0; ++ ++ if (E1 != ts->linemode) { ++ ts->span.channels = 24; ++ ts->span.deflaw = DAHDI_LAW_MULAW; ++ ts->span.linecompat = DAHDI_CONFIG_AMI | ++ DAHDI_CONFIG_B8ZS | DAHDI_CONFIG_D4 | ++ DAHDI_CONFIG_ESF; ++ } else { ++ ts->span.channels = 31; ++ ts->span.deflaw = DAHDI_LAW_ALAW; ++ ts->span.linecompat = DAHDI_CONFIG_AMI | ++ DAHDI_CONFIG_HDB3 | DAHDI_CONFIG_CCS | ++ DAHDI_CONFIG_CRC4; ++ } ++ ts->span.chans = ts->chans; ++ ts->span.flags = DAHDI_FLAG_RBS; ++ ++ ts->owner = wc; ++ ts->span.offset = x; ++ ts->writechunk = (void *)(wc->writechunk + x * 32 * 2); ++ ts->readchunk = (void *)(wc->readchunk + x * 32 * 2); ++ ++ if (gen2) { ++ ts->span.ops = &t4_gen2_span_ops; ++ } else { ++ ts->span.ops = &t4_gen1_span_ops; ++ } ++ ++ for (y=0;y<wc->tspans[x]->span.channels;y++) { ++ struct dahdi_chan *mychans = ts->chans[y]; ++ sprintf(mychans->name, "ALLO4/%d/%d/%d", wc->num, x + 1, y + 1);//cem:1port change, v002 ++ t4_chan_set_sigcap(&ts->span, x); ++ mychans->pvt = wc; ++ mychans->chanpos = y + 1; ++ } ++ ++ /* Start checking for alarms in 250 ms */ ++ ts->alarmcheck_time = jiffies + msecs_to_jiffies(250); ++ ++ /* Enable 1sec timer interrupt */ ++ spin_lock_irqsave(&wc->reglock, flags); ++ reg = __t4_framer_in(wc, x, FMR1_T); ++ __t4_framer_out(wc, x, FMR1_T, (reg | FMR1_ECM)); ++ ++ /* Enable Errored Second interrupt */ ++ __t4_framer_out(wc, x, ESM, 0); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ t4_reset_counters(&ts->span); ++ ++ } ++ ++ set_span_devicetype(wc); ++ setup_chunks(wc, 0); ++ wc->lastindex = 0; ++} ++ ++static int syncsrc = 0; ++static int syncnum = 0 /* -1 */; ++static int syncspan = 0; ++static DEFINE_SPINLOCK(synclock); ++ ++static void __t4_set_rclk_src(struct t4 *wc, int span) ++{ ++ if (is_octal(wc)) { ++ int cmr5 = 0x00 | (span << 5); ++ int cmr1 = 0x38; /* Clock Mode: RCLK sourced by DCO-R1 ++ by default, Disable Clock-Switching */ ++ ++ __t4_framer_out(wc, 0, 0x44, cmr1); ++ __t4_framer_out(wc, 0, FRMR_CMR5, cmr5); ++ } else { ++ int cmr1 = 0x38; /* Clock Mode: RCLK sourced by DCO-R1 ++ by default, Disable Clock-Switching */ ++ cmr1 |= (span << 6); ++ __t4_framer_out(wc, 0, 0x44, cmr1); ++ } ++ ++ dev_info(&wc->dev->dev, "RCLK source set to span %d\n", span+1); ++} ++ ++static void __t4_set_sclk_src(struct t4 *wc, int mode, int master, int slave) ++{ ++ if (slave) { ++ wc->dmactrl |= (1 << 25); ++ dev_info(&wc->dev->dev, "SCLK is slaved to timing cable\n"); ++ } else { ++ wc->dmactrl &= ~(1 << 25); ++ } ++ ++ if (master) { ++ wc->dmactrl |= (1 << 24); ++ dev_info(&wc->dev->dev, "SCLK is master to timing cable\n"); ++ } else { ++ wc->dmactrl &= ~(1 << 24); ++ } ++ ++ if (mode == WC_RECOVER) ++ wc->dmactrl |= (1 << 29); /* Recover timing from RCLK */ ++ ++ if (mode == WC_SELF) ++ wc->dmactrl &= ~(1 << 29);/* Provide timing from MCLK */ ++ ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++} ++ ++static ssize_t t4_timing_master_show(struct device *dev, ++ struct device_attribute *attr, ++ char *buf) ++{ ++ struct t4 *wc = dev_get_drvdata(dev); ++ if (wc->dmactrl & (1 << 29)) ++ return sprintf(buf, "%d\n", wc->syncsrc); ++ else ++ return sprintf(buf, "%d\n", -1); ++} ++ ++static DEVICE_ATTR(timing_master, 0400, t4_timing_master_show, NULL); ++ ++static void create_sysfs_files(struct t4 *wc) ++{ ++ int ret; ++ ret = device_create_file(&wc->dev->dev, ++ &dev_attr_timing_master); ++ if (ret) { ++ dev_info(&wc->dev->dev, ++ "Failed to create device attributes.\n"); ++ } ++} ++ ++static void remove_sysfs_files(struct t4 *wc) ++{ ++ device_remove_file(&wc->dev->dev, ++ &dev_attr_timing_master); ++} ++ ++static inline void __t4_update_timing(struct t4 *wc) ++{ ++ int i; ++ /* update sync src info */ ++ if (wc->syncsrc != syncsrc) { ++ dev_info(&wc->dev->dev, "Swapping card %d from %d to %d\n", ++ wc->num, wc->syncsrc, syncsrc); ++ wc->syncsrc = syncsrc; ++ /* Update sync sources */ ++ for (i = 0; i < wc->numspans; i++) { ++ wc->tspans[i]->span.syncsrc = wc->syncsrc; ++ } ++ if (syncnum == wc->num) { ++ __t4_set_rclk_src(wc, syncspan-1); ++ __t4_set_sclk_src(wc, WC_RECOVER, 1, 0); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Card %d, using sync " ++ "span %d, master\n", wc->num, syncspan); ++ } else { ++ __t4_set_sclk_src(wc, WC_RECOVER, 0, 1); ++ if (debug) ++ dev_notice(&wc->dev->dev, "Card %d, using " ++ "Timing Bus, NOT master\n", wc->num); ++ } ++ } ++} ++ ++static int __t4_findsync(struct t4 *wc) ++{ ++ int i; ++ int x; ++ unsigned long flags; ++ int p; ++ int nonzero; ++ int newsyncsrc = 0; /* DAHDI span number */ ++ int newsyncnum = 0; /* allo4xxp card number */ ++ int newsyncspan = 0; /* span on given allo4xxp card */ ++ spin_lock_irqsave(&synclock, flags); ++ if (!wc->num) { ++ /* If we're the first card, go through all the motions, up to 8 levels ++ of sync source */ ++ p = 1; ++ while (p < 8) { ++ nonzero = 0; ++ for (x=0;cards[x];x++) { ++ for (i = 0; i < cards[x]->numspans; i++) { ++ if (cards[x]->tspans[i]->syncpos) { ++ nonzero = 1; ++ if ((cards[x]->tspans[i]->syncpos == p) && ++ !(cards[x]->tspans[i]->span.alarms & (DAHDI_ALARM_RED | DAHDI_ALARM_BLUE | DAHDI_ALARM_LOOPBACK)) && ++ (cards[x]->tspans[i]->span.flags & DAHDI_FLAG_RUNNING)) { ++ /* This makes a good sync source */ ++ newsyncsrc = cards[x]->tspans[i]->span.spanno; ++ newsyncnum = x; ++ newsyncspan = i + 1; ++ /* Jump out */ ++ goto found; ++ } ++ } ++ } ++ } ++ if (nonzero) ++ p++; ++ else ++ break; ++ } ++found: ++ if ((syncnum != newsyncnum) || (syncsrc != newsyncsrc) || (newsyncspan != syncspan)) { ++ if (debug) ++ dev_notice(&wc->dev->dev, "New syncnum: %d " ++ "(was %d), syncsrc: %d (was %d), " ++ "syncspan: %d (was %d)\n", newsyncnum, ++ syncnum, newsyncsrc, syncsrc, ++ newsyncspan, syncspan); ++ syncnum = newsyncnum; ++ syncsrc = newsyncsrc; ++ syncspan = newsyncspan; ++ for (x=0;cards[x];x++) { ++ __t4_update_timing(cards[x]); ++ } ++ } ++ } ++ __t4_update_timing(wc); ++ spin_unlock_irqrestore(&synclock, flags); ++ return 0; ++} ++ ++static void __t4_set_timing_source_auto(struct t4 *wc) ++{ ++ int x, i; ++ int firstprio, secondprio; ++ firstprio = secondprio = 4; ++ ++ if (debug) ++ dev_info(&wc->dev->dev, "timing source auto\n"); ++ clear_bit(T4_CHECK_TIMING, &wc->checkflag); ++ if (timingcable) { ++ __t4_findsync(wc); ++ } else { ++ if (debug) ++ dev_info(&wc->dev->dev, "Evaluating spans for timing " ++ "source\n"); ++ for (x=0;x<wc->numspans;x++) { ++ if ((wc->tspans[x]->span.flags & DAHDI_FLAG_RUNNING) && ++ !(wc->tspans[x]->span.alarms & (DAHDI_ALARM_RED | ++ DAHDI_ALARM_BLUE))) { ++ if (debug) ++ dev_info(&wc->dev->dev, "span %d is " ++ "green : syncpos %d\n", x+1, ++ wc->tspans[x]->syncpos); ++ if (wc->tspans[x]->syncpos) { ++ /* Valid rsync source in recovered ++ timing mode */ ++ if (firstprio == 4) ++ firstprio = x; ++ else if (wc->tspans[x]->syncpos < ++ wc->tspans[firstprio]->syncpos) ++ firstprio = x; ++ } else { ++ /* Valid rsync source in system timing ++ mode */ ++ if (secondprio == 4) ++ secondprio = x; ++ } ++ } ++ } ++ if (firstprio != 4) { ++ wc->syncsrc = firstprio; ++ __t4_set_rclk_src(wc, firstprio); ++ __t4_set_sclk_src(wc, WC_RECOVER, 0, 0); ++ dev_info(&wc->dev->dev, "Recovered timing mode, "\ ++ "RCLK set to span %d\n", ++ firstprio+1); ++ } else if (secondprio != 4) { ++ wc->syncsrc = -1; ++ __t4_set_rclk_src(wc, secondprio); ++ __t4_set_sclk_src(wc, WC_SELF, 0, 0); ++ dev_info(&wc->dev->dev, "System timing mode, "\ ++ "RCLK set to span %d\n", ++ secondprio+1); ++ } else { ++ wc->syncsrc = -1; ++ dev_info(&wc->dev->dev, "All spans in alarm : No valid"\ ++ "span to source RCLK from\n"); ++ /* Default rclk to lock with span 1 */ ++ __t4_set_rclk_src(wc, 0); ++ __t4_set_sclk_src(wc, WC_SELF, 0, 0); ++ } ++ ++ /* Propagate sync selection to dahdi_span struct ++ * this is read by dahdi_tool to display the span's ++ * master/slave sync information */ ++ for (i = 0; i < wc->numspans; i++) { ++ wc->tspans[i]->span.syncsrc = wc->syncsrc + 1; ++ } ++ } ++} ++ ++static void __t4_configure_t1(struct t4 *wc, int unit, int lineconfig, int txlevel) ++{ ++ unsigned int fmr4, fmr2, fmr1, fmr0, lim2; ++ char *framing, *line; ++ int mytxlevel; ++ if ((txlevel > 7) || (txlevel < 4)) ++ mytxlevel = 0; ++ else ++ mytxlevel = txlevel - 4; ++ ++ if (is_octal(wc)) ++ fmr1 = 0x9c | 0x02; /* FMR1: Mode 1, T1 mode, CRC on for ESF, 8.192 Mhz system data rate, no XAIS */ ++ else ++ fmr1 = 0x9c; /* FMR1: Mode 1, T1 mode, CRC on for ESF, 8.192 Mhz system data rate, no XAIS */ ++ ++ fmr2 = 0x20; /* FMR2: no payload loopback, don't auto yellow */ ++ fmr4 = 0x0c; /* FMR4: Lose sync on 2 out of 5 framing bits, auto resync */ ++ lim2 = 0x21; /* LIM2: 50% peak is a "1", Advanced Loss recovery */ ++ lim2 |= (mytxlevel << 6); /* LIM2: Add line buildout */ ++ __t4_framer_out(wc, unit, 0x1d, fmr1); ++ __t4_framer_out(wc, unit, 0x1e, fmr2); ++ ++ /* Configure line interface */ ++ if (lineconfig & DAHDI_CONFIG_AMI) { ++ line = "AMI"; ++ /* workaround for errata #2 in ES v3 09-10-16 */ ++ fmr0 = (wc->falc31) ? 0xb0 : 0xa0; ++ } else { ++ line = "B8ZS"; ++ fmr0 = 0xf0; ++ } ++ if (lineconfig & DAHDI_CONFIG_D4) { ++ framing = "D4"; ++ } else { ++ framing = "ESF"; ++ fmr4 |= 0x2; ++ fmr2 |= 0xc0; ++ } ++ __t4_framer_out(wc, unit, 0x1c, fmr0); ++ __t4_framer_out(wc, unit, 0x20, fmr4); ++ __t4_framer_out(wc, unit, FMR5, FMR5_EIBR); /* FMR5: Enable RBS mode */ ++ ++ __t4_framer_out(wc, unit, 0x37, 0xf0 ); /* LIM1: Clear data in case of LOS, Set receiver threshold (0.5V), No remote loop, no DRS */ ++ __t4_framer_out(wc, unit, 0x36, 0x08); /* LIM0: Enable auto long haul mode, no local loop (must be after LIM1) */ ++ ++ __t4_framer_out(wc, unit, 0x02, 0x50); /* CMDR: Reset the receiver and transmitter line interface */ ++ __t4_framer_out(wc, unit, 0x02, 0x00); /* CMDR: Reset the receiver and transmitter line interface */ ++ ++ if (wc->falc31) { ++ if (debug) ++ dev_info(&wc->dev->dev, "card %d span %d: setting Rtx " ++ "to 0ohm for T1\n", wc->num, unit); ++ __t4_framer_out(wc, unit, 0x86, 0x00); /* PC6: set Rtx to 0ohm for T1 */ ++ ++ // Hitting the bugfix register to fix errata #3 ++ __t4_framer_out(wc, unit, 0xbd, 0x05); ++ } ++ ++ __t4_framer_out(wc, unit, 0x3a, lim2); /* LIM2: 50% peak amplitude is a "1" */ ++ __t4_framer_out(wc, unit, 0x38, 0x0a); /* PCD: LOS after 176 consecutive "zeros" */ ++ __t4_framer_out(wc, unit, 0x39, 0x15); /* PCR: 22 "ones" clear LOS */ ++ ++ /* Generate pulse mask for T1 */ ++ switch(mytxlevel) { ++ case 3: ++ __t4_framer_out(wc, unit, 0x26, 0x07); /* XPM0 */ ++ __t4_framer_out(wc, unit, 0x27, 0x01); /* XPM1 */ ++ __t4_framer_out(wc, unit, 0x28, 0x00); /* XPM2 */ ++ break; ++ case 2: ++ __t4_framer_out(wc, unit, 0x26, 0x8c); /* XPM0 */ ++ __t4_framer_out(wc, unit, 0x27, 0x11); /* XPM1 */ ++ __t4_framer_out(wc, unit, 0x28, 0x01); /* XPM2 */ ++ break; ++ case 1: ++ __t4_framer_out(wc, unit, 0x26, 0x8c); /* XPM0 */ ++ __t4_framer_out(wc, unit, 0x27, 0x01); /* XPM1 */ ++ __t4_framer_out(wc, unit, 0x28, 0x00); /* XPM2 */ ++ break; ++ case 0: ++ default: ++ __t4_framer_out(wc, unit, 0x26, 0xd7); /* XPM0 */ ++ __t4_framer_out(wc, unit, 0x27, 0x22); /* XPM1 */ ++ __t4_framer_out(wc, unit, 0x28, 0x01); /* XPM2 */ ++ break; ++ } ++ ++ /* Don't mask framer interrupts if hardware HDLC is in use */ ++ __t4_framer_out(wc, unit, FRMR_IMR0, 0xff & ~((wc->tspans[unit]->sigchan) ? HDLC_IMR0_MASK : 0)); /* IMR0: We care about CAS changes, etc */ ++ __t4_framer_out(wc, unit, FRMR_IMR1, 0xff & ~((wc->tspans[unit]->sigchan) ? HDLC_IMR1_MASK : 0)); /* IMR1: We care about nothing */ ++ __t4_framer_out(wc, unit, 0x16, 0x00); /* IMR2: All the alarm stuff! */ ++ __t4_framer_out(wc, unit, 0x17, 0x34); /* IMR3: AIS and friends */ ++ __t4_framer_out(wc, unit, 0x18, 0x3f); /* IMR4: Slips on transmit */ ++ ++ dev_info(&wc->dev->dev, "Span %d configured for %s/%s\n", unit + 1, ++ framing, line); ++} ++ ++static void __t4_configure_e1(struct t4 *wc, int unit, int lineconfig) ++{ ++ unsigned int fmr2, fmr1, fmr0; ++ unsigned int cas = 0; ++ unsigned int imr3extra=0; ++ char *crc4 = ""; ++ char *framing, *line; ++ if (is_octal(wc)) { ++ /* 16 MHz */ ++ fmr1 = 0x44 | 0x02; /* FMR1: E1 mode, Automatic force resync, PCM30 mode, 8.192 Mhz backplane, no XAIS */ ++ } else { ++ /* 8 MHz */ ++ fmr1 = 0x44; /* FMR1: E1 mode, Automatic force resync, PCM30 mode, 8.192 Mhz backplane, no XAIS */ ++ } ++ fmr2 = 0x03; /* FMR2: Auto transmit remote alarm, auto loss of multiframe recovery, no payload loopback */ ++ if (lineconfig & DAHDI_CONFIG_CRC4) { ++ fmr1 |= 0x08; /* CRC4 transmit */ ++ fmr2 |= 0xc0; /* CRC4 receive */ ++ crc4 = "/CRC4"; ++ } ++ __t4_framer_out(wc, unit, 0x1d, fmr1); ++ __t4_framer_out(wc, unit, 0x1e, fmr2); ++ ++ /* Configure line interface */ ++ if (lineconfig & DAHDI_CONFIG_AMI) { ++ line = "AMI"; ++ /* workaround for errata #2 in ES v3 09-10-16 */ ++ fmr0 = (wc->falc31) ? 0xb0 : 0xa0; ++ } else { ++ line = "HDB3"; ++ fmr0 = 0xf0; ++ } ++ if (lineconfig & DAHDI_CONFIG_CCS) { ++ framing = "CCS"; ++ imr3extra = 0x28; ++ } else { ++ framing = "CAS"; ++ cas = 0x40; ++ } ++ __t4_framer_out(wc, unit, 0x1c, fmr0); ++ ++ __t4_framer_out(wc, unit, 0x37, 0xf0 /*| 0x6 */ ); /* LIM1: Clear data in case of LOS, Set receiver threshold (0.5V), No remote loop, no DRS */ ++ __t4_framer_out(wc, unit, 0x36, 0x08); /* LIM0: Enable auto long haul mode, no local loop (must be after LIM1) */ ++ ++ __t4_framer_out(wc, unit, 0x02, 0x50); /* CMDR: Reset the receiver and transmitter line interface */ ++ __t4_framer_out(wc, unit, 0x02, 0x00); /* CMDR: Reset the receiver and transmitter line interface */ ++ ++ if (wc->falc31) { ++ if (debug) ++ dev_info(&wc->dev->dev, ++ "setting Rtx to 7.5ohm for E1\n"); ++ __t4_framer_out(wc, unit, 0x86, 0x40); /* PC6: turn on 7.5ohm Rtx for E1 */ ++ } ++ ++ /* Condition receive line interface for E1 after reset */ ++ __t4_framer_out(wc, unit, 0xbb, 0x17); ++ __t4_framer_out(wc, unit, 0xbc, 0x55); ++ __t4_framer_out(wc, unit, 0xbb, 0x97); ++ __t4_framer_out(wc, unit, 0xbb, 0x11); ++ __t4_framer_out(wc, unit, 0xbc, 0xaa); ++ __t4_framer_out(wc, unit, 0xbb, 0x91); ++ __t4_framer_out(wc, unit, 0xbb, 0x12); ++ __t4_framer_out(wc, unit, 0xbc, 0x55); ++ __t4_framer_out(wc, unit, 0xbb, 0x92); ++ __t4_framer_out(wc, unit, 0xbb, 0x0c); ++ __t4_framer_out(wc, unit, 0xbb, 0x00); ++ __t4_framer_out(wc, unit, 0xbb, 0x8c); ++ ++ __t4_framer_out(wc, unit, 0x3a, 0x20); /* LIM2: 50% peak amplitude is a "1" */ ++ __t4_framer_out(wc, unit, 0x38, 0x0a); /* PCD: LOS after 176 consecutive "zeros" */ ++ __t4_framer_out(wc, unit, 0x39, 0x15); /* PCR: 22 "ones" clear LOS */ ++ ++ __t4_framer_out(wc, unit, 0x20, 0x9f); /* XSW: Spare bits all to 1 */ ++ __t4_framer_out(wc, unit, 0x21, 0x1c|cas); /* XSP: E-bit set when async. AXS auto, XSIF to 1 */ ++ ++ ++ /* Generate pulse mask for E1 */ ++ __t4_framer_out(wc, unit, 0x26, 0x54); /* XPM0 */ ++ __t4_framer_out(wc, unit, 0x27, 0x02); /* XPM1 */ ++ __t4_framer_out(wc, unit, 0x28, 0x00); /* XPM2 */ ++ ++ /* Don't mask framer interrupts if hardware HDLC is in use */ ++ __t4_framer_out(wc, unit, FRMR_IMR0, 0xff & ~((wc->tspans[unit]->sigchan) ? HDLC_IMR0_MASK : 0)); /* IMR0: We care about CRC errors, CAS changes, etc */ ++ __t4_framer_out(wc, unit, FRMR_IMR1, 0x3f & ~((wc->tspans[unit]->sigchan) ? HDLC_IMR1_MASK : 0)); /* IMR1: We care about loopup / loopdown */ ++ __t4_framer_out(wc, unit, 0x16, 0x00); /* IMR2: We care about all the alarm stuff! */ ++ __t4_framer_out(wc, unit, 0x17, 0x04 | imr3extra); /* IMR3: AIS */ ++ __t4_framer_out(wc, unit, 0x18, 0x3f); /* IMR4: We care about slips on transmit */ ++ ++ __t4_framer_out(wc, unit, 0x2f, 0x00); ++ __t4_framer_out(wc, unit, 0x30, 0x00); ++ __t4_framer_out(wc, unit, 0x31, 0x00); ++ ++ dev_info(&wc->dev->dev, "ALLO%dXXP: Span %d configured for %s/%s%s\n", ++ wc->numspans, unit + 1, framing, line, crc4); ++} ++ ++/** ++ * t4_check_for_interrupts - Return 0 if the card is generating interrupts. ++ * @wc: The card to check. ++ * ++ * If the card is not generating interrupts, this function will also place all ++ * the spans on the card into red alarm. ++ * ++ */ ++static int t4_check_for_interrupts(struct t4 *wc) ++{ ++ unsigned int starting_intcount = wc->intcount; ++ unsigned long stop_time = jiffies + HZ*2; ++ unsigned long flags; ++ int x; ++ ++ msleep(20); ++ spin_lock_irqsave(&wc->reglock, flags); ++ while (starting_intcount == wc->intcount) { ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ if (time_after(jiffies, stop_time)) { ++ for (x = 0; x < wc->numspans; x++) ++ wc->tspans[x]->span.alarms = DAHDI_ALARM_RED; ++ dev_err(&wc->dev->dev, "Interrupts not detected.\n"); ++ return -EIO; ++ } ++ msleep(100); ++ spin_lock_irqsave(&wc->reglock, flags); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ return 0; ++} ++ ++static int _t4_startup(struct file *file, struct dahdi_span *span) ++{ ++#ifdef SUPPORT_GEN1 ++ int i; ++#endif ++ int tspan; ++ unsigned long flags; ++ int alreadyrunning; ++ struct t4_span *ts = container_of(span, struct t4_span, span); ++ struct t4 *wc = ts->owner; ++ ++ set_bit(T4_IGNORE_LATENCY, &wc->checkflag); ++ if (debug) ++ dev_info(&wc->dev->dev, "About to enter startup!\n"); ++ ++ tspan = span->offset + 1; ++ if (tspan < 0) { ++ dev_info(&wc->dev->dev, "ALLO%dXXP: Span '%d' isn't us?\n", ++ wc->numspans, span->spanno); ++ return -1; ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ alreadyrunning = span->flags & DAHDI_FLAG_RUNNING; ++ ++#ifdef SUPPORT_GEN1 ++ /* initialize the start value for the entire chunk of last ec buffer */ ++ for(i = 0; i < span->channels; i++) ++ { ++ memset(ts->ec_chunk1[i], ++ DAHDI_LIN2X(0,span->chans[i]),DAHDI_CHUNKSIZE); ++ memset(ts->ec_chunk2[i], ++ DAHDI_LIN2X(0,span->chans[i]),DAHDI_CHUNKSIZE); ++ } ++#endif ++ /* Force re-evaluation of timing source */ ++ wc->syncsrc = -1; ++ set_bit(T4_CHECK_TIMING, &wc->checkflag); ++ ++ if (E1 == ts->linemode) ++ __t4_configure_e1(wc, span->offset, span->lineconfig); ++ else ++ __t4_configure_t1(wc, span->offset, span->lineconfig, span->txlevel); ++ ++ /* Note clear channel status */ ++ wc->tspans[span->offset]->notclear = 0; ++ __set_clear(wc, span->offset); ++ ++ if (!alreadyrunning) { ++ span->flags |= DAHDI_FLAG_RUNNING; ++ wc->spansstarted++; ++ ++ if (wc->devtype->flags & FLAG_5THGEN) ++ __t4_pci_out(wc, 5, (ms_per_irq << 16) | wc->numbufs); ++ else ++ __t4_pci_out(wc, 5, (1 << 16) | 1); ++ /* enable interrupts */ ++ /* Start DMA, enabling DMA interrupts on read only */ ++ wc->dmactrl |= (ts->spanflags & FLAG_2NDGEN) ? 0xc0000000 : 0xc0000003; ++#ifdef VPM_SUPPORT ++ wc->dmactrl |= (wc->vpm) ? T4_VPM_PRESENT : 0; ++#endif ++ /* Seed interrupt register */ ++ __t4_pci_out(wc, WC_INTR, 0x0c); ++ if (noburst || !(ts->spanflags & FLAG_BURST)) ++ wc->dmactrl |= (1 << 26); ++ else ++ wc->dmactrl &= ~(1 << 26); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ ++ /* Startup HDLC controller too */ ++ } ++ ++ if (ts->sigchan) { ++ struct dahdi_chan *sigchan = ts->sigchan; ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ if (hdlc_start(wc, span->offset, sigchan, ts->sigmode)) { ++ dev_notice(&wc->dev->dev, "Error initializing " ++ "signalling controller\n"); ++ return -1; ++ } ++ spin_lock_irqsave(&wc->reglock, flags); ++ } ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ local_irq_save(flags); ++ t4_check_alarms(wc, span->offset); ++ t4_check_sigbits(wc, span->offset); ++ local_irq_restore(flags); ++ if (wc->numspans >= 1) { /*cem:1port change, v002*/ ++ if (wc->tspans[0]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Primary Sync Source\n", ++ span->spanno); ++ } ++ if (wc->tspans[1]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Secondary Sync Source\n", ++ span->spanno); ++ if (wc->numspans >= 4) { ++ if (wc->tspans[2]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Tertiary Sync Source" ++ "\n", span->spanno); ++ if (wc->tspans[3]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Quaternary Sync " ++ "Source\n", span->spanno); ++ } ++ if (wc->numspans == 8) { ++ if (wc->tspans[4]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Quinary Sync " ++ "Source\n", span->spanno); ++ if (wc->tspans[5]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Senary Sync " ++ "Source\n", span->spanno); ++ if (wc->tspans[6]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Septenary Sync " ++ "Source\n", span->spanno); ++ if (wc->tspans[7]->sync == span->spanno) ++ dev_info(&wc->dev->dev, "SPAN %d: Octonary Sync " ++ "Source\n", span->spanno); ++ } ++ ++ if (!alreadyrunning) { ++ if (t4_check_for_interrupts(wc)) ++ return -EIO; ++ } ++ ++ if (debug) ++ dev_info(&wc->dev->dev, "Completed startup!\n"); ++ clear_bit(T4_IGNORE_LATENCY, &wc->checkflag); ++ return 0; ++} ++ ++static int t4_startup(struct file *file, struct dahdi_span *span) ++{ ++ int ret; ++ struct dahdi_device *const ddev = span->parent; ++ struct dahdi_span *s; ++ ++ ret = _t4_startup(file, span); ++ list_for_each_entry(s, &ddev->spans, device_node) { ++ if (!test_bit(DAHDI_FLAGBIT_RUNNING, &s->flags)) { ++ _t4_startup(file, s); ++ } ++ } ++ return ret; ++} ++ ++#ifdef SUPPORT_GEN1 ++static inline void e1_check(struct t4 *wc, int span, int val) ++{ ++ struct t4_span *ts = wc->tspans[span]; ++ if ((ts->span.channels > 24) && ++ (ts->span.flags & DAHDI_FLAG_RUNNING) && ++ !(ts->span.alarms) && ++ (!wc->e1recover)) { ++ if (val != 0x1b) { ++ ts->e1check++; ++ } else ++ ts->e1check = 0; ++ if (ts->e1check > 100) { ++ /* Wait 1000 ms */ ++ wc->e1recover = 1000 * 8; ++ wc->tspans[0]->e1check = wc->tspans[1]->e1check = 0; ++ if (wc->numspans == 4) ++ wc->tspans[2]->e1check = wc->tspans[3]->e1check = 0; ++ if (debug & DEBUG_MAIN) ++ dev_notice(&wc->dev->dev, "Detected loss of " ++ "E1 alignment on span %d!\n", span); ++ t4_reset_dma(wc); ++ } ++ } ++} ++ ++static void t4_receiveprep(struct t4 *wc, int irq) ++{ ++ unsigned int *readchunk; ++ int dbl = 0; ++ int x,y,z; ++ unsigned int tmp; ++ int offset=0; ++ if (!has_e1_span(wc)) ++ offset = 4; ++ if (irq & 1) { ++ /* First part */ ++ readchunk = wc->readchunk; ++ if (!wc->last0) ++ dbl = 1; ++ wc->last0 = 0; ++ } else { ++ readchunk = wc->readchunk + DAHDI_CHUNKSIZE * 32; ++ if (wc->last0) ++ dbl = 1; ++ wc->last0 = 1; ++ } ++ if (unlikely(dbl && (debug & DEBUG_MAIN))) ++ dev_notice(&wc->dev->dev, "Double/missed interrupt detected\n"); ++ ++ for (x=0;x<DAHDI_CHUNKSIZE;x++) { ++ for (z=0;z<24;z++) { ++ /* All T1/E1 channels */ ++ tmp = readchunk[z+1+offset]; ++ if (wc->numspans == 4) { ++ wc->tspans[3]->span.chans[z]->readchunk[x] = tmp & 0xff; ++ wc->tspans[2]->span.chans[z]->readchunk[x] = (tmp & 0xff00) >> 8; ++ } ++ if (wc->numspans == 2) { /*cem:1port change, v002*/ ++ wc->tspans[1]->span.chans[z]->readchunk[x] = (tmp & 0xff0000) >> 16; ++ } ++ wc->tspans[0]->span.chans[z]->readchunk[x] = tmp >> 24; ++ } ++ if (has_e1_span(wc)) { ++ if (wc->e1recover > 0) ++ wc->e1recover--; ++ tmp = readchunk[0]; ++ if (wc->numspans == 4) { ++ e1_check(wc, 3, (tmp & 0x7f)); ++ e1_check(wc, 2, (tmp & 0x7f00) >> 8); ++ } ++ if (wc->numspans >=2) //cem:1port change ++ e1_check(wc, 1, (tmp & 0x7f0000) >> 16); ++ e1_check(wc, 0, (tmp & 0x7f000000) >> 24); ++ for (z=24;z<31;z++) { ++ /* Only E1 channels now */ ++ tmp = readchunk[z+1]; ++ if (wc->numspans == 4) { ++ if (wc->tspans[3]->span.channels > 24) ++ wc->tspans[3]->span.chans[z]->readchunk[x] = tmp & 0xff; ++ if (wc->tspans[2]->span.channels > 24) ++ wc->tspans[2]->span.chans[z]->readchunk[x] = (tmp & 0xff00) >> 8; ++ } ++ if (wc->numspans == 2) {/*cem:1port change, v002*/ ++ if (wc->tspans[1]->span.channels > 24) ++ wc->tspans[1]->span.chans[z]->readchunk[x] = (tmp & 0xff0000) >> 16; ++ } ++ if (wc->tspans[0]->span.channels > 24) ++ wc->tspans[0]->span.chans[z]->readchunk[x] = tmp >> 24; ++ } ++ } ++ /* Advance pointer by 4 TDM frame lengths */ ++ readchunk += 32; ++ } ++ for (x=0;x<wc->numspans;x++) { ++ if (wc->tspans[x]->span.flags & DAHDI_FLAG_RUNNING) { ++ for (y=0;y<wc->tspans[x]->span.channels;y++) { ++ /* Echo cancel double buffered data */ ++ dahdi_ec_chunk(wc->tspans[x]->span.chans[y], ++ wc->tspans[x]->span.chans[y]->readchunk, ++ wc->tspans[x]->ec_chunk2[y]); ++ memcpy(wc->tspans[x]->ec_chunk2[y],wc->tspans[x]->ec_chunk1[y], ++ DAHDI_CHUNKSIZE); ++ memcpy(wc->tspans[x]->ec_chunk1[y], ++ wc->tspans[x]->span.chans[y]->writechunk, ++ DAHDI_CHUNKSIZE); ++ } ++ _dahdi_receive(&wc->tspans[x]->span); ++ } ++ } ++} ++#endif ++ ++#if (DAHDI_CHUNKSIZE != 8) ++#error Sorry, nextgen does not support chunksize != 8 ++#endif ++ ++static void __receive_span(struct t4_span *ts) ++{ ++#ifdef VPM_SUPPORT ++ int y; ++ unsigned long merged; ++ merged = ts->dtmfactive & ts->dtmfmutemask; ++ if (merged) { ++ for (y=0;y<ts->span.channels;y++) { ++ /* Mute any DTMFs which are supposed to be muted */ ++ if (test_bit(y, &merged)) { ++ memset(ts->span.chans[y]->readchunk, DAHDI_XLAW(0, ts->span.chans[y]), DAHDI_CHUNKSIZE); ++ } ++ } ++ } ++#endif ++ _dahdi_ec_span(&ts->span); ++ _dahdi_receive(&ts->span); ++} ++ ++static inline void __transmit_span(struct t4_span *ts) ++{ ++ _dahdi_transmit(&ts->span); ++} ++ ++#ifdef ENABLE_WORKQUEUES ++static void workq_handlespan(void *data) ++{ ++ struct t4_span *ts = data; ++ struct t4 *wc = ts->owner; ++ ++ __receive_span(ts); ++ __transmit_span(ts); ++ atomic_dec(&wc->worklist); ++ if (!atomic_read(&wc->worklist)) ++ t4_pci_out(wc, WC_INTR, 0); ++} ++#else ++static void t4_prep_gen2(struct t4 *wc) ++{ ++ int x; ++ for (x=0;x<wc->numspans;x++) { ++ if (wc->tspans[x]->span.flags & DAHDI_FLAG_RUNNING) { ++ __receive_span(wc->tspans[x]); ++ __transmit_span(wc->tspans[x]); ++ } ++ } ++} ++ ++#endif ++#ifdef SUPPORT_GEN1 ++static void t4_transmitprep(struct t4 *wc, int irq) ++{ ++ u32 *writechunk; ++ int x, y, z; ++ unsigned int tmp; ++ int offset = 0; ++ if (!has_e1_span(wc)) ++ offset = 4; ++ if (irq & 1) { ++ /* First part */ ++ writechunk = wc->writechunk + 1; ++ } else { ++ writechunk = wc->writechunk + DAHDI_CHUNKSIZE * 32 + 1; ++ } ++ for (y=0;y<wc->numspans;y++) { ++ if (wc->tspans[y]->span.flags & DAHDI_FLAG_RUNNING) ++ _dahdi_transmit(&wc->tspans[y]->span); ++ } ++ ++ for (x=0;x<DAHDI_CHUNKSIZE;x++) { ++ /* Once per chunk */ ++ for (z=0;z<24;z++) { ++ /* All T1/E1 channels */ ++ tmp = (wc->tspans[3]->span.chans[z]->writechunk[x]) | ++ (wc->tspans[2]->span.chans[z]->writechunk[x] << 8) | ++ (wc->tspans[1]->span.chans[z]->writechunk[x] << 16) | ++ (wc->tspans[0]->span.chans[z]->writechunk[x] << 24); ++ writechunk[z+offset] = tmp; ++ } ++ if (has_e1_span(wc)) { ++ for (z=24;z<31;z++) { ++ /* Only E1 channels now */ ++ tmp = 0; ++ if (wc->numspans == 4) { ++ if (wc->tspans[3]->span.channels > 24) ++ tmp |= wc->tspans[3]->span.chans[z]->writechunk[x]; ++ if (wc->tspans[2]->span.channels > 24) ++ tmp |= (wc->tspans[2]->span.chans[z]->writechunk[x] << 8); ++ } ++ if (wc->numspans == 2) { /*cem:1port change, v002*/ ++ if (wc->tspans[1]->span.channels > 24) ++ tmp |= (wc->tspans[1]->span.chans[z]->writechunk[x] << 16); ++ } ++ if (wc->tspans[0]->span.channels > 24) ++ tmp |= (wc->tspans[0]->span.chans[z]->writechunk[x] << 24); ++ writechunk[z] = tmp; ++ } ++ } ++ /* Advance pointer by 4 TDM frame lengths */ ++ writechunk += 32; ++ } ++ ++} ++#endif ++ ++static void t4_dahdi_rbsbits(struct dahdi_chan *const chan, int rxs) ++{ ++ if ((debug & DEBUG_RBS) && printk_ratelimit()) { ++ const struct t4_span *tspan = container_of(chan->span, ++ struct t4_span, ++ span); ++ const struct t4 *const wc = tspan->owner; ++ dev_notice(&wc->dev->dev, "Detected sigbits change on " \ ++ "channel %s to %04x\n", chan->name, rxs); ++ } ++ dahdi_rbsbits(chan, rxs); ++} ++ ++static void t4_check_sigbits(struct t4 *wc, int span) ++{ ++ int a,i,rxs; ++ struct t4_span *ts = wc->tspans[span]; ++ ++ if (debug & DEBUG_RBS) ++ dev_notice(&wc->dev->dev, "Checking sigbits on span %d\n", ++ span + 1); ++ ++ if (!(ts->span.flags & DAHDI_FLAG_RUNNING)) ++ return; ++ if (E1 == ts->linemode) { ++ for (i = 0; i < 15; i++) { ++ a = t4_framer_in(wc, span, 0x71 + i); ++ /* Get high channel in low bits */ ++ rxs = (a & 0xf); ++ if (!(ts->span.chans[i+16]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i+16]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i+16], rxs); ++ } ++ rxs = (a >> 4) & 0xf; ++ if (!(ts->span.chans[i]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i], rxs); ++ } ++ } ++ } else if (ts->span.lineconfig & DAHDI_CONFIG_D4) { ++ for (i = 0; i < 24; i+=4) { ++ a = t4_framer_in(wc, span, 0x70 + (i>>2)); ++ /* Get high channel in low bits */ ++ rxs = (a & 0x3) << 2; ++ if (!(ts->span.chans[i+3]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i+3]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i+3], rxs); ++ } ++ rxs = (a & 0xc); ++ if (!(ts->span.chans[i+2]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i+2]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i+2], rxs); ++ } ++ rxs = (a >> 2) & 0xc; ++ if (!(ts->span.chans[i+1]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i+1]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i+1], rxs); ++ } ++ rxs = (a >> 4) & 0xc; ++ if (!(ts->span.chans[i]->sig & DAHDI_SIG_CLEAR)) { ++ if (ts->span.chans[i]->rxsig != rxs) ++ t4_dahdi_rbsbits(ts->span.chans[i], rxs); ++ } ++ } ++ } else { ++ for (i = 0; i < 24; i+=2) { ++ a = t4_framer_in(wc, span, 0x70 + (i>>1)); ++ /* Get high channel in low bits */ ++ rxs = (a & 0xf); ++ if (!(ts->span.chans[i+1]->sig & DAHDI_SIG_CLEAR)) { ++ /* XXX Not really reset on every trans! XXX */ ++ if (ts->span.chans[i+1]->rxsig != rxs) { ++ t4_dahdi_rbsbits(ts->span.chans[i+1], rxs); ++ } ++ } ++ rxs = (a >> 4) & 0xf; ++ if (!(ts->span.chans[i]->sig & DAHDI_SIG_CLEAR)) { ++ /* XXX Not really reset on every trans! XXX */ ++ if (ts->span.chans[i]->rxsig != rxs) { ++ t4_dahdi_rbsbits(ts->span.chans[i], rxs); ++ } ++ } ++ } ++ } ++} ++ ++/* Must be called from within hardirq context. */ ++static void t4_check_alarms(struct t4 *wc, int span) ++{ ++ unsigned char c, d, e; ++ int alarms; ++ int x,j; ++ struct t4_span *ts = wc->tspans[span]; ++ ++ if (time_before(jiffies, ts->alarmcheck_time)) ++ return; ++ ++ if (!(ts->span.flags & DAHDI_FLAG_RUNNING)) ++ return; ++ ++ spin_lock(&wc->reglock); ++ ++ c = __t4_framer_in(wc, span, 0x4c); ++ d = __t4_framer_in(wc, span, 0x4d); ++ ++ /* Assume no alarms */ ++ alarms = 0; ++ ++ /* And consider only carrier alarms */ ++ ts->span.alarms &= (DAHDI_ALARM_RED | DAHDI_ALARM_BLUE | DAHDI_ALARM_NOTOPEN); ++ ++ if (E1 == ts->linemode) { ++ if (c & 0x04) { ++ /* No multiframe found, force RAI high after 400ms only if ++ we haven't found a multiframe since last loss ++ of frame */ ++ if (!(ts->spanflags & FLAG_NMF)) { ++ __t4_framer_out(wc, span, 0x20, 0x9f | 0x20); /* LIM0: Force RAI High */ ++ ts->spanflags |= FLAG_NMF; ++ dev_notice(&wc->dev->dev, ++ "Lost crc4-multiframe alignment\n"); ++ } ++ __t4_framer_out(wc, span, 0x1e, 0xc3); /* Reset to CRC4 mode */ ++ __t4_framer_out(wc, span, 0x1c, 0xf2); /* Force Resync */ ++ __t4_framer_out(wc, span, 0x1c, 0xf0); /* Force Resync */ ++ } else if (!(c & 0x02)) { ++ if ((ts->spanflags & FLAG_NMF)) { ++ __t4_framer_out(wc, span, 0x20, 0x9f); /* LIM0: Clear forced RAI */ ++ ts->spanflags &= ~FLAG_NMF; ++ dev_notice(&wc->dev->dev, ++ "Obtained crc4-multiframe alignment\n"); ++ } ++ } ++ } else { ++ /* Detect loopup code if we're not sending one */ ++ if ((!ts->span.mainttimer) && (d & 0x08)) { ++ /* Loop-up code detected */ ++ if ((ts->loopupcnt++ > 80) && (ts->span.maintstat != DAHDI_MAINT_REMOTELOOP)) { ++ dev_notice(&wc->dev->dev, ++ "span %d: Loopup detected,"\ ++ " enabling remote loop\n", ++ span+1); ++ __t4_framer_out(wc, span, 0x36, 0x08); /* LIM0: Disable any local loop */ ++ __t4_framer_out(wc, span, 0x37, 0xf6 ); /* LIM1: Enable remote loop */ ++ ts->span.maintstat = DAHDI_MAINT_REMOTELOOP; ++ } ++ } else ++ ts->loopupcnt = 0; ++ /* Same for loopdown code */ ++ if ((!ts->span.mainttimer) && (d & 0x10)) { ++ /* Loop-down code detected */ ++ if ((ts->loopdowncnt++ > 80) && (ts->span.maintstat == DAHDI_MAINT_REMOTELOOP)) { ++ dev_notice(&wc->dev->dev, ++ "span %d: Loopdown detected,"\ ++ " disabling remote loop\n", ++ span+1); ++ __t4_framer_out(wc, span, 0x36, 0x08); /* LIM0: Disable any local loop */ ++ __t4_framer_out(wc, span, 0x37, 0xf0 ); /* LIM1: Disable remote loop */ ++ ts->span.maintstat = DAHDI_MAINT_NONE; ++ } ++ } else ++ ts->loopdowncnt = 0; ++ } ++ ++ if (ts->span.lineconfig & DAHDI_CONFIG_NOTOPEN) { ++ for (x=0,j=0;x < ts->span.channels;x++) ++ if ((ts->span.chans[x]->flags & DAHDI_FLAG_OPEN) || ++ dahdi_have_netdev(ts->span.chans[x])) ++ j++; ++ if (!j) ++ alarms |= DAHDI_ALARM_NOTOPEN; ++ } ++ ++ /* Loss of Frame Alignment */ ++ if (c & 0x20) { ++ if (!ts->alarm_time) { ++ if (unlikely(debug)) { ++ /* starting to debounce LOF/LFA */ ++ dev_info(&wc->dev->dev, "allo%dxxp: LOF/LFA " ++ "detected on span %d but debouncing " ++ "for %d ms\n", wc->numspans, span + 1, ++ alarmdebounce); ++ } ++ ts->alarm_time = jiffies + ++ msecs_to_jiffies(alarmdebounce); ++ } else if (time_after(jiffies, ts->alarm_time)) { ++ /* Disable Slip Interrupts */ ++ e = __t4_framer_in(wc, span, 0x17); ++ __t4_framer_out(wc, span, 0x17, (e|0x03)); ++ ++ alarms |= DAHDI_ALARM_RED; ++ } ++ } else { ++ ts->alarm_time = 0; ++ } ++ ++ /* Loss of Signal */ ++ if (c & 0x80) { ++ if (!ts->losalarm_time) { ++ if (unlikely(debug)) { ++ /* starting to debounce LOS */ ++ dev_info(&wc->dev->dev, "allo%dxxp: LOS " ++ "detected on span %d but debouncing " ++ "for %d ms\n", wc->numspans, ++ span + 1, losalarmdebounce); ++ } ++ ts->losalarm_time = jiffies + ++ msecs_to_jiffies(losalarmdebounce); ++ } else if (time_after(jiffies, ts->losalarm_time)) { ++ /* Disable Slip Interrupts */ ++ e = __t4_framer_in(wc, span, 0x17); ++ __t4_framer_out(wc, span, 0x17, (e|0x03)); ++ ++ alarms |= DAHDI_ALARM_RED; ++ } ++ } else { ++ ts->losalarm_time = 0; ++ } ++ ++ /* Alarm Indication Signal */ ++ if (c & 0x40) { ++ if (!ts->aisalarm_time) { ++ if (unlikely(debug)) { ++ /* starting to debounce AIS */ ++ dev_info(&wc->dev->dev, "allo%dxxp: AIS " ++ "detected on span %d but debouncing " ++ "for %d ms\n", wc->numspans, ++ span + 1, aisalarmdebounce); ++ } ++ ts->aisalarm_time = jiffies + ++ msecs_to_jiffies(aisalarmdebounce); ++ } else if (time_after(jiffies, ts->aisalarm_time)) { ++ alarms |= DAHDI_ALARM_BLUE; ++ } ++ } else { ++ ts->aisalarm_time = 0; ++ } ++ ++ /* Add detailed alarm status information to a red alarm state */ ++ if (alarms & DAHDI_ALARM_RED) { ++ if (c & FRS0_LOS) ++ alarms |= DAHDI_ALARM_LOS; ++ if (c & FRS0_LFA) ++ alarms |= DAHDI_ALARM_LFA; ++ if (c & FRS0_LMFA) ++ alarms |= DAHDI_ALARM_LMFA; ++ } ++ ++ if (unlikely(debug)) { ++ /* Check to ensure the xmit line isn't shorted */ ++ if (unlikely(d & FRS1_XLS)) { ++ dev_info(&wc->dev->dev, ++ "Detected a possible hardware malfunction"\ ++ " this card may need servicing\n"); ++ } ++ } ++ ++ if (((!ts->span.alarms) && alarms) || ++ (ts->span.alarms && (!alarms))) ++ set_bit(T4_CHECK_TIMING, &wc->checkflag); ++ ++ /* Keep track of recovering */ ++ if ((!alarms) && ts->span.alarms) ++ ts->alarmtimer = DAHDI_ALARMSETTLE_TIME; ++ if (ts->alarmtimer) ++ alarms |= DAHDI_ALARM_RECOVER; ++ ++ /* If receiving alarms, go into Yellow alarm state */ ++ if (alarms && !(ts->spanflags & FLAG_SENDINGYELLOW)) { ++ /* We manually do yellow alarm to handle RECOVER and NOTOPEN, otherwise it's auto anyway */ ++ unsigned char fmr4; ++ fmr4 = __t4_framer_in(wc, span, 0x20); ++ __t4_framer_out(wc, span, 0x20, fmr4 | 0x20); ++ dev_info(&wc->dev->dev, "Setting yellow alarm span %d\n", ++ span+1); ++ ts->spanflags |= FLAG_SENDINGYELLOW; ++ } else if ((!alarms) && (ts->spanflags & FLAG_SENDINGYELLOW)) { ++ unsigned char fmr4; ++ /* We manually do yellow alarm to handle RECOVER */ ++ fmr4 = __t4_framer_in(wc, span, 0x20); ++ __t4_framer_out(wc, span, 0x20, fmr4 & ~0x20); ++ dev_info(&wc->dev->dev, "Clearing yellow alarm span %d\n", ++ span+1); ++ ++ /* Re-enable timing slip interrupts */ ++ e = __t4_framer_in(wc, span, 0x17); ++ ++ __t4_framer_out(wc, span, 0x17, (e & ~(0x03))); ++ ++ ts->spanflags &= ~FLAG_SENDINGYELLOW; ++ } ++ ++ /* Re-check the timing source when we enter/leave alarm, not withstanding ++ yellow alarm */ ++ if (c & 0x10) { /* receiving yellow (RAI) */ ++ if (!ts->yelalarm_time) { ++ if (unlikely(debug)) { ++ /* starting to debounce AIS */ ++ dev_info(&wc->dev->dev, "allo%dxxp: yellow " ++ "(RAI) detected on span %d but " ++ "debouncing for %d ms\n", ++ wc->numspans, span + 1, ++ yelalarmdebounce); ++ } ++ ts->yelalarm_time = jiffies + ++ msecs_to_jiffies(yelalarmdebounce); ++ } else if (time_after(jiffies, ts->yelalarm_time)) { ++ alarms |= DAHDI_ALARM_YELLOW; ++ } ++ } else { ++ ts->yelalarm_time = 0; ++ } ++ ++ if (alarms) ++ ts->alarmcheck_time = jiffies + msecs_to_jiffies(100); ++ else ++ ts->alarmcheck_time = jiffies + msecs_to_jiffies(50); ++ ++ if (ts->span.mainttimer || ts->span.maintstat) ++ alarms |= DAHDI_ALARM_LOOPBACK; ++ ts->span.alarms = alarms; ++ ++ spin_unlock(&wc->reglock); ++ dahdi_alarm_notify(&ts->span); ++} ++ ++static void t4_do_counters(struct t4 *wc) ++{ ++ int span; ++ for (span = 0; span < wc->numspans; span++) { ++ struct t4_span *ts = wc->tspans[span]; ++ ++ spin_lock(&wc->reglock); ++ if (ts->alarmtimer && (0 == (--ts->alarmtimer))) ++ ts->span.alarms &= ~(DAHDI_ALARM_RECOVER); ++ spin_unlock(&wc->reglock); ++ ++ t4_check_alarms(wc, span); ++ } ++} ++ ++static inline void __handle_leds(struct t4 *wc) ++{ ++ int x; ++ ++ wc->blinktimer++; ++ for (x=0;x<wc->numspans;x++) { ++ struct t4_span *ts = wc->tspans[x]; ++ if (ts->span.flags & DAHDI_FLAG_RUNNING) { ++ if ((ts->span.alarms & (DAHDI_ALARM_RED | ++ DAHDI_ALARM_BLUE)) || ++ ts->losalarm_time) { ++#ifdef FANCY_ALARM ++ if (wc->blinktimer == (altab[wc->alarmpos] >> 1)) { ++ __t4_set_led(wc, x, WC_RED); ++ } ++ if (wc->blinktimer == 0xf) { ++ __t4_set_led(wc, x, WC_OFF); ++ } ++#else ++ if (wc->blinktimer == 160) { ++ __t4_set_led(wc, x, WC_RED); ++ } else if (wc->blinktimer == 480) { ++ __t4_set_led(wc, x, WC_OFF); ++ } ++#endif ++ } else if (ts->span.alarms & DAHDI_ALARM_YELLOW) { ++ /* Yellow Alarm */ ++ __t4_set_led(wc, x, WC_YELLOW); ++ } else if (ts->span.mainttimer || ts->span.maintstat) { ++#ifdef FANCY_ALARM ++ if (wc->blinktimer == (altab[wc->alarmpos] >> 1)) { ++ __t4_set_led(wc, x, WC_GREEN); ++ } ++ if (wc->blinktimer == 0xf) { ++ __t4_set_led(wc, x, WC_OFF); ++ } ++#else ++ if (wc->blinktimer == 160) { ++ __t4_set_led(wc, x, WC_GREEN); ++ } else if (wc->blinktimer == 480) { ++ __t4_set_led(wc, x, WC_OFF); ++ } ++#endif ++ } else { ++ /* No Alarm */ ++ __t4_set_led(wc, x, WC_GREEN); ++ } ++ } else ++ __t4_set_led(wc, x, WC_OFF); ++ ++ } ++#ifdef FANCY_ALARM ++ if (wc->blinktimer == 0xf) { ++ wc->blinktimer = -1; ++ wc->alarmpos++; ++ if (wc->alarmpos >= ARRAY_SIZE(altab)) ++ wc->alarmpos = 0; ++ } ++#else ++ if (wc->blinktimer == 480) ++ wc->blinktimer = 0; ++#endif ++} ++ ++static inline void t4_framer_interrupt(struct t4 *wc, int span) ++{ ++ /* Check interrupts for a given span */ ++ unsigned char gis, isr0, isr1, isr2, isr3, isr4; ++ int readsize = -1; ++ struct t4_span *ts = wc->tspans[span]; ++ struct dahdi_chan *sigchan; ++ unsigned long flags; ++ ++ ++ /* 1st gen cards isn't used interrupts */ ++ spin_lock_irqsave(&wc->reglock, flags); ++ gis = __t4_framer_in(wc, span, FRMR_GIS); ++ isr0 = (gis & FRMR_GIS_ISR0) ? __t4_framer_in(wc, span, FRMR_ISR0) : 0; ++ isr1 = (gis & FRMR_GIS_ISR1) ? __t4_framer_in(wc, span, FRMR_ISR1) : 0; ++ isr2 = (gis & FRMR_GIS_ISR2) ? __t4_framer_in(wc, span, FRMR_ISR2) : 0; ++ isr3 = (gis & FRMR_GIS_ISR3) ? __t4_framer_in(wc, span, FRMR_ISR3) : 0; ++ isr4 = (gis & FRMR_GIS_ISR4) ? __t4_framer_in(wc, span, FRMR_ISR4) : 0; ++ ++ if ((debug & DEBUG_FRAMER) && !(isr3 & ISR3_SEC)) { ++ dev_info(&wc->dev->dev, "gis: %02x, isr0: %02x, isr1: %02x, "\ ++ "isr2: %02x, isr3: %08x, isr4: %02x, intcount=%u\n", ++ gis, isr0, isr1, isr2, isr3, isr4, wc->intcount); ++ } ++ ++ /* Collect performance counters once per second */ ++ if (isr3 & ISR3_SEC) { ++ ts->span.count.fe += __t4_framer_in(wc, span, FECL_T); ++ ts->span.count.crc4 += __t4_framer_in(wc, span, CEC1L_T); ++ ts->span.count.cv += __t4_framer_in(wc, span, CVCL_T); ++ ts->span.count.ebit += __t4_framer_in(wc, span, EBCL_T); ++ ts->span.count.be += __t4_framer_in(wc, span, BECL_T); ++ ts->span.count.prbs = __t4_framer_in(wc, span, FRS1_T); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ /* Collect errored second counter once per second */ ++ if (isr3 & ISR3_ES) { ++ ts->span.count.errsec += 1; ++ } ++ ++ if (isr0) ++ t4_check_sigbits(wc, span); ++ ++ if (E1 == ts->linemode) { ++ /* E1 checks */ ++ if ((isr3 & 0x38) || isr2 || isr1) ++ t4_check_alarms(wc, span); ++ } else { ++ /* T1 checks */ ++ if (isr2 || (isr3 & 0x08)) ++ t4_check_alarms(wc, span); ++ } ++ if (!ts->span.alarms) { ++ if ((isr3 & 0x3) || (isr4 & 0xc0)) ++ ts->span.count.timingslips++; ++ ++ if (debug & DEBUG_MAIN) { ++ if (isr3 & 0x02) ++ dev_notice(&wc->dev->dev, "ALLO%d10P: RECEIVE " ++ "slip NEGATIVE on span %d\n", ++ wc->numspans, span + 1); ++ if (isr3 & 0x01) ++ dev_notice(&wc->dev->dev, "ALLO%d10P: RECEIVE " ++ "slip POSITIVE on span %d\n", ++ wc->numspans, span + 1); ++ if (isr4 & 0x80) ++ dev_notice(&wc->dev->dev, "ALLO%dXXP: TRANSMIT " ++ "slip POSITIVE on span %d\n", ++ wc->numspans, span + 1); ++ if (isr4 & 0x40) ++ dev_notice(&wc->dev->dev, "ALLO%d10P: TRANSMIT " ++ "slip NEGATIVE on span %d\n", ++ wc->numspans, span + 1); ++ } ++ } else ++ ts->span.count.timingslips = 0; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ /* HDLC controller checks - receive side */ ++ if (!ts->sigchan) { ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return; ++ } ++ ++ sigchan = ts->sigchan; ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ if (isr0 & FRMR_ISR0_RME) { ++ readsize = (t4_framer_in(wc, span, FRMR_RBCH) << 8) | t4_framer_in(wc, span, FRMR_RBCL); ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Received data length is %d " ++ "(%d)\n", readsize, ++ readsize & FRMR_RBCL_MAX_SIZE); ++ /* RPF isn't set on last part of frame */ ++ if ((readsize > 0) && ((readsize &= FRMR_RBCL_MAX_SIZE) == 0)) ++ readsize = FRMR_RBCL_MAX_SIZE + 1; ++ } else if (isr0 & FRMR_ISR0_RPF) ++ readsize = FRMR_RBCL_MAX_SIZE + 1; ++ ++ if (readsize > 0) { ++ int i; ++ unsigned char readbuf[FRMR_RBCL_MAX_SIZE + 1]; ++ ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Framer %d: Got RPF/RME! " ++ "readsize is %d\n", sigchan->span->offset, ++ readsize); ++ ++ for (i = 0; i < readsize; i++) ++ readbuf[i] = t4_framer_in(wc, span, FRMR_RXFIFO); ++ ++ /* Tell the framer to clear the RFIFO */ ++ t4_framer_cmd_wait(wc, span, FRMR_CMDR_RMC); ++ ++ if (debug & DEBUG_FRAMER) { ++ dev_notice(&wc->dev->dev, "RX("); ++ for (i = 0; i < readsize; i++) ++ dev_notice(&wc->dev->dev, "%s%02x", ++ (i ? " " : ""), readbuf[i]); ++ dev_notice(&wc->dev->dev, ")\n"); ++ } ++ ++ if (isr0 & FRMR_ISR0_RME) { ++ /* Do checks for HDLC problems */ ++ unsigned char rsis = readbuf[readsize-1]; ++ unsigned char rsis_reg = t4_framer_in(wc, span, FRMR_RSIS); ++ ++ ++ts->frames_in; ++ if ((debug & DEBUG_FRAMER) && !(ts->frames_in & 0x0f)) ++ dev_notice(&wc->dev->dev, "Received %d frames " ++ "on span %d\n", ts->frames_in, span); ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Received HDLC frame" ++ " %d. RSIS = 0x%x (%x)\n", ++ ts->frames_in, rsis, rsis_reg); ++ if (!(rsis & FRMR_RSIS_CRC16)) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "CRC check " ++ "failed %d\n", span); ++ //dahdi_hdlc_abort(sigchan, DAHDI_EVENT_BADFCS); ++ } else if (rsis & FRMR_RSIS_RAB) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "ABORT of " ++ "current frame due to " ++ "overflow %d\n", span); ++ //dahdi_hdlc_abort(sigchan, DAHDI_EVENT_ABORT); ++ } else if (rsis & FRMR_RSIS_RDO) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "HDLC " ++ "overflow occured %d\n", ++ span); ++ dahdi_hdlc_abort(sigchan, DAHDI_EVENT_OVERRUN); ++ } else if (!(rsis & FRMR_RSIS_VFR)) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Valid Frame" ++ " check failed on span %d\n", ++ span); ++ //dahdi_hdlc_abort(sigchan, DAHDI_EVENT_ABORT); ++ } else { ++ dahdi_hdlc_putbuf(sigchan, readbuf, readsize - 1); ++ dahdi_hdlc_finish(sigchan); ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Received " ++ "valid HDLC frame on span %d" ++ "\n", span); ++ } ++ } else if (isr0 & FRMR_ISR0_RPF) ++ dahdi_hdlc_putbuf(sigchan, readbuf, readsize); ++ } ++ ++ /* Transmit side */ ++ if (isr1 & FRMR_ISR1_XDU) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "XDU: Resetting signal " ++ "controller!\n"); ++ t4_framer_cmd_wait(wc, span, FRMR_CMDR_SRES); ++ } else if (isr1 & FRMR_ISR1_XPR) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Sigchan %d is %p\n", ++ sigchan->chanpos, sigchan); ++ ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "Framer %d: Got XPR!\n", ++ sigchan->span->offset); ++ t4_hdlc_xmit_fifo(wc, span, ts); ++ } ++ ++ if (isr1 & FRMR_ISR1_ALLS) { ++ if (debug & DEBUG_FRAMER) ++ dev_notice(&wc->dev->dev, "ALLS received\n"); ++ } ++} ++ ++#ifdef SUPPORT_GEN1 ++static irqreturn_t _t4_interrupt(int irq, void *dev_id) ++{ ++ struct t4 *wc = dev_id; ++ unsigned long flags; ++ int x; ++ ++ unsigned int status; ++ unsigned int status2; ++ ++ /* Make sure it's really for us */ ++ status = __t4_pci_in(wc, WC_INTR); ++ ++ /* Process framer interrupts */ ++ status2 = t4_framer_in(wc, 0, FRMR_CIS); ++ if (status2 & 0x0f) { ++ for (x = 0; x < wc->numspans; ++x) { ++ if (status2 & (1 << x)) ++ t4_framer_interrupt(wc, x); ++ } ++ } ++ ++ /* Ignore if it's not for us */ ++ if (!status) ++ return IRQ_NONE; ++ ++ __t4_pci_out(wc, WC_INTR, 0); ++ ++ if (!wc->spansstarted) { ++ dev_notice(&wc->dev->dev, "Not prepped yet!\n"); ++ return IRQ_NONE; ++ } ++ ++ wc->intcount++; ++ ++ if (status & 0x3) { ++ t4_receiveprep(wc, status); ++ t4_transmitprep(wc, status); ++ } ++ ++ t4_do_counters(wc); ++ ++ x = wc->intcount & 15 /* 63 */; ++ switch(x) { ++ case 0: ++ case 1: ++ case 2: ++ case 3: ++ t4_check_sigbits(wc, x); ++ break; ++ case 4: ++ case 5: ++ case 6: ++ case 7: ++ t4_check_alarms(wc, x - 4); ++ break; ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ __handle_leds(wc); ++ ++ if (test_bit(T4_CHECK_TIMING, &wc->checkflag)) ++ __t4_set_timing_source_auto(wc); ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ return IRQ_RETVAL(1); ++} ++ ++DAHDI_IRQ_HANDLER(t4_interrupt) ++{ ++ irqreturn_t ret; ++ unsigned long flags; ++ local_irq_save(flags); ++ ret = _t4_interrupt(irq, dev_id); ++ local_irq_restore(flags); ++ return ret; ++} ++#endif ++ ++static int t4_allocate_buffers(struct t4 *wc, int numbufs, ++ void **oldalloc, dma_addr_t *oldwritedma) ++{ ++ void *alloc; ++ dma_addr_t writedma; ++ ++ /* 32 channels, Double-buffer, Read/Write, 4 spans */ ++ alloc = pci_alloc_consistent(wc->dev, numbufs * T4_BASE_SIZE(wc) * 2, ++ &writedma); ++ ++ if (!alloc) { ++ dev_notice(&wc->dev->dev, "allo%dxxp: Unable to allocate " ++ "DMA-able memory\n", wc->numspans); ++ return -ENOMEM; ++ } ++ ++ if (oldwritedma) ++ *oldwritedma = wc->writedma; ++ if (oldalloc) ++ *oldalloc = wc->writechunk; ++ ++ wc->writechunk = alloc; ++ wc->writedma = writedma; ++ ++ /* Read is after the whole write piece (in words) */ ++ wc->readchunk = wc->writechunk + (T4_BASE_SIZE(wc) * numbufs) / 4; ++ ++ /* Same thing but in bytes... */ ++ wc->readdma = wc->writedma + (T4_BASE_SIZE(wc) * numbufs); ++ ++ wc->numbufs = numbufs; ++ ++ /* Initialize Write/Buffers to all blank data */ ++ memset(wc->writechunk, 0x00, T4_BASE_SIZE(wc) * numbufs); ++ memset(wc->readchunk, 0xff, T4_BASE_SIZE(wc) * numbufs); ++ ++ if (debug) { ++ dev_notice(&wc->dev->dev, "DMA memory base of size %d at " \ ++ "%p. Read: %p and Write %p\n", ++ numbufs * T4_BASE_SIZE(wc) * 2, wc->writechunk, ++ wc->readchunk, wc->writechunk); ++ } ++ ++ return 0; ++} ++ ++static void t4_increase_latency(struct t4 *wc, int newlatency) ++{ ++ unsigned long flags; ++ void *oldalloc; ++ dma_addr_t oldaddr; ++ int oldbufs; ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ ++ __t4_pci_out(wc, WC_DMACTRL, 0x00000000); ++ /* Acknowledge any pending interrupts */ ++ __t4_pci_out(wc, WC_INTR, 0x00000000); ++ ++ __t4_pci_in(wc, WC_VERSION); ++ ++ oldbufs = wc->numbufs; ++ ++ if (t4_allocate_buffers(wc, newlatency, &oldalloc, &oldaddr)) { ++ dev_info(&wc->dev->dev, "Error allocating latency buffers for " ++ "latency of %d\n", newlatency); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ return; ++ } ++ ++ __t4_pci_out(wc, WC_RDADDR, wc->readdma); ++ __t4_pci_out(wc, WC_WRADDR, wc->writedma); ++ ++ __t4_pci_in(wc, WC_VERSION); ++ ++ __t4_pci_out(wc, 5, (ms_per_irq << 16) | newlatency); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ ++ __t4_pci_in(wc, WC_VERSION); ++ ++ wc->rxident = 0; ++ wc->lastindex = 0; ++ ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ pci_free_consistent(wc->dev, T4_BASE_SIZE(wc) * oldbufs * 2, ++ oldalloc, oldaddr); ++ ++ dev_info(&wc->dev->dev, "Increased latency to %d\n", newlatency); ++ ++} ++ ++static void t4_isr_bh(unsigned long data) ++{ ++ struct t4 *wc = (struct t4 *)data; ++ ++ if (test_bit(T4_CHANGE_LATENCY, &wc->checkflag)) { ++ if (wc->needed_latency != wc->numbufs) { ++ t4_increase_latency(wc, wc->needed_latency); ++ clear_bit(T4_CHANGE_LATENCY, &wc->checkflag); ++ } ++ } ++#ifdef VPM_SUPPORT ++ if (wc->vpm) { ++ if (test_and_clear_bit(T4_CHECK_VPM, &wc->checkflag)) { ++ /* How stupid is it that the octasic can't generate an ++ * interrupt when there's a tone, in spite of what ++ * their documentation says? */ ++ t4_check_vpm(wc); ++ } ++ } ++#endif ++} ++ ++static irqreturn_t _t4_interrupt_gen2(int irq, void *dev_id) ++{ ++ struct t4 *wc = dev_id; ++ unsigned int status; ++ unsigned char rxident, expected; ++ ++ /* Check this first in case we get a spurious interrupt */ ++ if (unlikely(test_bit(T4_STOP_DMA, &wc->checkflag))) { ++ /* Stop DMA cleanly if requested */ ++ wc->dmactrl = 0x0; ++ t4_pci_out(wc, WC_DMACTRL, 0x00000000); ++ /* Acknowledge any pending interrupts */ ++ t4_pci_out(wc, WC_INTR, 0x00000000); ++ spin_lock(&wc->reglock); ++ __t4_set_sclk_src(wc, WC_SELF, 0, 0); ++ spin_unlock(&wc->reglock); ++ return IRQ_RETVAL(1); ++ } ++ ++ /* Make sure it's really for us */ ++ status = __t4_pci_in(wc, WC_INTR); ++ ++ /* Ignore if it's not for us */ ++ if (!(status & 0x7)) { ++ return IRQ_NONE; ++ } ++ ++#ifdef ENABLE_WORKQUEUES ++ __t4_pci_out(wc, WC_INTR, status & 0x00000008); ++#endif ++ ++ if (unlikely(!wc->spansstarted)) { ++ dev_info(&wc->dev->dev, "Not prepped yet!\n"); ++ return IRQ_NONE; ++ } ++ ++ wc->intcount++; ++ if ((wc->devtype->flags & FLAG_5THGEN) && (status & 0x2)) { ++ rxident = (status >> 16) & 0x7f; ++ expected = (wc->rxident + ms_per_irq) % 128; ++ ++ if ((rxident != expected) && !test_bit(T4_IGNORE_LATENCY, &wc->checkflag)) { ++ int needed_latency; ++ int smallest_max; ++ ++ if (debug & DEBUG_MAIN) ++ dev_warn(&wc->dev->dev, "Missed interrupt. " ++ "Expected ident of %d and got ident " ++ "of %d\n", expected, rxident); ++ ++ if (test_bit(T4_IGNORE_LATENCY, &wc->checkflag)) { ++ dev_info(&wc->dev->dev, ++ "Should have ignored latency\n"); ++ } ++ if (rxident > wc->rxident) { ++ needed_latency = rxident - wc->rxident; ++ } else { ++ needed_latency = (128 - wc->rxident) + rxident; ++ } ++ ++ needed_latency += 1; ++ ++ smallest_max = (max_latency >= GEN5_MAX_LATENCY) ? GEN5_MAX_LATENCY : max_latency; ++ ++ if (needed_latency > smallest_max) { ++ dev_info(&wc->dev->dev, "Truncating latency " ++ "request to %d instead of %d\n", ++ smallest_max, needed_latency); ++ needed_latency = smallest_max; ++ } ++ ++ if (needed_latency > wc->numbufs) { ++ dev_info(&wc->dev->dev, "Need to increase " ++ "latency. Estimated latency should " ++ "be %d\n", needed_latency); ++ wc->ddev->irqmisses++; ++ wc->needed_latency = needed_latency; ++ __t4_pci_out(wc, WC_DMACTRL, 0x00000000); ++ set_bit(T4_CHANGE_LATENCY, &wc->checkflag); ++ goto out; ++ } ++ } ++ ++ wc->rxident = rxident; ++ } ++ ++#ifdef DEBUG ++ if (unlikely((wc->intcount < 20))) ++ dev_dbg(&wc->dev->dev, "2G: Got interrupt, status = %08x, " ++ "CIS = %04x\n", status, t4_framer_in(wc, 0, FRMR_CIS)); ++#endif ++ ++ if (likely(status & 0x2)) { ++#ifdef ENABLE_WORKQUEUES ++ int cpus = num_online_cpus(); ++ atomic_set(&wc->worklist, wc->numspans); ++ if (wc->tspans[0]->span.flags & DAHDI_FLAG_RUNNING) ++ t4_queue_work(wc->workq, &wc->tspans[0]->swork, 0); ++ else ++ atomic_dec(&wc->worklist); ++ if (wc->numspans >= 2) { //cem:1port change ++ if (wc->tspans[1]->span.flags & DAHDI_FLAG_RUNNING) ++ t4_queue_work(wc->workq, &wc->tspans[1]->swork, 1 % cpus); ++ else ++ atomic_dec(&wc->worklist); ++ } ++ if (wc->numspans == 4) { ++ if (wc->tspans[2]->span.flags & DAHDI_FLAG_RUNNING) ++ t4_queue_work(wc->workq, &wc->tspans[2]->swork, 2 % cpus); ++ else ++ atomic_dec(&wc->worklist); ++ if (wc->tspans[3]->span.flags & DAHDI_FLAG_RUNNING) ++ t4_queue_work(wc->workq, &wc->tspans[3]->swork, 3 % cpus); ++ else ++ atomic_dec(&wc->worklist); ++ } ++#else ++ unsigned int reg5 = __t4_pci_in(wc, 5); ++ ++#ifdef DEBUG ++ if (wc->intcount < 20) ++ dev_info(&wc->dev->dev, "Reg 5 is %08x\n", reg5); ++#endif ++ ++ if (wc->devtype->flags & FLAG_5THGEN) { ++ unsigned int current_index = (reg5 >> 8) & 0x7f; ++ ++ while (((wc->lastindex + 1) % wc->numbufs) != current_index) { ++ wc->lastindex = (wc->lastindex + 1) % wc->numbufs; ++ setup_chunks(wc, wc->lastindex); ++ t4_prep_gen2(wc); ++ } ++ } else { ++ t4_prep_gen2(wc); ++ } ++ ++#endif ++ t4_do_counters(wc); ++ spin_lock(&wc->reglock); ++ __handle_leds(wc); ++ spin_unlock(&wc->reglock); ++ ++ } ++ ++ if (unlikely(status & 0x1)) { ++ unsigned char cis; ++ ++ cis = t4_framer_in(wc, 0, FRMR_CIS); ++ if (cis & FRMR_CIS_GIS1) ++ t4_framer_interrupt(wc, 0); ++ if (cis & FRMR_CIS_GIS2) ++ t4_framer_interrupt(wc, 1); ++ if (cis & FRMR_CIS_GIS3) ++ t4_framer_interrupt(wc, 2); ++ if (cis & FRMR_CIS_GIS4) ++ t4_framer_interrupt(wc, 3); ++ ++ if (is_octal(wc)) { ++ if (cis & FRMR_CIS_GIS5) ++ t4_framer_interrupt(wc, 4); ++ if (cis & FRMR_CIS_GIS6) ++ t4_framer_interrupt(wc, 5); ++ if (cis & FRMR_CIS_GIS7) ++ t4_framer_interrupt(wc, 6); ++ if (cis & FRMR_CIS_GIS8) ++ t4_framer_interrupt(wc, 7); ++ } ++ } ++ ++#ifdef VPM_SUPPORT ++ if (wc->vpm && vpmdtmfsupport) { ++ /* How stupid is it that the octasic can't generate an ++ * interrupt when there's a tone, in spite of what their ++ * documentation says? */ ++ if (!(wc->intcount & 0xf)) ++ set_bit(T4_CHECK_VPM, &wc->checkflag); ++ } ++#endif ++ ++ spin_lock(&wc->reglock); ++ ++ if (unlikely(test_bit(T4_CHECK_TIMING, &wc->checkflag))) { ++ __t4_set_timing_source_auto(wc); ++ } ++ ++ spin_unlock(&wc->reglock); ++ ++out: ++ if (unlikely(test_bit(T4_CHANGE_LATENCY, &wc->checkflag) || test_bit(T4_CHECK_VPM, &wc->checkflag))) ++ tasklet_schedule(&wc->t4_tlet); ++ ++#ifndef ENABLE_WORKQUEUES ++ __t4_pci_out(wc, WC_INTR, 0); ++#endif ++ ++ return IRQ_RETVAL(1); ++} ++ ++DAHDI_IRQ_HANDLER(t4_interrupt_gen2) ++{ ++ irqreturn_t ret; ++ unsigned long flags; ++ local_irq_save(flags); ++ ret = _t4_interrupt_gen2(irq, dev_id); ++ local_irq_restore(flags); ++ return ret; ++} ++ ++#ifdef SUPPORT_GEN1 ++static int t4_reset_dma(struct t4 *wc) ++{ ++ /* Turn off DMA and such */ ++ wc->dmactrl = 0x0; ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ t4_pci_out(wc, WC_COUNT, 0); ++ t4_pci_out(wc, WC_RDADDR, 0); ++ t4_pci_out(wc, WC_WRADDR, 0); ++ t4_pci_out(wc, WC_INTR, 0); ++ /* Turn it all back on */ ++ t4_pci_out(wc, WC_RDADDR, wc->readdma); ++ t4_pci_out(wc, WC_WRADDR, wc->writedma); ++ t4_pci_out(wc, WC_COUNT, ((DAHDI_MAX_CHUNKSIZE * 2 * 32 - 1) << 18) | ((DAHDI_MAX_CHUNKSIZE * 2 * 32 - 1) << 2)); ++ t4_pci_out(wc, WC_INTR, 0); ++#ifdef VPM_SUPPORT ++ wc->dmactrl = 0xc0000000 | (1 << 29) | ++ ((wc->vpm) ? T4_VPM_PRESENT : 0); ++#else ++ wc->dmactrl = 0xc0000000 | (1 << 29); ++#endif ++ if (noburst) ++ wc->dmactrl |= (1 << 26); ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ return 0; ++} ++#endif ++ ++#ifdef VPM_SUPPORT ++static void t4_vpm_init(struct t4 *wc) ++{ ++ int laws[8] = { 0, }; ++ int x; ++ unsigned int vpm_capacity; ++ struct firmware embedded_firmware; ++ const struct firmware *firmware = &embedded_firmware; ++#if !defined(HOTPLUG_FIRMWARE) ++ extern void _binary_dahdi_fw_oct6114_032_bin_size; //cem:1port change ++ extern void _binary_dahdi_fw_oct6114_064_bin_size; ++ extern void _binary_dahdi_fw_oct6114_128_bin_size; ++ extern void _binary_dahdi_fw_oct6114_256_bin_size; ++ extern u8 _binary_dahdi_fw_oct6114_032_bin_start[]; //cem:1port change ++ extern u8 _binary_dahdi_fw_oct6114_064_bin_start[]; ++ extern u8 _binary_dahdi_fw_oct6114_128_bin_start[]; ++ extern u8 _binary_dahdi_fw_oct6114_256_bin_start[]; ++#else ++ static const char oct032_firmware[] = "dahdi-fw-oct6114-032.bin"; //cem:1port change ++ static const char oct064_firmware[] = "dahdi-fw-oct6114-064.bin"; ++ static const char oct128_firmware[] = "dahdi-fw-oct6114-128.bin"; ++ static const char oct256_firmware[] = "dahdi-fw-oct6114-256.bin"; ++#endif ++ ++ if (!vpmsupport) { ++ dev_info(&wc->dev->dev, "VPM450: Support Disabled\n"); ++ return; ++ } ++ ++ /* Turn on GPIO/DATA mux if supported */ ++ t4_gpio_setdir(wc, (1 << 24), (1 << 24)); ++ __t4_raw_oct_out(wc, 0x000a, 0x5678); ++ __t4_raw_oct_out(wc, 0x0004, 0x1234); ++ __t4_raw_oct_in(wc, 0x0004); ++ __t4_raw_oct_in(wc, 0x000a); ++ if (debug) ++ dev_notice(&wc->dev->dev, "OCT Result: %04x/%04x\n", ++ __t4_raw_oct_in(wc, 0x0004), ++ __t4_raw_oct_in(wc, 0x000a)); ++ if (__t4_raw_oct_in(wc, 0x0004) != 0x1234) { ++ dev_notice(&wc->dev->dev, "VPM450: Not Present\n"); ++ return; ++ } ++ ++ /* Setup alaw vs ulaw rules */ ++ for (x = 0;x < wc->numspans; x++) { ++ if (wc->tspans[x]->span.channels > 24) ++ laws[x] = 1; ++ } ++ ++ vpm_capacity = get_vpm450m_capacity(&wc->dev->dev); ++ if (vpm_capacity != wc->numspans * 32) { ++ dev_info(&wc->dev->dev, "Disabling VPMOCT%03d. ALLO%dXXP"\ ++ " requires a VPMOCT%03d", vpm_capacity, ++ wc->numspans, wc->numspans*32); ++ return; ++ } ++ ++ switch (vpm_capacity) { ++ case 32:/*cem:1port change, v002*/ ++#if defined(HOTPLUG_FIRMWARE) ++ if ((request_firmware(&firmware, oct032_firmware, &wc->dev->dev) != 0) || ++ !firmware) { ++ dev_notice(&wc->dev->dev, "VPM450: firmware %s not " ++ "available from userspace\n", oct032_firmware); ++ return; ++ } ++#else ++ embedded_firmware.data = _binary_dahdi_fw_oct6114_032_bin_start; ++ /* Yes... this is weird. objcopy gives us a symbol containing ++ the size of the firmware, not a pointer a variable containing ++ the size. The only way we can get the value of the symbol ++ is to take its address, so we define it as a pointer and ++ then cast that value to the proper type. ++ */ ++ embedded_firmware.size = (size_t) &_binary_dahdi_fw_oct6114_032_bin_size; ++#endif ++ break; ++ ++ case 64: ++#if defined(HOTPLUG_FIRMWARE) ++ if ((request_firmware(&firmware, oct064_firmware, &wc->dev->dev) != 0) || ++ !firmware) { ++ dev_notice(&wc->dev->dev, "VPM450: firmware %s not " ++ "available from userspace\n", oct064_firmware); ++ return; ++ } ++#else ++ embedded_firmware.data = _binary_dahdi_fw_oct6114_064_bin_start; ++ /* Yes... this is weird. objcopy gives us a symbol containing ++ the size of the firmware, not a pointer a variable containing ++ the size. The only way we can get the value of the symbol ++ is to take its address, so we define it as a pointer and ++ then cast that value to the proper type. ++ */ ++ embedded_firmware.size = (size_t) &_binary_dahdi_fw_oct6114_064_bin_size; ++#endif ++ break; ++ case 128: ++#if defined(HOTPLUG_FIRMWARE) ++ if ((request_firmware(&firmware, oct128_firmware, &wc->dev->dev) != 0) || ++ !firmware) { ++ dev_notice(&wc->dev->dev, "VPM450: firmware %s not " ++ "available from userspace\n", oct128_firmware); ++ return; ++ } ++#else ++ embedded_firmware.data = _binary_dahdi_fw_oct6114_128_bin_start; ++ /* Yes... this is weird. objcopy gives us a symbol containing ++ the size of the firmware, not a pointer a variable containing ++ the size. The only way we can get the value of the symbol ++ is to take its address, so we define it as a pointer and ++ then cast that value to the proper type. ++ */ ++ embedded_firmware.size = (size_t) &_binary_dahdi_fw_oct6114_128_bin_size; ++#endif ++ break; ++ case 256: ++#if defined(HOTPLUG_FIRMWARE) ++ if ((request_firmware(&firmware, oct256_firmware, &wc->dev->dev) != 0) || ++ !firmware) { ++ dev_notice(&wc->dev->dev, "VPM450: firmware %s not " ++ "available from userspace\n", oct256_firmware); ++ return; ++ } ++#else ++ embedded_firmware.data = _binary_dahdi_fw_oct6114_256_bin_start; ++ /* Yes... this is weird. objcopy gives us a symbol containing ++ the size of the firmware, not a pointer a variable containing ++ the size. The only way we can get the value of the symbol ++ is to take its address, so we define it as a pointer and ++ then cast that value to the proper type. ++ */ ++ embedded_firmware.size = (size_t) &_binary_dahdi_fw_oct6114_256_bin_size; ++#endif ++ break; ++ default: ++ dev_notice(&wc->dev->dev, "Unsupported channel capacity found " ++ "on VPM module (%d).\n", vpm_capacity); ++ return; ++ } ++ ++ wc->vpm = init_vpm450m(&wc->dev->dev, laws, wc->numspans, firmware); ++ if (!wc->vpm) { ++ dev_notice(&wc->dev->dev, "VPM450: Failed to initialize\n"); ++ if (firmware != &embedded_firmware) ++ release_firmware(firmware); ++ return; ++ } ++ ++ if (firmware != &embedded_firmware) ++ release_firmware(firmware); ++ ++ if (vpmdtmfsupport == -1) { ++ dev_info(&wc->dev->dev, "VPM450: hardware DTMF disabled.\n"); ++ vpmdtmfsupport = 0; ++ } ++ ++ dev_info(&wc->dev->dev, "VPM450: Present and operational servicing %d " ++ "span(s)\n", wc->numspans); ++ ++} ++#endif /* VPM_SUPPORT */ ++ ++static void t4_tsi_reset(struct t4 *wc) ++{ ++ int x; ++ if (is_octal(wc)) { ++ for (x = 0; x < 256; x++) { ++ wc->dmactrl &= ~0x0001ffff; ++ wc->dmactrl |= (0x00004000 | ((x & 0x7f) << 7) | ((x >> 7) << 15)); ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } ++ wc->dmactrl &= ~0x0001ffff; ++ } else { ++ for (x = 0; x < 128; x++) { ++ wc->dmactrl &= ~0x00007fff; ++ wc->dmactrl |= (0x00004000 | (x << 7)); ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } ++ wc->dmactrl &= ~0x00007fff; ++ } ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++} ++ ++/* Note that channels here start from 1 */ ++static void t4_tsi_assign(struct t4 *wc, int fromspan, int fromchan, int tospan, int tochan) ++{ ++ unsigned long flags; ++ int fromts, tots; ++ ++ fromts = (fromspan << 5) |(fromchan); ++ tots = (tospan << 5) | (tochan); ++ ++ if (!has_e1_span(wc)) { ++ fromts += 4; ++ tots += 4; ++ } ++ spin_lock_irqsave(&wc->reglock, flags); ++ if (is_octal(wc)) { ++ int fromts_b = fromts & 0x7f; ++ int fromts_t = fromts >> 7; ++ int tots_b = tots & 0x7f; ++ int tots_t = tots >> 7; ++ ++ wc->dmactrl &= ~0x0001ffff; ++ wc->dmactrl |= ((fromts_t << 16) | (tots_t << 15) | 0x00004000 | (tots_b << 7) | (fromts_b)); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ wc->dmactrl &= ~0x0001ffff; ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } else { ++ wc->dmactrl &= ~0x00007fff; ++ wc->dmactrl |= (0x00004000 | (tots << 7) | (fromts)); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ wc->dmactrl &= ~0x00007fff; ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++static void t4_tsi_unassign(struct t4 *wc, int tospan, int tochan) ++{ ++ unsigned long flags; ++ int tots; ++ ++ tots = (tospan << 5) | (tochan); ++ ++ if (!has_e1_span(wc)) ++ tots += 4; ++ spin_lock_irqsave(&wc->reglock, flags); ++ if (is_octal(wc)) { ++ int tots_b = tots & 0x7f; ++ int tots_t = tots >> 7; ++ ++ wc->dmactrl &= ~0x0001ffff; ++ wc->dmactrl |= ((tots_t << 15) | 0x00004000 | (tots_b << 7)); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ if (debug & DEBUG_TSI) ++ dev_notice(&wc->dev->dev, "Sending '%08x\n", wc->dmactrl); ++ wc->dmactrl &= ~0x0001ffff; ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } else { ++ wc->dmactrl &= ~0x00007fff; ++ wc->dmactrl |= (0x00004000 | (tots << 7)); ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ if (debug & DEBUG_TSI) ++ dev_notice(&wc->dev->dev, "Sending '%08x\n", wc->dmactrl); ++ wc->dmactrl &= ~0x00007fff; ++ __t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ } ++ spin_unlock_irqrestore(&wc->reglock, flags); ++} ++ ++#ifndef CONFIG_NOEXTENDED_RESET ++static void t4_extended_reset(struct t4 *wc) ++{ ++ unsigned int oldreg = t4_pci_in(wc, 0x4); ++ ++ udelay(1000); ++ ++ t4_pci_out(wc, 0x4, 0x42000000); ++ t4_pci_out(wc, 0xa, 0x42000000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0xa, 0x00080000); ++ t4_pci_out(wc, 0xa, 0x00180000); ++ t4_pci_out(wc, 0x4, oldreg); ++ ++ udelay(1000); ++} ++#endif ++ ++#define SPI_CS (0) ++#define SPI_CLK (1) ++#define SPI_IO0 (2) ++#define SPI_IO1 (3) ++#define SPI_IO3 (4) ++#define SPI_IO2 (5) ++#define SPI_IN (0) ++#define SPI_OUT (1) ++#define ESPI_REG 13 ++ ++static void t8_clear_bit(struct t4 *wc, int whichb) ++{ ++ wc->st.wrreg &= ~(1 << whichb); ++} ++ ++static void t8_set_bit(struct t4 *wc, int whichb, int val) ++{ ++ t8_clear_bit(wc, whichb); ++ wc->st.wrreg |= (val << whichb); ++} ++ ++static int t8_get_bit(struct t4 *wc, int whichb) ++{ ++ return (wc->st.rdreg >> whichb) & 1; ++} ++ ++static void set_iodir(struct t4 *wc, int whichb, int dir) ++{ ++ whichb += 16; ++ t8_clear_bit(wc, whichb); ++ t8_set_bit(wc, whichb, dir); ++} ++ ++static void write_hwreg(struct t4 *wc) ++{ ++ t4_pci_out(wc, ESPI_REG, wc->st.wrreg); ++} ++ ++static void read_hwreg(struct t4 *wc) ++{ ++ wc->st.rdreg = t4_pci_in(wc, ESPI_REG); ++} ++ ++static void set_cs(struct t4 *wc, int state) ++{ ++ t8_set_bit(wc, SPI_CS, state); ++ write_hwreg(wc); ++} ++ ++static void set_clk(struct t4 *wc, int clk) ++{ ++ t8_set_bit(wc, SPI_CLK, clk); ++ write_hwreg(wc); ++} ++ ++static void clk_bit_out(struct t4 *wc, int val) ++{ ++ t8_set_bit(wc, SPI_IO0, val & 1); ++ set_clk(wc, 0); ++ set_clk(wc, 1); ++} ++ ++static void shift_out(struct t4 *wc, int val) ++{ ++ int i; ++ for (i = 7; i >= 0; i--) ++ clk_bit_out(wc, (val >> i) & 1); ++} ++ ++static int clk_bit_in(struct t4 *wc) ++{ ++ int ret; ++ set_clk(wc, 0); ++ read_hwreg(wc); ++ ret = t8_get_bit(wc, SPI_IO1); ++ set_clk(wc, 1); ++ return ret; ++} ++ ++static int shift_in(struct t4 *wc) ++{ ++ int ret = 0; ++ int i; ++ int bit; ++ ++ for (i = 7; i >= 0; i--) { ++ bit = clk_bit_in(wc); ++ ret |= ((bit & 1) << i); ++ } ++ return ret; ++} ++ ++static void write_enable(struct t4 *wc) ++{ ++ int cmd = 0x06; ++ set_cs(wc, 0); ++ shift_out(wc, cmd); ++ set_cs(wc, 1); ++} ++ ++static int read_sr1(struct t4 *wc) ++{ ++ int cmd = 0x05; ++ int ret; ++ set_cs(wc, 0); ++ shift_out(wc, cmd); ++ ret = shift_in(wc); ++ set_cs(wc, 1); ++ return ret; ++} ++ ++static void clear_busy(struct t4 *wc) ++{ ++ static const int SR1_BUSY = (1 << 0); ++ unsigned long stop; ++ ++ stop = jiffies + 2*HZ; ++ while (read_sr1(wc) & SR1_BUSY) { ++ if (time_after(jiffies, stop)) { ++ if (printk_ratelimit()) { ++ dev_err(&wc->dev->dev, ++ "Lockup in %s\n", __func__); ++ } ++ break; ++ } ++ cond_resched(); ++ } ++} ++ ++static void sector_erase(struct t4 *wc, uint32_t addr) ++{ ++ int cmd = 0x20; ++ write_enable(wc); ++ set_cs(wc, 0); ++ shift_out(wc, cmd); ++ shift_out(wc, (addr >> 16) & 0xff); ++ shift_out(wc, (addr >> 8) & 0xff); ++ shift_out(wc, (addr >> 0) & 0xff); ++ set_cs(wc, 1); ++ clear_busy(wc); ++} ++ ++static void erase_half(struct t4 *wc) ++{ ++ uint32_t addr = 0x00010000; ++ uint32_t primary_firmware = 0x00250000; ++ uint32_t i; ++ ++ if(wc->devtype->flags & FLAG_2PORT | wc->devtype->flags & FLAG_1PORT) ++ primary_firmware = 0x000E0000; ++ dev_info(&wc->dev->dev, "Erasing fpag primary firmware from Address:%x\n",primary_firmware); ++ ++ for (i = addr; i < (addr + primary_firmware); i += 4096) ++ sector_erase(wc, i); ++} ++ ++ ++#define T8_FLASH_PAGE_SIZE 256UL ++ ++static void t8_update_firmware_page(struct t4 *wc, u32 address, ++ const u8 *page_data, size_t size) ++{ ++ int i; ++ ++ write_enable(wc); ++ set_cs(wc, 0); ++ shift_out(wc, 0x02); ++ shift_out(wc, (address >> 16) & 0xff); ++ shift_out(wc, (address >> 8) & 0xff); ++ shift_out(wc, (address >> 0) & 0xff); ++ ++ for (i = 0; i < size; ++i) ++ shift_out(wc, page_data[i]); ++ ++ set_cs(wc, 1); ++ clear_busy(wc); ++} ++ ++static int t8_update_firmware(struct t4 *wc, const struct firmware *fw, ++ const char *t8_firmware) ++{ ++ int res; ++ size_t offset = 0; ++ const u32 BASE_ADDRESS = 0x00010000; ++ const u8 *data, *end; ++ size_t size = 0; ++ ++ /* Erase flash */ ++ erase_half(wc); ++ ++ dev_info(&wc->dev->dev, ++ "Uploading %s. This can take up to 30 seconds.\n", t8_firmware); ++ ++ data = &fw->data[sizeof(struct t8_firm_header)]; ++ end = &fw->data[fw->size]; ++ ++ while (data < end) { ++ /* Calculate the tail end of data that's shorter than a page */ ++ size = min(T8_FLASH_PAGE_SIZE, (unsigned long)(end - data)); ++ ++ t8_update_firmware_page(wc, BASE_ADDRESS + offset, ++ data, size); ++ data += T8_FLASH_PAGE_SIZE; ++ offset += T8_FLASH_PAGE_SIZE; ++ ++ cond_resched(); ++ } ++ ++ /* Reset te820 fpga after loading firmware */ ++#if 0 ++ dev_info(&wc->dev->dev, "Firmware load complete. Reseting device.\n"); ++/*allo:*/ ++ res = pci_save_state(wc->dev); ++ if (res) ++ goto error_exit; ++ /* Set the fpga reset bits and clobber the remainder of the ++ * register, device will be reset anyway */ ++ t4_pci_out(wc, WC_LEDS, 0xe0000000); ++ msleep(1000); ++ ++ pci_restore_state(wc->dev); ++#endif ++ ++ /* Signal the driver to restart initialization. ++ * This will back out all initialization so far and ++ * restart the driver load process */ ++ return -EAGAIN; ++ ++error_exit: ++ return res; ++} ++ ++static char read_flash_byte(struct t4 *wc, unsigned int addr) ++{ ++ int cmd = 0x03; ++ char data; ++ ++ set_cs(wc, 0); ++ ++ shift_out(wc, cmd); ++ ++ shift_out(wc, (addr >> 16) & 0xff); ++ shift_out(wc, (addr >> 8) & 0xff); ++ shift_out(wc, (addr >> 0) & 0xff); ++ ++ data = shift_in(wc) & 0xff; ++ ++ set_cs(wc, 1); ++ ++ return data; ++} ++ ++/** ++ * t8_read_serial - Returns the serial number of the board. ++ * @wc: The board whos serial number we are reading. ++ * ++ * The buffer returned is dynamically allocated and must be kfree'd by the ++ * caller. If memory could not be allocated, NULL is returned. ++ * ++ * Must be called in process context. ++ * ++ */ ++static char *t8_read_serial(struct t4 *wc) ++{ ++ int i; ++ static const int MAX_SERIAL = 16; ++ static const u32 base_addr = 0x00000000; ++ unsigned char c; ++ unsigned char *serial = kzalloc(MAX_SERIAL + 1, GFP_KERNEL); ++ ++ if (!serial) ++ return NULL; ++ ++ for (i = 0; i < MAX_SERIAL; ++i) { ++ c = read_flash_byte(wc, base_addr+i); ++ if ((c >= 'a' && c <= 'z') || ++ (c >= 'A' && c <= 'Z') || ++ (c >= '0' && c <= '9')) ++ serial[i] = c; ++ else ++ break; ++ ++ } ++ ++ if (!i) { ++ kfree(serial); ++ serial = NULL; ++ } ++ ++ return serial; ++} ++ ++ ++static void setup_spi(struct t4 *wc) ++{ ++ wc->st.rdreg = wc->st.wrreg = 0; ++ ++ set_iodir(wc, SPI_IO0, SPI_OUT); ++ set_iodir(wc, SPI_IO1, SPI_IN); ++ set_iodir(wc, SPI_CS, SPI_OUT); ++ set_iodir(wc, SPI_CLK, SPI_OUT); ++ ++ t8_set_bit(wc, SPI_CS, 1); ++ t8_set_bit(wc, SPI_CLK, 1); ++ ++ write_hwreg(wc); ++} ++ ++static int t8_check_firmware(struct t4 *wc, unsigned int version) ++{ ++ int res = 0; ++ u32 crc; ++ const struct t8_firm_header *header; ++ const struct firmware *fw; ++ static const char *t8_firmware; ++ static char t8_firm[128] = "allo-dahdi-fw-2aCP4.bin"; ++ ++ if (is_octal(wc)){ ++ strcpy(t8_firm,"allo-dahdi-fw-2aCP8e.bin"); ++ } ++ else if(wc->devtype->flags & FLAG_2PORT){ ++ strcpy(t8_firm, "allo-dahdi-fw-2aCP2e.bin"); ++ } ++ else if(wc->devtype->flags & FLAG_1PORT){ ++ strcpy(t8_firm, "allo-dahdi-fw-2aCP1e.bin"); ++ } ++ else if (wc->devtype->flags & FLAG_EXPRESS){ ++ strcpy(t8_firm,"allo-dahdi-fw-2aCP4e.bin"); ++ } ++ ++ ++ t8_firmware = (const char *)t8_firm; ++ ++ res = request_firmware(&fw, t8_firmware, &wc->dev->dev); ++ if (res) { ++ dev_info(&wc->dev->dev, "firmware %s not " ++ "available from userspace\n", t8_firmware); ++ goto cleanup; ++ } ++ ++ header = (const struct t8_firm_header *)fw->data; ++ ++/* Check the crc before anything else */ ++ crc = crc32(~0, &fw->data[10], fw->size - 10) ^ ~0; ++ if (memcmp("CEMBLR", header->header, sizeof(header->header)) || ++ (le32_to_cpu(header->chksum) != crc)) { ++ dev_info(&wc->dev->dev, ++ "%s is invalid. Please reinstall.:%08X %08X\n", t8_firmware, le32_to_cpu(header->chksum), crc); ++ goto cleanup; ++ } ++ ++ /* Spi struct must be setup before any access to flash memory */ ++ setup_spi(wc); ++ ++ /* Check the two firmware versions */ ++ if (le32_to_cpu(header->version) == version) ++ goto cleanup; ++ ++ dev_info(&wc->dev->dev, "%s Version: %08x available for flash\n", ++ t8_firmware, header->version); ++ ++ res = t8_update_firmware(wc, fw, t8_firmware); ++ if (res && res != -EAGAIN) { ++ dev_info(&wc->dev->dev, "Failed to load firmware %s\n", ++ t8_firmware); ++ } ++/*allo:*/ ++#if 1 ++ release_firmware(fw); ++ return 0; ++#endif ++cleanup: ++ release_firmware(fw); ++ return res; ++} ++ ++static int ++__t4_hardware_init_1(struct t4 *wc, unsigned int cardflags, bool first_time) ++{ ++ unsigned int version; ++ int res; ++ ++ version = t4_pci_in(wc, WC_VERSION); ++ if (is_octal(wc) && first_time) { ++ dev_info(&wc->dev->dev, "Firmware Version: %01x.%02x\n", ++ (version & 0xf00) >> 8, ++ version & 0xff); ++ } else if (first_time) { ++ dev_info(&wc->dev->dev, "Firmware Version: %08x\n", version); ++ } ++ if (debug) { ++ dev_info(&wc->dev->dev, "Burst Mode: %s\n", ++ (!(cardflags & FLAG_BURST) && noburst) ? "Off" : "On"); ++#ifdef ENABLE_WORKQUEUES ++ dev_info(&wc->dev->dev, "Work Queues: Enabled\n"); ++#endif ++ } ++ ++ /* Check the field updatable firmware for the wcte820 */ ++ res = t8_check_firmware(wc, version); ++ if (res) ++ return res; ++ wc->ddev->hardware_id = t8_read_serial(wc); ++ ++#if defined(CONFIG_FORCE_EXTENDED_RESET) ++ t4_extended_reset(wc); ++#elif !defined(CONFIG_NOEXTENDED_RESET) ++ if (wc->devtype->flags & FLAG_EXPRESS) ++ t4_extended_reset(wc); ++#endif ++ ++ /* Make sure DMA engine is not running and interrupts are acknowledged */ ++ wc->dmactrl = 0x0; ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ /* Reset Framer and friends */ ++ t4_pci_out(wc, WC_LEDS, 0x00000000); ++ ++ /* Set DMA addresses */ ++ t4_pci_out(wc, WC_RDADDR, wc->readdma); ++ t4_pci_out(wc, WC_WRADDR, wc->writedma); ++ ++ /* Setup counters, interrupt flags (ignored in Gen2) */ ++ if (cardflags & FLAG_2NDGEN) { ++ t4_tsi_reset(wc); ++ } else { ++ t4_pci_out(wc, WC_COUNT, ((DAHDI_MAX_CHUNKSIZE * 2 * 32 - 1) << 18) | ((DAHDI_MAX_CHUNKSIZE * 2 * 32 - 1) << 2)); ++ } ++ ++ /* Reset pending interrupts */ ++ t4_pci_out(wc, WC_INTR, 0x00000000); ++ ++ /* Read T1/E1 status */ ++ if (first_time) { ++ if (!strcasecmp("auto", default_linemode)) { ++ if (-1 == t1e1override) { ++ wc->t1e1 = (((t4_pci_in(wc, WC_LEDS)) & ++ 0x0f00) >> 8); ++ wc->t1e1 &= 0xf; ++ if (is_octal(wc)) { ++ wc->t1e1 |= ((t4_pci_in(wc, WC_LEDS2)) & ++ 0x0f00) >> 4; ++ } ++ } else { ++ dev_warn(&wc->dev->dev, ++ "'t1e1override' is deprecated. Please use 'default_linemode'.\n"); ++ wc->t1e1 = t1e1override & 0xf; ++ } ++ } else if (!strcasecmp("t1", default_linemode)) { ++ wc->t1e1 = 0; ++ } else if (!strcasecmp("e1", default_linemode)) { ++ wc->t1e1 = 0xff; ++ } else if (!strcasecmp("j1", default_linemode)) { ++ wc->t1e1 = 0; ++ j1mode = 1; ++ } else { ++ dev_err(&wc->dev->dev, "'%s' is an unknown linemode.\n", ++ default_linemode); ++ wc->t1e1 = 0; ++ } ++ } ++ ++ wc->order = ((t4_pci_in(wc, WC_LEDS)) & 0xf0000000) >> 28; ++ order_index[wc->order]++; ++ ++ /* TE820 Auth Check */ ++ if (is_octal(wc)) { ++ unsigned long stop = jiffies + HZ; ++ uint32_t donebit; ++ ++ t4_pci_out(wc, WC_LEDS2, WC_SET_AUTH); ++ donebit = t4_pci_in(wc, WC_LEDS2); ++ while (!(donebit & WC_GET_AUTH)) { ++ if (time_after(jiffies, stop)) { ++ /* Encryption check failed, stop operation */ ++ dev_info(&wc->dev->dev, ++ "Failed encryption check. " ++ "Unloading driver.\n"); ++ return -EIO; ++ } ++ msleep(20); ++ donebit = t4_pci_in(wc, WC_LEDS2); ++ } ++ } ++ ++ return 0; ++} ++ ++static int t4_hardware_init_1(struct t4 *wc, unsigned int cardflags) ++{ ++ return __t4_hardware_init_1(wc, cardflags, true); ++} ++ ++static int __t4_hardware_init_2(struct t4 *wc, bool first_time) ++{ ++ int x; ++ unsigned int regval; ++ unsigned long flags; ++ ++ if (t4_pci_in(wc, WC_VERSION) >= 0xc01a0165) { ++ wc->tspans[0]->spanflags |= FLAG_OCTOPT; ++ } ++ /* Setup LEDS, take out of reset */ ++ t4_pci_out(wc, WC_LEDS, 0x000000ff); ++ udelay(100); ++ t4_activate(wc); ++ udelay(100); ++ ++ /* In order to find out the QFALC framer version, we have to ++ * temporarily turn off compat mode and take a peak at VSTR. We turn ++ * compat back on when we are done. ++ * ++ */ ++ spin_lock_irqsave(&wc->reglock, flags); ++ regval = __t4_framer_in(wc, 0, 0xd6); ++ if (is_octal(wc)) ++ regval |= (1 << 4); /* SSI16 - For 16 MHz multiplex mode with comp = 1 */ ++ else ++ regval |= (1 << 5); /* set COMP_DIS*/ ++ ++ __t4_framer_out(wc, 0, 0xd6, regval); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ if (!is_octal(wc)) { ++ regval = t4_framer_in(wc, 0, 0x4a); ++ if (first_time && regval == 0x05) { ++ dev_info(&wc->dev->dev, "FALC Framer Version: 2.1 or " ++ "earlier\n"); ++ } else if (regval == 0x20) { ++ if (first_time) ++ dev_info(&wc->dev->dev, "FALC Framer Version: 3.1\n"); ++ wc->falc31 = 1; ++ } else if (first_time) { ++ dev_info(&wc->dev->dev, "FALC Framer Version: Unknown " ++ "(VSTR = 0x%02x)\n", regval); ++ } ++ } ++ ++ spin_lock_irqsave(&wc->reglock, flags); ++ regval = __t4_framer_in(wc, 0, 0xd6); ++ regval &= ~(1 << 5); /* clear COMP_DIS*/ ++ __t4_framer_out(wc, 0, 0xd6, regval); ++ __t4_framer_out(wc, 0, 0x4a, 0xaa); ++ spin_unlock_irqrestore(&wc->reglock, flags); ++ ++ if (debug) { ++ dev_info(&wc->dev->dev, "Board ID: %02x\n", wc->order); ++ for (x = 0; x < 11; x++) { ++ dev_info(&wc->dev->dev, "Reg %d: 0x%08x\n", x, ++ t4_pci_in(wc, x)); ++ } ++ } ++ ++ wc->gpio = 0x00000000; ++ t4_pci_out(wc, WC_GPIO, wc->gpio); ++ t4_gpio_setdir(wc, (1 << 17), (1 << 17)); ++ t4_gpio_setdir(wc, (0xff), (0xff)); ++ ++ return 0; ++} ++ ++static int t4_hardware_init_2(struct t4 *wc) ++{ ++ return __t4_hardware_init_2(wc, true); ++} ++ ++static int __devinit t4_launch(struct t4 *wc) ++{ ++ int x; ++ int res; ++ ++ if (test_bit(DAHDI_FLAGBIT_REGISTERED, &wc->tspans[0]->span.flags)) ++ return 0; ++ ++ if (debug) { ++ dev_info(&wc->dev->dev, ++ "ALLO%dXXP: Launching card: %d\n", wc->numspans, ++ wc->order); ++ } ++ ++ wc->ddev->manufacturer = "Allo"; ++ if (!ignore_rotary && (1 == order_index[wc->order])) { ++ wc->ddev->location = kasprintf(GFP_KERNEL, ++ "Board ID Switch %d", wc->order); ++ } else { ++ bool express = ((wc->tspans[0]->spanflags & FLAG_EXPRESS) > 0); ++ wc->ddev->location = kasprintf(GFP_KERNEL, ++ "PCI%s Bus %02d Slot %02d", ++ (express) ? " Express" : "", ++ wc->dev->bus->number, ++ PCI_SLOT(wc->dev->devfn) + 1); ++ } ++ ++ if (!wc->ddev->location) ++ return -ENOMEM; ++ ++ for (x = 0; x < wc->numspans; ++x) { ++ list_add_tail(&wc->tspans[x]->span.device_node, ++ &wc->ddev->spans); ++ } ++ ++ tasklet_init(&wc->t4_tlet, t4_isr_bh, (unsigned long)wc); ++ ++ res = dahdi_register_device(wc->ddev, &wc->dev->dev); ++ if (res) { ++ dev_err(&wc->dev->dev, "Failed to register with DAHDI.\n"); ++ return res; ++ } ++ ++ return 0; ++} ++ ++/** ++ * allo4xxp_sort_cards - Sort the cards in card array by rotary switch settings. ++ * ++ */ ++static void allo4xxp_sort_cards(void) ++{ ++ int x; ++ ++ /* get the current number of probed cards and run a slice of a tail ++ * insertion sort */ ++ for (x = 0; x < MAX_T4_CARDS; x++) { ++ if (!cards[x+1]) ++ break; ++ } ++ for ( ; x > 0; x--) { ++ if (cards[x]->order < cards[x-1]->order) { ++ struct t4 *tmp = cards[x]; ++ cards[x] = cards[x-1]; ++ cards[x-1] = tmp; ++ } else { ++ /* if we're not moving it, we won't move any more ++ * since all cards are sorted on addition */ ++ break; ++ } ++ } ++} ++ ++static int __devinit ++t4_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) ++{ ++ int res; ++ struct t4 *wc; ++ unsigned int x; ++ int init_latency; ++ ++ if (pci_enable_device(pdev)) { ++ return -EIO; ++ } ++ ++ wc = kzalloc(sizeof(*wc), GFP_KERNEL); ++ if (!wc) ++ return -ENOMEM; ++ ++ wc->ddev = dahdi_create_device(); ++ if (!wc->ddev) { ++ kfree(wc); ++ return -ENOMEM; ++ } ++ ++ spin_lock_init(&wc->reglock); ++ wc->devtype = (const struct devtype *)(ent->driver_data); ++ ++#ifdef CONFIG_WCT4XXP_DISABLE_ASPM ++ if (is_pcie(wc)) { ++ pci_disable_link_state(pdev->bus->self, PCIE_LINK_STATE_L0S | ++ PCIE_LINK_STATE_L1 | PCIE_LINK_STATE_CLKPM); ++ }; ++#endif ++ ++ if (is_octal(wc)) ++ wc->numspans = 8; ++ else if (wc->devtype->flags & FLAG_2PORT) ++ wc->numspans = 2; ++ else if (wc->devtype->flags & FLAG_1PORT)/*cem:1port change, v002*/ ++ wc->numspans = 1; ++ else ++ wc->numspans = 4; ++ ++ wc->membase = pci_iomap(pdev, 0, 0); ++ /* This rids of the Double missed interrupt message after loading */ ++ wc->last0 = 1; ++ if (pci_request_regions(pdev, wc->devtype->desc)) ++ dev_info(&pdev->dev, "allo%dxxp: Unable to request regions\n", ++ wc->numspans); ++ ++ if (debug) ++ dev_info(&pdev->dev, "Found ALLO%dXXP\n", wc->numspans); ++ ++ wc->dev = pdev; ++ ++ /* Enable bus mastering */ ++ pci_set_master(pdev); ++ ++ /* Keep track of which device we are */ ++ pci_set_drvdata(pdev, wc); ++ ++ if (wc->devtype->flags & FLAG_5THGEN) { ++ if ((ms_per_irq > 1) && (latency <= ((ms_per_irq) << 1))) { ++ init_latency = ms_per_irq << 1; ++ } else { ++ if (latency > 2) ++ init_latency = latency; ++ else ++ init_latency = 2; ++ } ++/*ALLO:*/ ++ dev_info(&wc->dev->dev, "2nd gen card with initial latency of " ++ "%d and %d ms per IRQ\n", init_latency, ms_per_irq); ++ } else { ++ if (wc->devtype->flags & FLAG_2NDGEN) ++ init_latency = 1; ++ else ++ init_latency = 2; ++ } ++ ++ if (max_latency < init_latency) { ++ printk(KERN_INFO "maxlatency must be set to something greater than %d ms, increasing it to %d\n", init_latency, init_latency); ++ max_latency = init_latency; ++ } ++ ++ if (t4_allocate_buffers(wc, init_latency, NULL, NULL)) { ++ return -ENOMEM; ++ } ++ ++ /* Initialize hardware */ ++ res = t4_hardware_init_1(wc, wc->devtype->flags); ++ if (res) { ++ /* If this function returns -EAGAIN, we expect ++ * to attempt another driver load. Clean everything ++ * up first */ ++ pci_iounmap(wc->dev, wc->membase); ++ pci_release_regions(wc->dev); ++ pci_free_consistent(wc->dev, T4_BASE_SIZE(wc) * wc->numbufs * 2, ++ wc->writechunk, wc->writedma); ++ pci_set_drvdata(wc->dev, NULL); ++ free_wc(wc); ++ return res; ++ } ++ ++ for(x = 0; x < MAX_T4_CARDS; x++) { ++ if (!cards[x]) ++ break; ++ } ++ ++ if (x >= MAX_T4_CARDS) { ++ dev_notice(&wc->dev->dev, "No cards[] slot available!!\n"); ++ kfree(wc); ++ return -ENOMEM; ++ } ++ ++ wc->num = x; ++ cards[x] = wc; ++ ++#ifdef ENABLE_WORKQUEUES ++ if (wc->devtype->flags & FLAG_2NDGEN) { ++ char tmp[20]; ++ ++ sprintf(tmp, "te%dxxp[%d]", wc->numspans, wc->num); ++ wc->workq = create_workqueue(tmp); ++ } ++#endif ++ ++ /* Allocate pieces we need here */ ++ for (x = 0; x < ports_on_framer(wc); x++) { ++ struct t4_span *ts; ++ enum linemode linemode; ++ ++ ts = kzalloc(sizeof(*ts), GFP_KERNEL); ++ if (!ts) { ++ free_wc(wc); ++ return -ENOMEM; ++ } ++ wc->tspans[x] = ts; ++ ++#ifdef ENABLE_WORKQUEUES ++ INIT_WORK(&ts->swork, workq_handlespan, ts); ++#endif ++ ts->spanflags |= wc->devtype->flags; ++ linemode = (wc->t1e1 & (1 << x)) ? E1 : ((j1mode) ? J1 : T1); ++ t4_alloc_channels(wc, wc->tspans[x], linemode); ++ } ++ ++ /* Continue hardware intiialization */ ++ t4_hardware_init_2(wc); ++ ++#ifdef SUPPORT_GEN1 ++ if (request_irq(pdev->irq, (wc->devtype->flags & FLAG_2NDGEN) ? ++ t4_interrupt_gen2 : t4_interrupt, ++ DAHDI_IRQ_SHARED, ++ (wc->numspans == 8) ? "allo8xxp" : ++ (wc->numspans == 2) ? "allo2xxp" : ++ (wc->numspans == 1)? "allo1xxp" : ++ "allo4xxp", ++ wc)) { ++#else ++ if (!(wc->tspans[0]->spanflags & FLAG_2NDGEN)) { ++ dev_notice(&wc->dev->dev, "This driver does not " ++ "support 1st gen modules\n"); ++ free_wc(wc); ++ return -ENODEV; ++ } ++ ++ if (request_irq(pdev->irq, t4_interrupt_gen2, ++ DAHDI_IRQ_SHARED, "t4xxp", wc)) { ++#endif ++ dev_notice(&wc->dev->dev, "Unable to request IRQ %d\n", ++ pdev->irq); ++ free_wc(wc); ++ return -EIO; ++ } ++ ++ init_spans(wc); ++ ++ if (!ignore_rotary) ++ allo4xxp_sort_cards(); ++ ++ if (wc->ddev->hardware_id) { ++ dev_info(&wc->dev->dev, ++ "Found a Allocard: %s (SN: %s)\n", wc->devtype->desc, ++ wc->ddev->hardware_id); ++ } else { ++ dev_info(&wc->dev->dev, ++ "Found a Allocard: %s\n", wc->devtype->desc); ++ } ++ ++#ifdef VPM_SUPPORT ++ if (!wc->vpm) { ++ t4_vpm_init(wc); ++ wc->dmactrl |= (wc->vpm) ? T4_VPM_PRESENT : 0; ++ t4_pci_out(wc, WC_DMACTRL, wc->dmactrl); ++ if (wc->vpm) ++ set_span_devicetype(wc); ++ } ++#endif ++ ++ create_sysfs_files(wc); ++ ++ res = 0; ++ if (ignore_rotary) ++ res = t4_launch(wc); ++ ++ return res; ++} ++ ++static int t4_hardware_stop(struct t4 *wc) ++{ ++ ++ /* Turn off DMA, leave interrupts enabled */ ++ set_bit(T4_STOP_DMA, &wc->checkflag); ++ ++ /* Wait for interrupts to stop */ ++ msleep(25); ++ ++ /* Turn off counter, address, etc */ ++ if (wc->tspans[0]->spanflags & FLAG_2NDGEN) { ++ t4_tsi_reset(wc); ++ } else { ++ t4_pci_out(wc, WC_COUNT, 0x000000); ++ } ++ t4_pci_out(wc, WC_RDADDR, 0x0000000); ++ t4_pci_out(wc, WC_WRADDR, 0x0000000); ++ wc->gpio = 0x00000000; ++ t4_pci_out(wc, WC_GPIO, wc->gpio); ++ t4_pci_out(wc, WC_LEDS, 0x00000000); ++ ++ if (debug) { ++ dev_notice(&wc->dev->dev, "Stopped ALLO%dXXP, Turned off DMA\n", ++ wc->numspans); ++ } ++ return 0; ++} ++ ++static int __devinit ++t4_init_one_retry(struct pci_dev *pdev, const struct pci_device_id *ent) ++{ ++ int res; ++ res = t4_init_one(pdev, ent); ++ ++ /* If the driver was reset by a firmware load, ++ * try to load once again */ ++ if (-EAGAIN == res) { ++ res = t4_init_one(pdev, ent); ++ if (-EAGAIN == res) { ++ dev_err(&pdev->dev, "Failed to update firmware.\n"); ++ res = -EIO; ++ } ++ } ++ ++ return res; ++} ++ ++static void _t4_remove_one(struct t4 *wc) ++{ ++ int basesize; ++ ++ if (!wc) ++ return; ++ ++ dahdi_unregister_device(wc->ddev); ++ ++ remove_sysfs_files(wc); ++ ++ /* Stop hardware */ ++ t4_hardware_stop(wc); ++ ++#ifdef VPM_SUPPORT ++ /* Release vpm */ ++ if (wc->vpm) ++ release_vpm450m(wc->vpm); ++ wc->vpm = NULL; ++#endif ++ /* Unregister spans */ ++ ++ basesize = DAHDI_MAX_CHUNKSIZE * 32 * 4; ++ if (!(wc->tspans[0]->spanflags & FLAG_2NDGEN)) ++ basesize = basesize * 2; ++ ++#ifdef ENABLE_WORKQUEUES ++ if (wc->workq) { ++ flush_workqueue(wc->workq); ++ destroy_workqueue(wc->workq); ++ } ++#endif ++ ++ free_irq(wc->dev->irq, wc); ++ ++ if (wc->membase) ++ pci_iounmap(wc->dev, wc->membase); ++ ++ pci_release_regions(wc->dev); ++ ++ /* Immediately free resources */ ++ pci_free_consistent(wc->dev, T4_BASE_SIZE(wc) * wc->numbufs * 2, ++ wc->writechunk, wc->writedma); ++ ++ order_index[wc->order]--; ++ ++ cards[wc->num] = NULL; ++ pci_set_drvdata(wc->dev, NULL); ++ free_wc(wc); ++} ++ ++static void __devexit t4_remove_one(struct pci_dev *pdev) ++{ ++ struct t4 *wc = pci_get_drvdata(pdev); ++ if (!wc) ++ return; ++ ++ _t4_remove_one(wc); ++} ++ ++ ++static DEFINE_PCI_DEVICE_TABLE(t4_pci_tbl) = ++{ ++ ++/*ALLO:*/ ++ { 0x1d21, 0x1280, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (unsigned long)&allo1280p2 }, ++ { 0x1d21, 0x1240, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (unsigned long)&allo1240p2 }, ++ { 0x1d21, 0x1241, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (unsigned long)&allo1241p2 }, ++ { 0x1d21, 0x1220, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (unsigned long)&allo1220p2 }, ++ { 0x1d21, 0x1210, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (unsigned long)&allo1210p2 }, ++ { 0, } ++}; ++ ++static void _t4_shutdown(struct pci_dev *pdev) ++{ ++ struct t4 *wc = pci_get_drvdata(pdev); ++ t4_hardware_stop(wc); ++} ++ ++static int t4_suspend(struct pci_dev *pdev, pm_message_t state) ++{ ++ return -ENOSYS; ++} ++ ++static struct pci_driver t4_driver = { ++ .name = "allo4xxp", ++ .probe = t4_init_one_retry, ++ .remove = __devexit_p(t4_remove_one), ++ .shutdown = _t4_shutdown, ++ .suspend = t4_suspend, ++ .id_table = t4_pci_tbl, ++}; ++ ++static int __init t4_init(void) ++{ ++ int i; ++ int res; ++ ++ if (-1 != t1e1override) { ++ pr_info("'t1e1override' module parameter is deprecated. " ++ "Please use 'default_linemode' instead.\n"); ++ } ++ ++ res = dahdi_pci_module(&t4_driver); ++ if (res) ++ return -ENODEV; ++ ++ /* If we're ignoring the rotary switch settings, then we've already ++ * registered in the context of .probe */ ++ if (!ignore_rotary) { ++ ++ /* Initialize cards since we have all of them. Warn for ++ * missing zero and duplicate numbers. */ ++ ++ if (cards[0] && cards[0]->order != 0) { ++ printk(KERN_NOTICE "allo4xxp: Ident of first card is not zero (%d)\n", ++ cards[0]->order); ++ } ++ ++ for (i = 0; cards[i]; i++) { ++ /* warn the user of duplicate ident values it is ++ * probably unintended */ ++ if (debug && res < 15 && cards[i+1] && ++ cards[res]->order == cards[i+1]->order) { ++ printk(KERN_NOTICE "allo4xxp: Duplicate ident " ++ "value found (%d)\n", cards[i]->order); ++ } ++ res = t4_launch(cards[i]); ++ if (res) { ++ int j; ++ for (j = 0; j < i; ++j) ++ _t4_remove_one(cards[j]); ++ break; ++ } ++ } ++ } ++ return res; ++} ++ ++static void __exit t4_cleanup(void) ++{ ++ pci_unregister_driver(&t4_driver); ++} ++ ++MODULE_AUTHOR("Allo Incorporated <support@allo.com>"); ++MODULE_DESCRIPTION("Wildcard Dual/Quad/Octal-port Digital Card Driver"); ++MODULE_ALIAS("allo2xxp"); ++MODULE_LICENSE("GPL v2"); ++ ++module_param(debug, int, 0600); ++module_param(noburst, int, 0600); ++module_param(timingcable, int, 0600); ++module_param(t1e1override, int, 0400); ++module_param(default_linemode, charp, S_IRUGO); ++MODULE_PARM_DESC(default_linemode, "\"auto\"(default), \"e1\", \"t1\", " ++ "or \"j1\". \"auto\" will use the value from any hardware " ++ "jumpers."); ++module_param(alarmdebounce, int, 0600); ++module_param(losalarmdebounce, int, 0600); ++module_param(aisalarmdebounce, int, 0600); ++module_param(yelalarmdebounce, int, 0600); ++module_param(max_latency, int, 0600); ++module_param(j1mode, int, 0600); ++module_param(sigmode, int, 0600); ++module_param(latency, int, 0600); ++module_param(ms_per_irq, int, 0600); ++module_param(ignore_rotary, int, 0400); ++MODULE_PARM_DESC(ignore_rotary, "Set to > 0 to ignore the rotary switch when " \ ++ "registering with DAHDI."); ++ ++#ifdef VPM_SUPPORT ++module_param(vpmsupport, int, 0600); ++module_param(vpmdtmfsupport, int, 0600); ++#endif ++ ++MODULE_DEVICE_TABLE(pci, t4_pci_tbl); ++ ++module_init(t4_init); ++module_exit(t4_cleanup); +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/vpm450m.c dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/vpm450m.c +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/vpm450m.c 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/vpm450m.c 2015-02-10 15:01:15.889462151 +0100 +@@ -0,0 +1,624 @@ ++/* ++ * Copyright (C) 2005-2012 Digium, Inc. ++ * ++ * Mark Spencer <markster@digium.com> ++ * ++ * All Rights Reserved ++ */ ++ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++ ++#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt ++ ++#include <linux/slab.h> ++#include <linux/vmalloc.h> ++#include <linux/string.h> ++#include <linux/time.h> ++#include <linux/version.h> ++ ++#include <dahdi/kernel.h> ++#include <stdbool.h> ++ ++#include "vpm450m.h" ++#include <oct612x.h> ++ ++static int wct4xxp_oct612x_write(struct oct612x_context *context, ++ u32 address, u16 value) ++{ ++ struct t4 *wc = dev_get_drvdata(context->dev); ++ oct_set_reg(wc, address, value); ++ return 0; ++} ++ ++static int wct4xxp_oct612x_read(struct oct612x_context *context, u32 address, ++ u16 *value) ++{ ++ struct t4 *wc = dev_get_drvdata(context->dev); ++ *value = (u16)oct_get_reg(wc, address); ++ return 0; ++} ++ ++static int wct4xxp_oct612x_write_smear(struct oct612x_context *context, ++ u32 address, u16 value, size_t count) ++{ ++ struct t4 *wc = dev_get_drvdata(context->dev); ++ int i; ++ for (i = 0; i < count; ++i) ++ oct_set_reg(wc, address + (i << 1), value); ++ return 0; ++} ++ ++static int wct4xxp_oct612x_write_burst(struct oct612x_context *context, ++ u32 address, const u16 *buffer, ++ size_t count) ++{ ++ struct t4 *wc = dev_get_drvdata(context->dev); ++ int i; ++ for (i = 0; i < count; ++i) ++ oct_set_reg(wc, address + (i << 1), buffer[i]); ++ return 0; ++} ++ ++static int wct4xxp_oct612x_read_burst(struct oct612x_context *context, ++ u32 address, u16 *buffer, size_t count) ++{ ++ struct t4 *wc = dev_get_drvdata(context->dev); ++ int i; ++ for (i = 0; i < count; ++i) ++ buffer[i] = oct_get_reg(wc, address + (i << 1)); ++ return 0; ++} ++ ++static const struct oct612x_ops wct4xxp_oct612x_ops = { ++ .write = wct4xxp_oct612x_write, ++ .read = wct4xxp_oct612x_read, ++ .write_smear = wct4xxp_oct612x_write_smear, ++ .write_burst = wct4xxp_oct612x_write_burst, ++ .read_burst = wct4xxp_oct612x_read_burst, ++}; ++ ++#define SOUT_G168_1100GB_ON 0x40000004 ++#define SOUT_DTMF_1 0x40000011 ++#define SOUT_DTMF_2 0x40000012 ++#define SOUT_DTMF_3 0x40000013 ++#define SOUT_DTMF_A 0x4000001A ++#define SOUT_DTMF_4 0x40000014 ++#define SOUT_DTMF_5 0x40000015 ++#define SOUT_DTMF_6 0x40000016 ++#define SOUT_DTMF_B 0x4000001B ++#define SOUT_DTMF_7 0x40000017 ++#define SOUT_DTMF_8 0x40000018 ++#define SOUT_DTMF_9 0x40000019 ++#define SOUT_DTMF_C 0x4000001C ++#define SOUT_DTMF_STAR 0x4000001E ++#define SOUT_DTMF_0 0x40000010 ++#define SOUT_DTMF_POUND 0x4000001F ++#define SOUT_DTMF_D 0x4000001D ++ ++#define ROUT_G168_2100GB_ON 0x10000000 ++#define ROUT_G168_2100GB_WSPR 0x10000002 ++#define ROUT_SOUT_G168_2100HB_END 0x50000003 ++#define ROUT_G168_1100GB_ON 0x10000004 ++ ++#define ROUT_DTMF_1 0x10000011 ++#define ROUT_DTMF_2 0x10000012 ++#define ROUT_DTMF_3 0x10000013 ++#define ROUT_DTMF_A 0x1000001A ++#define ROUT_DTMF_4 0x10000014 ++#define ROUT_DTMF_5 0x10000015 ++#define ROUT_DTMF_6 0x10000016 ++#define ROUT_DTMF_B 0x1000001B ++#define ROUT_DTMF_7 0x10000017 ++#define ROUT_DTMF_8 0x10000018 ++#define ROUT_DTMF_9 0x10000019 ++#define ROUT_DTMF_C 0x1000001C ++#define ROUT_DTMF_STAR 0x1000001E ++#define ROUT_DTMF_0 0x10000010 ++#define ROUT_DTMF_POUND 0x1000001F ++#define ROUT_DTMF_D 0x1000001D ++ ++#if 0 ++#define cOCT6100_ECHO_OP_MODE_DIGITAL cOCT6100_ECHO_OP_MODE_HT_FREEZE ++#else ++#define cOCT6100_ECHO_OP_MODE_DIGITAL cOCT6100_ECHO_OP_MODE_POWER_DOWN ++#endif ++ ++struct vpm450m { ++ tPOCT6100_INSTANCE_API pApiInstance; ++ struct oct612x_context context; ++ UINT32 aulEchoChanHndl[256]; ++ int chanflags[256]; ++ int ecmode[256]; ++ int numchans; ++}; ++ ++#define FLAG_DTMF (1 << 0) ++#define FLAG_MUTE (1 << 1) ++#define FLAG_ECHO (1 << 2) ++#define FLAG_ALAW (1 << 3) ++ ++static unsigned int tones[] = { ++ SOUT_DTMF_1, ++ SOUT_DTMF_2, ++ SOUT_DTMF_3, ++ SOUT_DTMF_A, ++ SOUT_DTMF_4, ++ SOUT_DTMF_5, ++ SOUT_DTMF_6, ++ SOUT_DTMF_B, ++ SOUT_DTMF_7, ++ SOUT_DTMF_8, ++ SOUT_DTMF_9, ++ SOUT_DTMF_C, ++ SOUT_DTMF_STAR, ++ SOUT_DTMF_0, ++ SOUT_DTMF_POUND, ++ SOUT_DTMF_D, ++ SOUT_G168_1100GB_ON, ++ ++ ROUT_DTMF_1, ++ ROUT_DTMF_2, ++ ROUT_DTMF_3, ++ ROUT_DTMF_A, ++ ROUT_DTMF_4, ++ ROUT_DTMF_5, ++ ROUT_DTMF_6, ++ ROUT_DTMF_B, ++ ROUT_DTMF_7, ++ ROUT_DTMF_8, ++ ROUT_DTMF_9, ++ ROUT_DTMF_C, ++ ROUT_DTMF_STAR, ++ ROUT_DTMF_0, ++ ROUT_DTMF_POUND, ++ ROUT_DTMF_D, ++ ROUT_G168_1100GB_ON, ++}; ++ ++void vpm450m_set_alaw_companding(struct vpm450m *vpm450m, int channel, ++ bool alaw) ++{ ++ tOCT6100_CHANNEL_MODIFY *modify; ++ UINT32 ulResult; ++ UINT32 law_to_use = (alaw) ? cOCT6100_PCM_A_LAW : ++ cOCT6100_PCM_U_LAW; ++ ++ if (channel >= ARRAY_SIZE(vpm450m->chanflags)) { ++ pr_err("Channel out of bounds in %s\n", __func__); ++ return; ++ } ++ /* If we're already in this companding mode, no need to do anything. */ ++ if (alaw == ((vpm450m->chanflags[channel] & FLAG_ALAW) > 0)) ++ return; ++ ++ modify = kzalloc(sizeof(tOCT6100_CHANNEL_MODIFY), GFP_ATOMIC); ++ if (!modify) { ++ pr_notice("Unable to allocate memory for setec!\n"); ++ return; ++ } ++ ++ Oct6100ChannelModifyDef(modify); ++ modify->ulChannelHndl = vpm450m->aulEchoChanHndl[channel]; ++ modify->fTdmConfigModified = TRUE; ++ modify->TdmConfig.ulSinPcmLaw = law_to_use; ++ modify->TdmConfig.ulRinPcmLaw = law_to_use; ++ modify->TdmConfig.ulSoutPcmLaw = law_to_use; ++ modify->TdmConfig.ulRoutPcmLaw = law_to_use; ++ ulResult = Oct6100ChannelModify(vpm450m->pApiInstance, modify); ++ if (ulResult != GENERIC_OK) { ++ pr_notice("Failed to apply echo can changes on channel %d %d %08x!\n", ++ vpm450m->aulEchoChanHndl[channel], channel, ulResult); ++ } else { ++ if (debug) { ++ pr_info("Changed companding on channel %d to %s.\n", ++ channel, (alaw) ? "alaw" : "ulaw"); ++ } ++ if (alaw) ++ vpm450m->chanflags[channel] |= FLAG_ALAW; ++ else ++ vpm450m->chanflags[channel] &= ~(FLAG_ALAW); ++ } ++ kfree(modify); ++} ++ ++static void vpm450m_setecmode(struct vpm450m *vpm450m, int channel, int mode) ++{ ++ tOCT6100_CHANNEL_MODIFY *modify; ++ UINT32 ulResult; ++ ++ if (vpm450m->ecmode[channel] == mode) ++ return; ++ modify = kmalloc(sizeof(tOCT6100_CHANNEL_MODIFY), GFP_ATOMIC); ++ if (!modify) { ++ printk(KERN_NOTICE "wct4xxp: Unable to allocate memory for setec!\n"); ++ return; ++ } ++ Oct6100ChannelModifyDef(modify); ++ modify->ulEchoOperationMode = mode; ++ modify->ulChannelHndl = vpm450m->aulEchoChanHndl[channel]; ++ ulResult = Oct6100ChannelModify(vpm450m->pApiInstance, modify); ++ if (ulResult != GENERIC_OK) { ++ printk(KERN_NOTICE "Failed to apply echo can changes on channel %d %08x!\n", channel, ulResult); ++ } else { ++#ifdef OCTASIC_DEBUG ++ printk(KERN_DEBUG "Echo can on channel %d set to %d\n", channel, mode); ++#endif ++ vpm450m->ecmode[channel] = mode; ++ } ++ kfree(modify); ++} ++ ++void vpm450m_setdtmf(struct vpm450m *vpm450m, int channel, int detect, int mute) ++{ ++ tOCT6100_CHANNEL_MODIFY *modify; ++ UINT32 ulResult; ++ ++ if (channel >= ARRAY_SIZE(vpm450m->chanflags)) { ++ pr_err("Channel out of bounds in %s\n", __func__); ++ return; ++ } ++ ++ modify = kmalloc(sizeof(tOCT6100_CHANNEL_MODIFY), GFP_KERNEL); ++ if (!modify) { ++ printk(KERN_NOTICE "wct4xxp: Unable to allocate memory for setdtmf!\n"); ++ return; ++ } ++ Oct6100ChannelModifyDef(modify); ++ modify->ulChannelHndl = vpm450m->aulEchoChanHndl[channel]; ++ if (mute) { ++ vpm450m->chanflags[channel] |= FLAG_MUTE; ++ modify->VqeConfig.fDtmfToneRemoval = TRUE; ++ } else { ++ vpm450m->chanflags[channel] &= ~FLAG_MUTE; ++ modify->VqeConfig.fDtmfToneRemoval = FALSE; ++ } ++ if (detect) ++ vpm450m->chanflags[channel] |= FLAG_DTMF; ++ else ++ vpm450m->chanflags[channel] &= ~FLAG_DTMF; ++ if (vpm450m->chanflags[channel] & (FLAG_DTMF|FLAG_MUTE)) { ++ if (!(vpm450m->chanflags[channel] & FLAG_ECHO)) { ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_HT_RESET); ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_HT_FREEZE); ++ } ++ } else { ++ if (!(vpm450m->chanflags[channel] & FLAG_ECHO)) ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_DIGITAL); ++ } ++ ++ ulResult = Oct6100ChannelModify(vpm450m->pApiInstance, modify); ++ if (ulResult != GENERIC_OK) { ++ printk(KERN_NOTICE "Failed to apply dtmf mute changes on channel %d!\n", channel); ++ } ++/* printk(KERN_DEBUG "VPM450m: Setting DTMF on channel %d: %s / %s\n", channel, (detect ? "DETECT" : "NO DETECT"), (mute ? "MUTE" : "NO MUTE")); */ ++ kfree(modify); ++} ++ ++void vpm450m_setec(struct vpm450m *vpm450m, int channel, int eclen) ++{ ++ if (channel >= ARRAY_SIZE(vpm450m->chanflags)) { ++ pr_err("Channel out of bounds in %s\n", __func__); ++ return; ++ } ++ ++ if (eclen) { ++ vpm450m->chanflags[channel] |= FLAG_ECHO; ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_HT_RESET); ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_NORMAL); ++ } else { ++ vpm450m->chanflags[channel] &= ~FLAG_ECHO; ++ if (vpm450m->chanflags[channel] & (FLAG_DTMF | FLAG_MUTE)) { ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_HT_RESET); ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_HT_FREEZE); ++ } else ++ vpm450m_setecmode(vpm450m, channel, cOCT6100_ECHO_OP_MODE_DIGITAL); ++ } ++/* printk(KERN_DEBUG "VPM450m: Setting EC on channel %d to %d\n", channel, eclen); */ ++} ++ ++int vpm450m_checkirq(struct vpm450m *vpm450m) ++{ ++ tOCT6100_INTERRUPT_FLAGS InterruptFlags; ++ ++ Oct6100InterruptServiceRoutineDef(&InterruptFlags); ++ Oct6100InterruptServiceRoutine(vpm450m->pApiInstance, &InterruptFlags); ++ ++ return InterruptFlags.fToneEventsPending ? 1 : 0; ++} ++ ++int vpm450m_getdtmf(struct vpm450m *vpm450m, int *channel, int *tone, int *start) ++{ ++ tOCT6100_TONE_EVENT tonefound; ++ tOCT6100_EVENT_GET_TONE tonesearch; ++ ++ Oct6100EventGetToneDef(&tonesearch); ++ tonesearch.pToneEvent = &tonefound; ++ tonesearch.ulMaxToneEvent = 1; ++ Oct6100EventGetTone(vpm450m->pApiInstance, &tonesearch); ++ if (tonesearch.ulNumValidToneEvent) { ++ if (channel) ++ *channel = tonefound.ulUserChanId; ++ if (tone) { ++ switch(tonefound.ulToneDetected) { ++ case SOUT_DTMF_1: ++ *tone = '1'; ++ break; ++ case SOUT_DTMF_2: ++ *tone = '2'; ++ break; ++ case SOUT_DTMF_3: ++ *tone = '3'; ++ break; ++ case SOUT_DTMF_A: ++ *tone = 'A'; ++ break; ++ case SOUT_DTMF_4: ++ *tone = '4'; ++ break; ++ case SOUT_DTMF_5: ++ *tone = '5'; ++ break; ++ case SOUT_DTMF_6: ++ *tone = '6'; ++ break; ++ case SOUT_DTMF_B: ++ *tone = 'B'; ++ break; ++ case SOUT_DTMF_7: ++ *tone = '7'; ++ break; ++ case SOUT_DTMF_8: ++ *tone = '8'; ++ break; ++ case SOUT_DTMF_9: ++ *tone = '9'; ++ break; ++ case SOUT_DTMF_C: ++ *tone = 'C'; ++ break; ++ case SOUT_DTMF_STAR: ++ *tone = '*'; ++ break; ++ case SOUT_DTMF_0: ++ *tone = '0'; ++ break; ++ case SOUT_DTMF_POUND: ++ *tone = '#'; ++ break; ++ case SOUT_DTMF_D: ++ *tone = 'D'; ++ break; ++ case SOUT_G168_1100GB_ON: ++ *tone = 'f'; ++ break; ++ default: ++#ifdef OCTASIC_DEBUG ++ printk(KERN_DEBUG "Unknown tone value %08x\n", tonefound.ulToneDetected); ++#endif ++ *tone = 'u'; ++ break; ++ } ++ } ++ if (start) ++ *start = (tonefound.ulEventType == cOCT6100_TONE_PRESENT); ++ return 1; ++ } ++ return 0; ++} ++ ++unsigned int get_vpm450m_capacity(struct device *device) ++{ ++ struct oct612x_context context; ++ UINT32 ulResult; ++ ++ tOCT6100_API_GET_CAPACITY_PINS CapacityPins; ++ ++ context.dev = device; ++ context.ops = &wct4xxp_oct612x_ops; ++ ++ Oct6100ApiGetCapacityPinsDef(&CapacityPins); ++ CapacityPins.pProcessContext = &context; ++ CapacityPins.ulMemoryType = cOCT6100_MEM_TYPE_DDR; ++ CapacityPins.fEnableMemClkOut = TRUE; ++ CapacityPins.ulMemClkFreq = cOCT6100_MCLK_FREQ_133_MHZ; ++ ++ ulResult = Oct6100ApiGetCapacityPins(&CapacityPins); ++ if (ulResult != cOCT6100_ERR_OK) { ++ printk(KERN_DEBUG "Failed to get chip capacity, code %08x!\n", ulResult); ++ return 0; ++ } ++ ++ return CapacityPins.ulCapacityValue; ++} ++ ++struct vpm450m *init_vpm450m(struct device *device, int *isalaw, ++ int numspans, const struct firmware *firmware) ++{ ++ tOCT6100_CHIP_OPEN *ChipOpen; ++ tOCT6100_GET_INSTANCE_SIZE InstanceSize; ++ tOCT6100_CHANNEL_OPEN *ChannelOpen; ++ UINT32 ulResult; ++ const unsigned int mask = (8 == numspans) ? 0x7 : 0x3; ++ unsigned int sout_stream, rout_stream; ++ struct vpm450m *vpm450m; ++ int x,y,law; ++ ++ if (!(vpm450m = kmalloc(sizeof(struct vpm450m), GFP_KERNEL))) ++ return NULL; ++ ++ memset(vpm450m, 0, sizeof(struct vpm450m)); ++ vpm450m->context.dev = device; ++ vpm450m->context.ops = &wct4xxp_oct612x_ops; ++ ++ if (!(ChipOpen = kmalloc(sizeof(tOCT6100_CHIP_OPEN), GFP_KERNEL))) { ++ kfree(vpm450m); ++ return NULL; ++ } ++ ++ memset(ChipOpen, 0, sizeof(tOCT6100_CHIP_OPEN)); ++ ++ if (!(ChannelOpen = kmalloc(sizeof(tOCT6100_CHANNEL_OPEN), GFP_KERNEL))) { ++ kfree(vpm450m); ++ kfree(ChipOpen); ++ return NULL; ++ } ++ ++ memset(ChannelOpen, 0, sizeof(tOCT6100_CHANNEL_OPEN)); ++ ++ for (x = 0; x < ARRAY_SIZE(vpm450m->ecmode); x++) ++ vpm450m->ecmode[x] = -1; ++ ++ vpm450m->numchans = numspans * 32; ++ printk(KERN_INFO "VPM450: echo cancellation for %d channels\n", vpm450m->numchans); ++ ++ Oct6100ChipOpenDef(ChipOpen); ++ ++ /* Setup Chip Open Parameters */ ++ ChipOpen->ulUpclkFreq = cOCT6100_UPCLK_FREQ_33_33_MHZ; ++ Oct6100GetInstanceSizeDef(&InstanceSize); ++ ++ ChipOpen->pProcessContext = &vpm450m->context; ++ ++ ChipOpen->pbyImageFile = firmware->data; ++ ChipOpen->ulImageSize = firmware->size; ++ ChipOpen->fEnableMemClkOut = TRUE; ++ ChipOpen->ulMemClkFreq = cOCT6100_MCLK_FREQ_133_MHZ; ++ ChipOpen->ulMaxChannels = vpm450m->numchans; ++ ChipOpen->ulMemoryType = cOCT6100_MEM_TYPE_DDR; ++ ChipOpen->ulMemoryChipSize = cOCT6100_MEMORY_CHIP_SIZE_32MB; ++ ChipOpen->ulNumMemoryChips = 1; ++ ChipOpen->aulTdmStreamFreqs[0] = cOCT6100_TDM_STREAM_FREQ_8MHZ; ++ ChipOpen->ulMaxFlexibleConfParticipants = 0; ++ ChipOpen->ulMaxConfBridges = 0; ++ ChipOpen->ulMaxRemoteDebugSessions = 0; ++ ChipOpen->fEnableChannelRecording = FALSE; ++ ChipOpen->ulSoftToneEventsBufSize = 64; ++ ++ if (vpm450m->numchans <= 128) { ++ ChipOpen->ulMaxTdmStreams = 4; ++ ChipOpen->ulTdmSampling = cOCT6100_TDM_SAMPLE_AT_FALLING_EDGE; ++ } else { ++ ChipOpen->ulMaxTdmStreams = 32; ++ ChipOpen->fEnableFastH100Mode = TRUE; ++ ChipOpen->ulTdmSampling = cOCT6100_TDM_SAMPLE_AT_RISING_EDGE; ++ } ++ ++#if 0 ++ ChipOpen->fEnableAcousticEcho = TRUE; ++#endif ++ ++ ulResult = Oct6100GetInstanceSize(ChipOpen, &InstanceSize); ++ if (ulResult != cOCT6100_ERR_OK) { ++ printk(KERN_NOTICE "Failed to get instance size, code %08x!\n", ulResult); ++ kfree(vpm450m); ++ kfree(ChipOpen); ++ kfree(ChannelOpen); ++ return NULL; ++ } ++ ++ vpm450m->pApiInstance = vmalloc(InstanceSize.ulApiInstanceSize); ++ if (!vpm450m->pApiInstance) { ++ printk(KERN_NOTICE "Out of memory (can't allocate %d bytes)!\n", InstanceSize.ulApiInstanceSize); ++ kfree(vpm450m); ++ kfree(ChipOpen); ++ kfree(ChannelOpen); ++ return NULL; ++ } ++ ++ ulResult = Oct6100ChipOpen(vpm450m->pApiInstance, ChipOpen); ++ if (ulResult != cOCT6100_ERR_OK) { ++ printk(KERN_NOTICE "Failed to open chip, code %08x!\n", ulResult); ++ vfree(vpm450m->pApiInstance); ++ kfree(vpm450m); ++ kfree(ChipOpen); ++ kfree(ChannelOpen); ++ return NULL; ++ } ++ ++ sout_stream = (8 == numspans) ? 29 : 2; ++ rout_stream = (8 == numspans) ? 24 : 3; ++ ++ for (x = 0; x < ((8 == numspans) ? 256 : 128); x++) { ++ /* execute this loop always on 4 span cards but ++ * on 2 span cards only execute for the channels related to our spans */ ++ if (( numspans > 2) || ((x & 0x03) <numspans)) { /*allo: 1port support*/ ++ /* span timeslots are interleaved 12341234... ++ * therefore, the lower 2 bits tell us which span this ++ * timeslot/channel ++ */ ++ if (isalaw[x & mask]) { ++ law = cOCT6100_PCM_A_LAW; ++ vpm450m->chanflags[x] |= FLAG_ALAW; ++ } else { ++ law = cOCT6100_PCM_U_LAW; ++ vpm450m->chanflags[x] &= ~(FLAG_ALAW); ++ } ++ Oct6100ChannelOpenDef(ChannelOpen); ++ ChannelOpen->pulChannelHndl = &vpm450m->aulEchoChanHndl[x]; ++ ChannelOpen->ulUserChanId = x; ++ ChannelOpen->TdmConfig.ulRinPcmLaw = law; ++ ChannelOpen->TdmConfig.ulRinStream = 0; ++ ChannelOpen->TdmConfig.ulRinTimeslot = x; ++ ChannelOpen->TdmConfig.ulSinPcmLaw = law; ++ ChannelOpen->TdmConfig.ulSinStream = 1; ++ ChannelOpen->TdmConfig.ulSinTimeslot = x; ++ ChannelOpen->TdmConfig.ulSoutPcmLaw = law; ++ ChannelOpen->TdmConfig.ulSoutStream = sout_stream; ++ ChannelOpen->TdmConfig.ulSoutTimeslot = x; ++#if 1 ++ ChannelOpen->TdmConfig.ulRoutPcmLaw = law; ++ ChannelOpen->TdmConfig.ulRoutStream = rout_stream; ++ ChannelOpen->TdmConfig.ulRoutTimeslot = x; ++#endif ++ ChannelOpen->VqeConfig.fEnableNlp = TRUE; ++ ChannelOpen->VqeConfig.fRinDcOffsetRemoval = TRUE; ++ ChannelOpen->VqeConfig.fSinDcOffsetRemoval = TRUE; ++ ++ ChannelOpen->fEnableToneDisabler = TRUE; ++ ChannelOpen->ulEchoOperationMode = cOCT6100_ECHO_OP_MODE_DIGITAL; ++ ++ ulResult = Oct6100ChannelOpen(vpm450m->pApiInstance, ChannelOpen); ++ if (ulResult != GENERIC_OK) { ++ printk(KERN_NOTICE "Failed to open channel %d %x!\n", x, ulResult); ++ continue; ++ } ++ for (y = 0; y < ARRAY_SIZE(tones); y++) { ++ tOCT6100_TONE_DETECTION_ENABLE enable; ++ Oct6100ToneDetectionEnableDef(&enable); ++ enable.ulChannelHndl = vpm450m->aulEchoChanHndl[x]; ++ enable.ulToneNumber = tones[y]; ++ if (Oct6100ToneDetectionEnable(vpm450m->pApiInstance, &enable) != GENERIC_OK) ++ printk(KERN_NOTICE "Failed to enable tone detection on channel %d for tone %d!\n", x, y); ++ } ++ } ++ } ++ ++ kfree(ChipOpen); ++ kfree(ChannelOpen); ++ return vpm450m; ++} ++ ++void release_vpm450m(struct vpm450m *vpm450m) ++{ ++ UINT32 ulResult; ++ tOCT6100_CHIP_CLOSE ChipClose; ++ ++ Oct6100ChipCloseDef(&ChipClose); ++ ulResult = Oct6100ChipClose(vpm450m->pApiInstance, &ChipClose); ++ if (ulResult != cOCT6100_ERR_OK) { ++ printk(KERN_NOTICE "Failed to close chip, code %08x!\n", ulResult); ++ } ++ vfree(vpm450m->pApiInstance); ++ kfree(vpm450m); ++} +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/vpm450m.h dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/vpm450m.h +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/vpm450m.h 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/vpm450m.h 2015-02-10 15:01:15.889462151 +0100 +@@ -0,0 +1,49 @@ ++/* ++ * Copyright (C) 2005-2006 Digium, Inc. ++ * ++ * Mark Spencer <markster@digium.com> ++ * ++ * All Rights Reserved ++ * ++ */ ++ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++ ++#ifndef _VPM450M_H ++#define _VPM450M_H ++ ++#include <linux/firmware.h> ++ ++struct t4; ++struct vpm450m; ++ ++/* From driver */ ++unsigned int oct_get_reg(void *data, unsigned int reg); ++void oct_set_reg(void *data, unsigned int reg, unsigned int val); ++ ++/* From vpm450m */ ++struct vpm450m *init_vpm450m(struct device *device, int *isalaw, ++ int numspans, const struct firmware *firmware); ++unsigned int get_vpm450m_capacity(struct device *device); ++void vpm450m_setec(struct vpm450m *instance, int channel, int eclen); ++void vpm450m_setdtmf(struct vpm450m *instance, int channel, int dtmfdetect, int dtmfmute); ++int vpm450m_checkirq(struct vpm450m *vpm450m); ++int vpm450m_getdtmf(struct vpm450m *vpm450m, int *channel, int *tone, int *start); ++void release_vpm450m(struct vpm450m *instance); ++void vpm450m_set_alaw_companding(struct vpm450m *vpm450m, ++ int channel, bool alaw); ++ ++extern int debug; ++ ++#endif +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/wct4xxp-diag.c dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/wct4xxp-diag.c +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/wct4xxp-diag.c 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/wct4xxp-diag.c 2015-02-10 15:01:15.892795252 +0100 +@@ -0,0 +1,427 @@ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++ ++#include <fcntl.h> ++#include <stdio.h> ++#include <stdlib.h> ++#include <unistd.h> ++#include <sys/ioctl.h> ++#include <errno.h> ++#include <string.h> ++#include <dahdi/user.h> ++#include "wct4xxp.h" ++ ++struct t4_reg_def { ++ int reg; ++ char *name; ++ int global; ++}; ++static struct t4_reg_def xreginfo[] = { ++ { 0x00, "RDADDR" }, ++ { 0x01, "WRADDR" }, ++ { 0x02, "COUNT" }, ++ { 0x03, "DMACTRL" }, ++ { 0x04, "WCINTR" }, ++ { 0x06, "VERSION" }, ++ { 0x07, "LEDS" }, ++ { 0x08, "GPIOCTL" }, ++ { 0x09, "GPIO" }, ++ { 0x0A, "LADDR" }, ++ { 0x0b, "LDATA" }, ++}; ++ ++static struct t4_reg_def reginfo[] = { ++ { 0x00, "XFIFO" }, ++ { 0x01, "XFIFO" }, ++ { 0x02, "CMDR" }, ++ { 0x03, "MODE" }, ++ { 0x04, "RAH1" }, ++ { 0x05, "RAH2" }, ++ { 0x06, "RAL1" }, ++ { 0x07, "RAL2" }, ++ { 0x08, "IPC", 1 }, ++ { 0x09, "CCR1" }, ++ { 0x0a, "CCR2" }, ++ { 0x0c, "RTR1" }, ++ { 0x0d, "RTR2" }, ++ { 0x0e, "RTR3" }, ++ { 0x0f, "RTR4" }, ++ { 0x10, "TTR1" }, ++ { 0x11, "TTR2" }, ++ { 0x12, "TTR3" }, ++ { 0x13, "TTR4" }, ++ { 0x14, "IMR0" }, ++ { 0x15, "IMR1" }, ++ { 0x16, "IMR2" }, ++ { 0x17, "IMR3" }, ++ { 0x18, "IMR4" }, ++ { 0x1b, "IERR" }, ++ { 0x1c, "FMR0" }, ++ { 0x1d, "FMR1" }, ++ { 0x1e, "FMR2" }, ++ { 0x1f, "LOOP" }, ++ { 0x20, "XSW" }, ++ { 0x21, "XSP" }, ++ { 0x22, "XC0" }, ++ { 0x23, "XC1" }, ++ { 0x24, "RC0" }, ++ { 0x25, "RC1" }, ++ { 0x26, "XPM0" }, ++ { 0x27, "XPM1" }, ++ { 0x28, "XPM2" }, ++ { 0x29, "TSWM" }, ++ { 0x2b, "IDLE" }, ++ { 0x2c, "XSA4" }, ++ { 0x2d, "XSA5" }, ++ { 0x2e, "XSA6" }, ++ { 0x2f, "XSA7" }, ++ { 0x30, "XSA8" }, ++ { 0x31, "FMR3" }, ++ { 0x32, "ICB1" }, ++ { 0x33, "ICB2" }, ++ { 0x34, "ICB3" }, ++ { 0x35, "ICB4" }, ++ { 0x36, "LIM0" }, ++ { 0x37, "LIM1" }, ++ { 0x38, "PCD" }, ++ { 0x39, "PCR" }, ++ { 0x3a, "LIM2" }, ++ { 0x3b, "LCR1" }, ++ { 0x3c, "LCR2" }, ++ { 0x3d, "LCR3" }, ++ { 0x3e, "SIC1" }, ++ { 0x3f, "SIC2" }, ++ { 0x40, "SIC3" }, ++ { 0x44, "CMR1" }, ++ { 0x45, "CMR2" }, ++ { 0x46, "GCR" }, ++ { 0x47, "ESM" }, ++ { 0x60, "DEC" }, ++ { 0x70, "XS1" }, ++ { 0x71, "XS2" }, ++ { 0x72, "XS3" }, ++ { 0x73, "XS4" }, ++ { 0x74, "XS5" }, ++ { 0x75, "XS6" }, ++ { 0x76, "XS7" }, ++ { 0x77, "XS8" }, ++ { 0x78, "XS9" }, ++ { 0x79, "XS10" }, ++ { 0x7a, "XS11" }, ++ { 0x7b, "XS12" }, ++ { 0x7c, "XS13" }, ++ { 0x7d, "XS14" }, ++ { 0x7e, "XS15" }, ++ { 0x7f, "XS16" }, ++ { 0x80, "PC1" }, ++ { 0x81, "PC2" }, ++ { 0x82, "PC3" }, ++ { 0x83, "PC4" }, ++ { 0x84, "PC5" }, ++ { 0x85, "GPC1", 1 }, ++ { 0x87, "CMDR2" }, ++ { 0x8d, "CCR5" }, ++ { 0x92, "GCM1", 1 }, ++ { 0x93, "GCM2", 1 }, ++ { 0x94, "GCM3", 1 }, ++ { 0x95, "GCM4", 1 }, ++ { 0x96, "GCM5", 1 }, ++ { 0x97, "GCM6", 1 }, ++ { 0x98, "GCM7", 1 }, ++ { 0x99, "GCM8", 1 }, ++ { 0xa0, "TSEO" }, ++ { 0xa1, "TSBS1" }, ++ { 0xa8, "TPC0" }, ++}; ++ ++static struct t4_reg_def t1_reginfo[] = { ++ { 0x00, "XFIFO" }, ++ { 0x01, "XFIFO" }, ++ { 0x02, "CMDR" }, ++ { 0x03, "MODE" }, ++ { 0x04, "RAH1" }, ++ { 0x05, "RAH2" }, ++ { 0x06, "RAL1" }, ++ { 0x07, "RAL2" }, ++ { 0x08, "IPC", 1 }, ++ { 0x09, "CCR1" }, ++ { 0x0a, "CCR2" }, ++ { 0x0c, "RTR1" }, ++ { 0x0d, "RTR2" }, ++ { 0x0e, "RTR3" }, ++ { 0x0f, "RTR4" }, ++ { 0x10, "TTR1" }, ++ { 0x11, "TTR2" }, ++ { 0x12, "TTR3" }, ++ { 0x13, "TTR4" }, ++ { 0x14, "IMR0" }, ++ { 0x15, "IMR1" }, ++ { 0x16, "IMR2" }, ++ { 0x17, "IMR3" }, ++ { 0x18, "IMR4" }, ++ { 0x1b, "IERR" }, ++ { 0x1c, "FMR0" }, ++ { 0x1d, "FMR1" }, ++ { 0x1e, "FMR2" }, ++ { 0x1f, "LOOP" }, ++ { 0x20, "FMR4" }, ++ { 0x21, "FMR5" }, ++ { 0x22, "XC0" }, ++ { 0x23, "XC1" }, ++ { 0x24, "RC0" }, ++ { 0x25, "RC1" }, ++ { 0x26, "XPM0" }, ++ { 0x27, "XPM1" }, ++ { 0x28, "XPM2" }, ++ { 0x2b, "IDLE" }, ++ { 0x2c, "XDL1" }, ++ { 0x2d, "XDL2" }, ++ { 0x2e, "XDL3" }, ++ { 0x2f, "CCB1" }, ++ { 0x30, "CCB2" }, ++ { 0x31, "CCB3" }, ++ { 0x32, "ICB1" }, ++ { 0x33, "ICB2" }, ++ { 0x34, "ICB3" }, ++ { 0x36, "LIM0" }, ++ { 0x37, "LIM1" }, ++ { 0x38, "PCD" }, ++ { 0x39, "PCR" }, ++ { 0x3a, "LIM2" }, ++ { 0x3b, "LCR1" }, ++ { 0x3c, "LCR2" }, ++ { 0x3d, "LCR3" }, ++ { 0x3e, "SIC1" }, ++ { 0x3f, "SIC2" }, ++ { 0x40, "SIC3" }, ++ { 0x44, "CMR1" }, ++ { 0x45, "CMR2" }, ++ { 0x46, "GCR" }, ++ { 0x47, "ESM" }, ++ { 0x60, "DEC" }, ++ { 0x70, "XS1" }, ++ { 0x71, "XS2" }, ++ { 0x72, "XS3" }, ++ { 0x73, "XS4" }, ++ { 0x74, "XS5" }, ++ { 0x75, "XS6" }, ++ { 0x76, "XS7" }, ++ { 0x77, "XS8" }, ++ { 0x78, "XS9" }, ++ { 0x79, "XS10" }, ++ { 0x7a, "XS11" }, ++ { 0x7b, "XS12" }, ++ { 0x80, "PC1" }, ++ { 0x81, "PC2" }, ++ { 0x82, "PC3" }, ++ { 0x83, "PC4" }, ++ { 0x84, "PC5" }, ++ { 0x85, "GPC1", 1 }, ++ { 0x87, "CMDR2" }, ++ { 0x8d, "CCR5" }, ++ { 0x92, "GCM1", 1 }, ++ { 0x93, "GCM2", 1 }, ++ { 0x94, "GCM3", 1 }, ++ { 0x95, "GCM4", 1 }, ++ { 0x96, "GCM5", 1 }, ++ { 0x97, "GCM6", 1 }, ++ { 0x98, "GCM7", 1 }, ++ { 0x99, "GCM8", 1 }, ++ { 0xa0, "TSEO" }, ++ { 0xa1, "TSBS1" }, ++ { 0xa8, "TPC0" }, ++}; ++ ++static struct t4_reg_def t1_sreginfo[] = { ++ { 0x00, "RFIFO" }, ++ { 0x01, "RFIFO" }, ++ { 0x49, "RBD" }, ++ { 0x4a, "VSTR", 1 }, ++ { 0x4b, "RES" }, ++ { 0x4c, "FRS0" }, ++ { 0x4d, "FRS1" }, ++ { 0x4e, "FRS2" }, ++ { 0x4f, "Old FRS1" }, ++ { 0x50, "FECL" }, ++ { 0x51, "FECH" }, ++ { 0x52, "CVCL" }, ++ { 0x53, "CVCH" }, ++ { 0x54, "CECL" }, ++ { 0x55, "CECH" }, ++ { 0x56, "EBCL" }, ++ { 0x57, "EBCH" }, ++ { 0x58, "BECL" }, ++ { 0x59, "BECH" }, ++ { 0x5a, "COEC" }, ++ { 0x5c, "RDL1" }, ++ { 0x5d, "RDL2" }, ++ { 0x5e, "RDL3" }, ++ { 0x62, "RSP1" }, ++ { 0x63, "RSP2" }, ++ { 0x64, "SIS" }, ++ { 0x65, "RSIS" }, ++ { 0x66, "RBCL" }, ++ { 0x67, "RBCH" }, ++ { 0x68, "ISR0" }, ++ { 0x69, "ISR1" }, ++ { 0x6a, "ISR2" }, ++ { 0x6b, "ISR3" }, ++ { 0x6c, "ISR4" }, ++ { 0x6e, "GIS" }, ++ { 0x6f, "CIS", 1 }, ++ { 0x70, "RS1" }, ++ { 0x71, "RS2" }, ++ { 0x72, "RS3" }, ++ { 0x73, "RS4" }, ++ { 0x74, "RS5" }, ++ { 0x75, "RS6" }, ++ { 0x76, "RS7" }, ++ { 0x77, "RS8" }, ++ { 0x78, "RS9" }, ++ { 0x79, "RS10" }, ++ { 0x7a, "RS11" }, ++ { 0x7b, "RS12" }, ++}; ++ ++static struct t4_reg_def sreginfo[] = { ++ { 0x00, "RFIFO" }, ++ { 0x01, "RFIFO" }, ++ { 0x49, "RBD" }, ++ { 0x4a, "VSTR", 1 }, ++ { 0x4b, "RES" }, ++ { 0x4c, "FRS0" }, ++ { 0x4d, "FRS1" }, ++ { 0x4e, "RSW" }, ++ { 0x4f, "RSP" }, ++ { 0x50, "FECL" }, ++ { 0x51, "FECH" }, ++ { 0x52, "CVCL" }, ++ { 0x53, "CVCH" }, ++ { 0x54, "CEC1L" }, ++ { 0x55, "CEC1H" }, ++ { 0x56, "EBCL" }, ++ { 0x57, "EBCH" }, ++ { 0x58, "CEC2L" }, ++ { 0x59, "CEC2H" }, ++ { 0x5a, "CEC3L" }, ++ { 0x5b, "CEC3H" }, ++ { 0x5c, "RSA4" }, ++ { 0x5d, "RSA5" }, ++ { 0x5e, "RSA6" }, ++ { 0x5f, "RSA7" }, ++ { 0x60, "RSA8" }, ++ { 0x61, "RSA6S" }, ++ { 0x62, "RSP1" }, ++ { 0x63, "RSP2" }, ++ { 0x64, "SIS" }, ++ { 0x65, "RSIS" }, ++ { 0x66, "RBCL" }, ++ { 0x67, "RBCH" }, ++ { 0x68, "ISR0" }, ++ { 0x69, "ISR1" }, ++ { 0x6a, "ISR2" }, ++ { 0x6b, "ISR3" }, ++ { 0x6c, "ISR4" }, ++ { 0x6e, "GIS" }, ++ { 0x6f, "CIS", 1 }, ++ { 0x70, "RS1" }, ++ { 0x71, "RS2" }, ++ { 0x72, "RS3" }, ++ { 0x73, "RS4" }, ++ { 0x74, "RS5" }, ++ { 0x75, "RS6" }, ++ { 0x76, "RS7" }, ++ { 0x77, "RS8" }, ++ { 0x78, "RS9" }, ++ { 0x79, "RS10" }, ++ { 0x7a, "RS11" }, ++ { 0x7b, "RS12" }, ++ { 0x7c, "RS13" }, ++ { 0x7d, "RS14" }, ++ { 0x7e, "RS15" }, ++ { 0x7f, "RS16" }, ++}; ++ ++static char *tobin(int x) ++{ ++ static char s[9] = ""; ++ int y,z=0; ++ for (y=7;y>=0;y--) { ++ if (x & (1 << y)) ++ s[z++] = '1'; ++ else ++ s[z++] = '0'; ++ } ++ s[z] = '\0'; ++ return s; ++} ++ ++static char *tobin32(unsigned int x) ++{ ++ static char s[33] = ""; ++ int y,z=0; ++ for (y=31;y>=0;y--) { ++ if (x & (1 << y)) ++ s[z++] = '1'; ++ else ++ s[z++] = '0'; ++ } ++ s[z] = '\0'; ++ return s; ++} ++ ++int main(int argc, char *argv[]) ++{ ++ int fd; ++ int x; ++ char fn[256]; ++ struct t4_regs regs; ++ if ((argc < 2) || ((*(argv[1]) != '/') && !atoi(argv[1]))) { ++ fprintf(stderr, "Usage: wct4xxp-diag <channel>\n"); ++ exit(1); ++ } ++ if (*(argv[1]) == '/') ++ dahdi_copy_string(fn, argv[1], sizeof(fn)); ++ else ++ snprintf(fn, sizeof(fn), "/dev/dahdi/%d", atoi(argv[1])); ++ fd = open(fn, O_RDWR); ++ if (fd <0) { ++ fprintf(stderr, "Unable to open '%s': %s\n", fn, strerror(errno)); ++ exit(1); ++ } ++ if (ioctl(fd, WCT4_GET_REGS, ®s)) { ++ fprintf(stderr, "Unable to get registers: %s\n", strerror(errno)); ++ exit(1); ++ } ++ printf("PCI Registers:\n"); ++ for (x=0;x<sizeof(xreginfo) / sizeof(xreginfo[0]);x++) { ++ fprintf(stdout, "%s (%02x): %08x (%s)\n", xreginfo[x].name, xreginfo[x].reg, regs.pci[xreginfo[x].reg], tobin32(regs.pci[xreginfo[x].reg])); ++ } ++ printf("\nE1 Control Registers:\n"); ++ for (x=0;x<sizeof(reginfo) / sizeof(reginfo[0]);x++) { ++ fprintf(stdout, "%s (%02x): %02x (%s)\n", reginfo[x].name, reginfo[x].reg, regs.regs[reginfo[x].reg], tobin(regs.regs[reginfo[x].reg])); ++ } ++ printf("\nE1 Status Registers:\n"); ++ for (x=0;x<sizeof(sreginfo) / sizeof(sreginfo[0]);x++) { ++ fprintf(stdout, "%s (%02x): %02x (%s)\n", sreginfo[x].name, sreginfo[x].reg, regs.regs[sreginfo[x].reg], tobin(regs.regs[sreginfo[x].reg])); ++ } ++ printf("\nT1 Control Registers:\n"); ++ for (x=0;x<sizeof(t1_reginfo) / sizeof(t1_reginfo[0]);x++) { ++ fprintf(stdout, "%s (%02x): %02x (%s)\n", t1_reginfo[x].name, t1_reginfo[x].reg, regs.regs[t1_reginfo[x].reg], tobin(regs.regs[t1_reginfo[x].reg])); ++ } ++ printf("\nT1 Status Registers:\n"); ++ for (x=0;x<sizeof(t1_sreginfo) / sizeof(t1_sreginfo[0]);x++) { ++ fprintf(stdout, "%s (%02x): %02x (%s)\n", t1_sreginfo[x].name, t1_sreginfo[x].reg, regs.regs[t1_sreginfo[x].reg], tobin(regs.regs[t1_sreginfo[x].reg])); ++ } ++ exit(0); ++} +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/wct4xxp.h dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/wct4xxp.h +--- dahdi-linux-2.10.0.1/drivers/dahdi/allo4xxp/wct4xxp.h 1970-01-01 01:00:00.000000000 +0100 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/allo4xxp/wct4xxp.h 2015-02-10 15:01:15.892795252 +0100 +@@ -0,0 +1,144 @@ ++/* ++ * Wildcard T400P FXS Interface Driver for DAHDI Telephony interface ++ * ++ * Written by Mark Spencer <markster@linux-support.net> ++ * ++ * Copyright (C) 2001-2010, Digium, Inc. ++ * ++ * All rights reserved. ++ * ++ */ ++ ++/* ++ * See http://www.asterisk.org for more information about ++ * the Asterisk project. Please do not directly contact ++ * any of the maintainers of this project for assistance; ++ * the project provides a web site, mailing lists and IRC ++ * channels for your use. ++ * ++ * This program is free software, distributed under the terms of ++ * the GNU General Public License Version 2 as published by the ++ * Free Software Foundation. See the LICENSE file included with ++ * this program for more details. ++ */ ++ ++#include <linux/ioctl.h> ++ ++#define FRMR_TTR_BASE 0x10 ++#define FRMR_RTR_BASE 0x0c ++#define FRMR_TSEO 0xa0 ++#define FRMR_TSBS1 0xa1 ++#define FRMR_CCR1 0x09 ++#define FRMR_CCR1_ITF 0x08 ++#define FRMR_CCR1_EITS 0x10 ++#define FRMR_CCR2 0x0a ++#define FRMR_CCR2_RCRC 0x04 ++#define FRMR_CCR2_RADD 0x10 ++#define FRMR_MODE 0x03 ++#define FRMR_MODE_NO_ADDR_CMP 0x80 ++#define FRMR_MODE_SS7 0x20 ++#define FRMR_MODE_HRAC 0x08 ++#define FRMR_IMR0 0x14 ++#define FRMR_IMR0_RME 0x80 ++#define FRMR_IMR0_RPF 0x01 ++#define FRMR_IMR1 0x15 ++#define FRMR_IMR1_ALLS 0x20 ++#define FRMR_IMR1_XDU 0x10 ++#define FRMR_IMR1_XPR 0x01 ++#define FRMR_XC0 0x22 ++#define FRMR_XC1 0x23 ++#define FRMR_RC0 0x24 ++#define FRMR_RC1 0x25 ++#define FRMR_SIC1 0x3e ++#define FRMR_SIC2 0x3f ++#define FRMR_SIC3 0x40 ++#define FRMR_CMR1 0x44 ++#define FRMR_CMR2 0x45 ++/* OctalFALC Only */ ++#define FRMR_CMR4 0x41 ++#define FRMR_CMR5 0x42 ++#define FRMR_CMR6 0x43 ++#define FRMR_GPC2 0x8a ++/* End Octal */ ++#define FRMR_GCR 0x46 ++#define FRMR_ISR0 0x68 ++#define FRMR_ISR0_RME 0x80 ++#define FRMR_ISR0_RPF 0x01 ++#define FRMR_ISR1 0x69 ++#define FRMR_ISR1_ALLS 0x20 ++#define FRMR_ISR1_XDU 0x10 ++#define FRMR_ISR1_XPR 0x01 ++#define FRMR_ISR2 0x6a ++#define FRMR_ISR3 0x6b ++#define FRMR_ISR4 0x6c ++#define FRMR_GIS 0x6e ++#define FRMR_GIS_ISR0 0x01 ++#define FRMR_GIS_ISR1 0x02 ++#define FRMR_GIS_ISR2 0x04 ++#define FRMR_GIS_ISR3 0x08 ++#define FRMR_GIS_ISR4 0x10 ++#define FRMR_CIS 0x6f ++#define FRMR_CIS_GIS1 0x01 ++#define FRMR_CIS_GIS2 0x02 ++#define FRMR_CIS_GIS3 0x04 ++#define FRMR_CIS_GIS4 0x08 ++ ++/* CIS - Octal falc bits */ ++#define FRMR_CIS_GIS5 0x10 ++#define FRMR_CIS_GIS6 0x20 ++#define FRMR_CIS_GIS7 0x40 ++#define FRMR_CIS_GIS8 0x80 ++ ++#define FRMR_CMDR 0x02 ++#define FRMR_CMDR_SRES 0x01 ++#define FRMR_CMDR_XRES 0x10 ++#define FRMR_CMDR_RMC 0x80 ++#define FRMR_CMDR_XTF 0x04 ++#define FRMR_CMDR_XHF 0x08 ++#define FRMR_CMDR_XME 0x02 ++#define FRMR_RSIS 0x65 ++#define FRMR_RSIS_VFR 0x80 ++#define FRMR_RSIS_RDO 0x40 ++#define FRMR_RSIS_CRC16 0x20 ++#define FRMR_RSIS_RAB 0x10 ++#define FRMR_RBCL 0x66 ++#define FRMR_RBCL_MAX_SIZE 0x1f ++#define FRMR_RBCH 0x67 ++#define FRMR_RXFIFO 0x00 ++#define FRMR_SIS 0x64 ++#define FRMR_SIS_XFW 0x40 ++#define FRMR_TXFIFO 0x00 ++ ++#define FRS0 0x4c ++#define FRS0_LOS (1<<7) ++#define FRS0_LFA (1<<5) ++#define FRS0_LMFA (1<<1) ++ ++#define FRS1 0x4d ++#define FRS1_XLS (1<<1) ++#define FRS1_XLO (1<<0) ++ ++#define NUM_REGS 0xa9 ++#define NUM_PCI 12 ++ ++struct t4_regs { ++ unsigned int pci[NUM_PCI]; ++ unsigned char regs[NUM_REGS]; ++}; ++ ++struct t4_reg { ++ unsigned int reg; ++ unsigned int val; ++}; ++ ++#define T4_CHECK_VPM 0 ++#define T4_LOADING_FW 1 ++#define T4_STOP_DMA 2 ++#define T4_CHECK_TIMING 3 ++#define T4_CHANGE_LATENCY 4 ++#define T4_IGNORE_LATENCY 5 ++ ++#define WCT4_GET_REGS _IOW(DAHDI_CODE, 60, struct t4_regs) ++#define WCT4_GET_REG _IOW(DAHDI_CODE, 61, struct t4_reg) ++#define WCT4_SET_REG _IOW(DAHDI_CODE, 62, struct t4_reg) ++ +diff -Nur dahdi-linux-2.10.0.1/drivers/dahdi/firmware/Makefile dahdi-linux-2.10.0.1-allo/drivers/dahdi/firmware/Makefile +--- dahdi-linux-2.10.0.1/drivers/dahdi/firmware/Makefile 2014-09-22 20:40:19.000000000 +0200 ++++ dahdi-linux-2.10.0.1-allo/drivers/dahdi/firmware/Makefile 2015-02-10 15:01:15.892795252 +0100 +@@ -41,10 +41,11 @@ + A4B_VERSION:=b0019 + + FIRMWARE_URL:=http://downloads.digium.com/pub/telephony/firmware/releases ++FIRMWARE_URL_ALLO:=http://www.allo.com/firmware/pri-card-second-gen + + ALL_FIRMWARE=FIRMWARE-OCT6114-032 FIRMWARE-OCT6114-064 FIRMWARE-OCT6114-128 FIRMWARE-OCT6114-256 + ALL_FIRMWARE+=FIRMWARE-TC400M FIRMWARE-HX8 FIRMWARE-VPMOCT032 FIRMWARE-TE820 FIRMWARE-TE133 FIRMWARE-TE134 +-ALL_FIRMWARE+=FIRMWARE-A8A FIRMWARE-A8B FIRMWARE-A4A FIRMWARE-A4B FIRMWARE-TE435 FIRMWARE-TE436 ++ALL_FIRMWARE+=FIRMWARE-A8A FIRMWARE-A8B FIRMWARE-A4A FIRMWARE-A4B FIRMWARE-TE435 FIRMWARE-TE436 FIRMWARE-allo-1PPCIe FIRMWARE-allo-2PPCIe FIRMWARE-allo-4PPCI FIRMWARE-allo-4PPCIe FIRMWARE-allo-8PPCIe + + # Firmware files should use the naming convention: dahdi-fw-<base name>-<sub name>-<version> or dahdi-fw-<base name>-<version> + # First example: dahdi-fw-oct6114-064-1.05.01 +@@ -69,6 +70,12 @@ + FIRMWARE:=$(FIRMWARE:FIRMWARE-A8B=dahdi-fw-a8a-$(A8A_VERSION).tar.gz) + FIRMWARE:=$(FIRMWARE:FIRMWARE-A4A=dahdi-fw-a4b-$(A4B_VERSION).tar.gz) + FIRMWARE:=$(FIRMWARE:FIRMWARE-A4B=dahdi-fw-a4a-$(A4A_VERSION).tar.gz) ++FIRMWARE:=$(FIRMWARE:FIRMWARE-allo-1PPCIe=allo-dahdi-fw-2aCP1e.tar.gz) ++FIRMWARE:=$(FIRMWARE:FIRMWARE-allo-2PPCIe=allo-dahdi-fw-2aCP2e.tar.gz) ++FIRMWARE:=$(FIRMWARE:FIRMWARE-allo-4PPCI=allo-dahdi-fw-2aCP4.tar.gz) ++FIRMWARE:=$(FIRMWARE:FIRMWARE-allo-4PPCIe=allo-dahdi-fw-2aCP4e.tar.gz) ++FIRMWARE:=$(FIRMWARE:FIRMWARE-allo-8PPCIe=allo-dahdi-fw-2aCP8e.tar.gz) ++ + + FWLOADERS:=dahdi-fwload-vpmadt032-$(VPMADT032_VERSION).tar.gz + +@@ -100,6 +107,9 @@ + rm -f dahdi-fw-*.tar.gz + rm -f dahdi-fwload-*.tar.gz + rm -f make_firmware_object ++ rm -f allo-dahdi-fw-*.bin ++ rm -f allo-dahdi-fw-*.tar.gz ++ + + # Clean up anything we built + clean: +@@ -113,6 +123,15 @@ + if test ! -f $@; then exit 1; fi; \ + fi + ++allo-dahdi-fw%.tar.gz: ++ @if ( [ "$(HOTPLUG_FIRMWARE)" = "no" ] ) || ( [ -d $(DESTDIR)/usr/lib/hotplug/firmware ] && ! [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.$(subst .tar.gz,,$*) ] ) || ( [ -d $(DESTDIR)/lib/firmware ] && ! [ -f $(DESTDIR)/lib/firmware/.$(subst .tar.gz,,$*) ] ); then \ ++ echo "Attempting to download $@"; \ ++ if test ! -f $@; then $(DOWNLOAD) $(WGET_ARGS) $(FIRMWARE_URL_ALLO)/$@; fi; \ ++ if test ! -f $@; then exit 1; fi; \ ++ (cat $@ | gzip -d | tar --no-same-owner -xf -) \ ++ fi ++ ++ + firmware-loaders: $(FWLOADERS) + + .PHONY: dahdi-fwload-vpmadt032-$(VPMADT032_VERSION).tar.gz +@@ -145,16 +164,90 @@ + @$(call RUN_INST,dahdi-fw-a8b,$(A8B_VERSION)) + @$(call RUN_INST,dahdi-fw-a4a,$(A4A_VERSION)) + @$(call RUN_INST,dahdi-fw-a4b,$(A4B_VERSION)) ++######################ALLO################################# ++ifeq ($(shell if ( [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP1e ] ) && ( [ -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP1e ] ); then echo "no"; else echo "yes"; fi),yes) ++ ++ @echo "Installing ALLO firmwares to hotplug firmware directories" ++ @install -m 644 allo-dahdi-fw-2aCP1e.bin $(DESTDIR)/usr/lib/hotplug/firmware ++ @rm -rf $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP1e ++ @touch $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP1e ++ @install -m 644 allo-dahdi-fw-2aCP1e.bin $(DESTDIR)/lib/firmware ++ @rm -rf $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP1e ++ @touch $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP1e ++else ++ @echo "Firmware of ALLO card allo-dahdi-fw-2aCP1e.bin is already installed with required version" ++endif ++ ++ ++ifeq ($(shell if ( [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP2e ] ) && ( [ -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP2e ] ); then echo "no"; else echo "yes"; fi),yes) ++ ++ @echo "Installing ALLO firmwares to hotplug firmware directories" ++ @install -m 644 allo-dahdi-fw-2aCP2e.bin $(DESTDIR)/usr/lib/hotplug/firmware ++ @rm -rf $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP2e ++ @touch $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP2e ++ @install -m 644 allo-dahdi-fw-2aCP2e.bin $(DESTDIR)/lib/firmware ++ @rm -rf $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP2e ++ @touch $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP2e ++else ++ @echo "Firmware of ALLO cards allo-dahdi-fw-2aCP2e.bin is already installed with required version" ++endif ++ ++ ++ifeq ($(shell if ( [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4 ] ) && ( [ -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4 ] ); then echo "no"; else echo "yes"; fi),yes) ++ ++ @echo "Installing ALLO firmwares to hotplug firmware directories" ++ @install -m 644 allo-dahdi-fw-2aCP4.bin $(DESTDIR)/usr/lib/hotplug/firmware ++ @rm -rf $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4 ++ @touch $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4 ++ @install -m 644 allo-dahdi-fw-2aCP4.bin $(DESTDIR)/lib/firmware ++ @rm -rf $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4 ++ @touch $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4 ++else ++ @echo "Firmware of ALLO cards allo-dahdi-fw-2aCP4.bin is already installed with required version" ++endif ++ ++ifeq ($(shell if ( [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4e ] ) && ( [ -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4e ] ); then echo "no"; else echo "yes"; fi),yes) ++ ++ @echo "Installing ALLO firmwares to hotplug firmware directories" ++ @install -m 644 allo-dahdi-fw-2aCP4e.bin $(DESTDIR)/usr/lib/hotplug/firmware ++ @rm -rf $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4e ++ @touch $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP4e ++ @install -m 644 allo-dahdi-fw-2aCP4e.bin $(DESTDIR)/lib/firmware ++ @rm -rf $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4e ++ @touch $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP4e ++else ++ @echo "Firmware of ALLO cards allo-dahdi-fw-2aCP4e.bin is already installed with required version" ++endif ++ ++ ++ifeq ($(shell if ( [ -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP8e ] ) && ( [ -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP8e ] ); then echo "no"; else echo "yes"; fi),yes) ++ ++ @echo "Installing ALLO firmwares to hotplug firmware directories" ++ @install -m 644 allo-dahdi-fw-2aCP8e.bin $(DESTDIR)/usr/lib/hotplug/firmware ++ @rm -rf $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP8e* ++ @touch $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw-2aCP8e ++ @install -m 644 allo-dahdi-fw-2aCP8e.bin $(DESTDIR)/lib/firmware ++ @rm -rf $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP8e* ++ @touch $(DESTDIR)/lib/firmware/.allo-dahdi-fw-2aCP8e ++else ++ @echo "Firmware of ALLO cards allo-dahdi-fw-2aCP8e.bin is already installed with required version" ++endif ++ ++#########################END###################################### + + # Uninstall any installed dahdi firmware images from hotplug firmware directories + hotplug-uninstall: + if [ -d $(DESTDIR)/usr/lib/hotplug/firmware ]; then \ + rm -f $(DESTDIR)/usr/lib/hotplug/firmware/dahdi-fw-*.bin; \ + rm -f $(DESTDIR)/usr/lib/hotplug/firmware/.dahdi-fw*; \ ++ rm -f $(DESTDIR)/usr/lib/hotplug/firmware/allo-dahdi-fw-*.bin; \ ++ rm -f $(DESTDIR)/usr/lib/hotplug/firmware/.allo-dahdi-fw*; \ + fi + if [ -d $(DESTDIR)/lib/firmware ]; then \ + rm -f $(DESTDIR)/lib/firmware/dahdi-fw-*.bin; \ + rm -f $(DESTDIR)/lib/firmware/.dahdi-fw*; \ ++ rm -f $(DESTDIR)/lib/firmware/allo-dahdi-fw-*.bin; \ ++ rm -f $(DESTDIR)/lib/firmware/.allo-dahdi-fw*; \ + fi + + make_firmware_object: make_firmware_object.in ../dahdi-base.o |