summarylogtreecommitdiffstats
path: root/vmmon.patch
blob: 98a26413975985754be699ccabe56fb93ab5eb27 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
--- a/vmmon/Makefile
+++ b/vmmon/Makefile
@@ -43,7 +43,11 @@
 endif
 
 
+ifdef KVERSION
+VM_UNAME = $(KVERSION)
+else
 VM_UNAME = $(shell uname -r)
+endif
 
 # Header directory for the running kernel
 ifdef LINUXINCLUDE
@@ -98,6 +102,13 @@ auto-build: $(DRIVER_KO)
 $(DRIVER): $(DRIVER_KO)
 	if [ $< -nt $@ ] || [ ! -e $@ ] ; then cp -f $< $@; fi
 
+# Use SUBDIRS on 2.x, 3.x, 4.x.  Use M on newer kernels.
+ifeq ($(filter-out 2 3 4,$(firstword $(subst ., ,$(VM_UNAME)))),)
+DIRVAR := SUBDIRS
+else
+DIRVAR := M
+endif
+
 #
 # Define a setup target that gets built before the actual driver.
 # This target may not be used at all, but if it is then it will be defined
@@ -107,7 +118,7 @@ prebuild:: ;
 postbuild:: ;
 
 $(DRIVER_KO): prebuild
-	$(MAKE) -C $(BUILD_DIR) SUBDIRS=$$PWD SRCROOT=$$PWD/$(SRCROOT) \
+	$(MAKE) -C $(BUILD_DIR) $(DIRVAR)=$$PWD SRCROOT=$$PWD/$(SRCROOT) \
 	  MODULEBUILDDIR=$(MODULEBUILDDIR) modules
 	$(MAKE) -C $$PWD SRCROOT=$$PWD/$(SRCROOT) \
 	  MODULEBUILDDIR=$(MODULEBUILDDIR) postbuild
--- a/vmmon/include/pgtbl.h
+++ b/vmmon/include/pgtbl.h
@@ -91,7 +91,11 @@
          } else {
             pte_t *pte;
 
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(6,5,0)
+            pte = pte_offset_kernel(pmd, addr);
+#else
             pte = pte_offset_map(pmd, addr);
+#endif
             if (pte_present(*pte) == 0) {
                pte_unmap(pte);
                return INVALID_MPN;
--- a/vmmon/linux/driver.c
+++ b/vmmon/linux/driver.c
@@ -96,7 +96,9 @@ long LinuxDriver_Ioctl(struct file *filp
                        unsigned long ioarg);
 
 static int LinuxDriver_Close(struct inode *inode, struct file *filp);
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0)
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 17, 0)
+static vm_fault_t LinuxDriverFault(struct vm_fault *fault);
+#elif LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0)
 static int LinuxDriverFault(struct vm_fault *fault);
 #else
 static int LinuxDriverFault(struct vm_area_struct *vma, struct vm_fault *fault);
@@ -259,7 +261,7 @@
 /*
  *----------------------------------------------------------------------
  *
- * init_module --
+ * LinuxDriverInit --
  *
  *      linux module entry point. Called by /sbin/insmod command
  *
@@ -272,7 +274,7 @@
  */
 
 int
-init_module(void)
+LinuxDriverInit(void)
 {
    int retval;
 
@@ -351,7 +353,7 @@
 /*
  *----------------------------------------------------------------------
  *
- * cleanup_module --
+ * LinuxDriverExit --
  *
  *      Called by /sbin/rmmod
  *
@@ -360,7 +362,7 @@
  */
 
 void
-cleanup_module(void)
+LinuxDriverExit(void)
 {
    /*
     * XXX smp race?
@@ -594,7 +596,12 @@ LinuxDriver_Close(struct inode *inode, /
  *-----------------------------------------------------------------------------
  */
 
-static int
+static
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 17, 0)
+vm_fault_t
+#else
+int
+#endif
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0)
 LinuxDriverFault(struct vm_fault *fault)     //IN/OUT
 #else
@@ -761,7 +768,11 @@ LinuxDriverMmap(struct file *filp,
       return err;
    }
    /* Clear VM_IO, otherwise SuSE's kernels refuse to do get_user_pages */
+#if COMPAT_LINUX_VERSION_CHECK_LT(6, 3, 0)
    vma->vm_flags &= ~VM_IO;
+#else
+   vm_flags_clear(vma, VM_IO);
+#endif
 
    return 0;
 }
@@ -1751,3 +1762,5 @@
  * by default (i.e., neither mkinitrd nor modprobe will accept it).
  */
 MODULE_INFO(supported, "external");
+module_init(LinuxDriverInit);
+module_exit(LinuxDriverExit);
--- a/vmmon/linux/hostif.c
+++ a/vmmon/linux/hostif.c
@@ -47,6 +47,8 @@
 #include <asm/asm.h>
 #include <asm/io.h>
 #include <asm/page.h>
+#include <asm/tlbflush.h>
+#include <asm/irq_vectors.h>
 #include <asm/uaccess.h>
 #include <linux/capability.h>
 #include <linux/kthread.h>
@@ -54,6 +56,7 @@
 #include <linux/hrtimer.h>
 #include <linux/signal.h>
 #include <linux/taskstats_kern.h> // For linux/sched/signal.h without version check
+#include <linux/eventfd.h>
 
 #include "vmware.h"
 #include "x86apic.h"
@@ -79,6 +82,13 @@
 #error CONFIG_HIGH_RES_TIMERS required for acceptable performance
 #endif
 
+/* task's state is read-once rather than volatile from 5.14-rc2. */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 15, 0) || defined(get_current_state)
+#define get_task_state(task) READ_ONCE((task)->__state)
+#else
+#define get_task_state(task) ((task)->state)
+#endif
+
 /*
  * Although this is not really related to kernel-compatibility, I put this
  * helper macro here for now for a lack of better place --hpreg
@@ -442,7 +452,7 @@
    while ((vcpuid = VCPUSet_FindFirst(&req)) != VCPUID_INVALID) {
       struct task_struct *t = vm->vmhost->vcpuSemaTask[vcpuid];
       VCPUSet_Remove(&req, vcpuid);
-      if (t && (t->state & TASK_INTERRUPTIBLE)) {
+      if (t && (get_task_state(t) & TASK_INTERRUPTIBLE)) {
          wake_up_process(t);
       }
    }
@@ -593,6 +596,15 @@ HostIF_FastClockUnlock(int callerID) //
    MutexUnlock(&fastClockMutex, callerID);
 }
 
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 8, 0)
+static int crosspage_set_exec(pte_t *pte, unsigned long addr, void *data)
+{
+	struct page *p = data;
+
+	set_pte(pte, mk_pte(p, VM_PAGE_KERNEL_EXEC));
+	return 0;
+}
+#endif
 
 /*
  *----------------------------------------------------------------------
@@ -613,7 +625,29 @@ HostIF_FastClockUnlock(int callerID) //
 static void *
 MapCrossPage(struct page *p)  // IN:
 {
+#if COMPAT_LINUX_VERSION_CHECK_LT(5, 8, 0)
    return vmap(&p, 1, VM_MAP, VM_PAGE_KERNEL_EXEC);
+#else
+   void *addr;
+
+   addr = vmap(&p, 1, VM_MAP, VM_PAGE_KERNEL_EXEC);
+   if (!addr)
+	   return NULL;
+
+   /* Starting with 5.8, vmap() always sets the NX bit, but the cross
+    * page needs to be executable. */
+   if (apply_to_page_range(current->mm, (unsigned long)addr, PAGE_SIZE,
+			   crosspage_set_exec, p)) {
+      vunmap(addr);
+      return NULL;
+   }
+
+   preempt_disable();
+   __flush_tlb_all();
+   preempt_enable();
+
+   return addr;
+#endif
 }
 
 
@@ -1499,9 +1533,13 @@ HostIF_EstimateLockedPageLimit(const VMD
     * since at least 2.6.0.
     */
 
+#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 0, 0)
    extern unsigned long totalram_pages;
 
    unsigned int totalPhysicalPages = totalram_pages;
+#else
+   unsigned int totalPhysicalPages = totalram_pages();
+#endif
 
    /*
     * Use the memory information linux exports as of late for a more
@@ -1527,7 +1565,10 @@ HostIF_EstimateLockedPageLimit(const VMD
    lockedPages += global_page_state(NR_PAGETABLE);
 #endif
    /* NR_SLAB_* moved from zone to node in 4.13. */
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 13, 0)
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 9, 0)
+   lockedPages += global_node_page_state_pages(NR_SLAB_UNRECLAIMABLE_B);
+#elif LINUX_VERSION_CODE >= KERNEL_VERSION(4, 13, 0)
    lockedPages += global_node_page_state(NR_SLAB_UNRECLAIMABLE);
 #else
    lockedPages += global_page_state(NR_SLAB_UNRECLAIMABLE);
@@ -1602,6 +1643,49 @@ HostIF_WaitForFreePages(unsigned int tim
 /*
  *----------------------------------------------------------------------
  *
+ * HostIFGetTime --
+ *
+ *      Reads the current time in UPTIME_FREQ units.
+ *
+ * Results:
+ *      The uptime, in units of UPTIME_FREQ.
+ *
+ * Side effects:
+ *      None.
+ *
+ *----------------------------------------------------------------------
+ */
+
+static uint64
+HostIFGetTime(void)
+{
+#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 0, 0)
+   struct timeval tv;
+
+   do_gettimeofday(&tv);
+   return tv.tv_usec * (UPTIME_FREQ / 1000000) + tv.tv_sec * UPTIME_FREQ;
+#else
+   struct timespec64 now;
+
+   /*
+    * Use raw time used by Posix timers.  This time is not affected by
+    * NTP adjustments, so it may drift from real time and monotonic time,
+    * but it will stay in sync with other timers.
+    */
+   ktime_get_raw_ts64(&now);
+   /*
+    * UPTIME_FREQ resolution is lower than tv_nsec,
+    * so we have to do division...
+    */
+   ASSERT_ON_COMPILE(1000000000 % UPTIME_FREQ == 0);
+   return now.tv_nsec / (1000000000 / UPTIME_FREQ) + now.tv_sec * UPTIME_FREQ;
+#endif
+}
+
+
+/*
+ *----------------------------------------------------------------------
+ *
  * HostIFReadUptimeWork --
  *
  *      Reads the current uptime.  The uptime is based on getimeofday,
@@ -1630,7 +1714,6 @@ HostIF_WaitForFreePages(unsigned int tim
 static uint64
 HostIFReadUptimeWork(unsigned long *j)  // OUT: current jiffies
 {
-   struct timeval tv;
    uint64 monotime, uptime, upBase, monoBase;
    int64 diff;
    uint32 version;
@@ -1645,13 +1728,12 @@ HostIFReadUptimeWork(unsigned long *j)
       monoBase = uptimeState.monotimeBase;
    } while (!VersionedAtomic_EndTryRead(&uptimeState.version, version));
 
-   do_gettimeofday(&tv);
+   uptime = HostIFGetTime();
    upBase = Atomic_Read64(&uptimeState.uptimeBase);
 
    monotime = (uint64)(jifs - jifBase) * (UPTIME_FREQ / HZ);
    monotime += monoBase;
 
-   uptime = tv.tv_usec * (UPTIME_FREQ / 1000000) + tv.tv_sec * UPTIME_FREQ;
    uptime += upBase;
 
    /*
@@ -1756,13 +1838,11 @@ HostIFUptimeResyncMono(struct timer_list
 void
 HostIF_InitUptime(void)
 {
-   struct timeval tv;
+   uint64 tm;
 
    uptimeState.jiffiesBase = jiffies;
-   do_gettimeofday(&tv);
-   Atomic_Write64(&uptimeState.uptimeBase,
-                  -(tv.tv_usec * (UPTIME_FREQ / 1000000) +
-                    tv.tv_sec * UPTIME_FREQ));
+   tm = HostIFGetTime();
+   Atomic_Write64(&uptimeState.uptimeBase, -tm);
 
 #if LINUX_VERSION_CODE < KERNEL_VERSION(4, 15, 0) && !defined(timer_setup)
    init_timer(&uptimeState.timer);
@@ -2159,16 +2239,26 @@ HostIF_VMLockIsHeld(VMDriver *vm) // IN
 static Bool
 isVAReadable(VA r)  // IN:
 {
-   mm_segment_t old_fs;
    uint32 dummy;
    int ret;
 
+#if defined(HAVE_GET_KERNEL_NOFAULT) || LINUX_VERSION_CODE >= KERNEL_VERSION(5, 17, 0)
+   /*
+    * Exists from 5.10, first indicated by HAVE_GET_KERNEL_NOFAULT,
+    * and from post-5.17 just existing everywhere.
+    */
+   ret = get_kernel_nofault(dummy, (void *)r);
+#else
+   {
+   mm_segment_t old_fs;
+ 
    old_fs = get_fs();
-   set_fs(get_ds());
+   set_fs(KERNEL_DS);
    r = APICR_TO_ADDR(r, APICR_VERSION);
    ret = HostIF_CopyFromUser(&dummy, r, sizeof dummy);
    set_fs(old_fs);
-
+   }
+#endif
    return ret == 0;
 }
 
@@ -2197,7 +2283,7 @@ SetVMAPICAddr(VMDriver *vm, // IN/OUT: d
    volatile void *hostapic;
 
    ASSERT_ON_COMPILE(APICR_SIZE <= PAGE_SIZE);
-   hostapic = (volatile void *) ioremap_nocache(ma, PAGE_SIZE);
+   hostapic = (volatile void *) ioremap(ma, PAGE_SIZE);
    if (hostapic) {
       if ((APIC_VERSIONREG(hostapic) & 0xF0) == 0x10) {
          vm->hostAPIC.base = (volatile uint32 (*)[4]) hostapic;
@@ -2353,7 +2439,6 @@ HostIF_SemaphoreWait(VMDriver *vm,   //
                      uint64 *args)   // IN:
 {
    struct file *file;
-   mm_segment_t old_fs;
    int res;
    int waitFD = args[0];
    int timeoutms = args[2];
@@ -2364,22 +2449,19 @@ HostIF_SemaphoreWait(VMDriver *vm,   //
       return MX_WAITERROR;
    }
 
-   old_fs = get_fs();
-   set_fs(get_ds());
-
    {
       struct poll_wqueues table;
       unsigned int mask;
 
       poll_initwait(&table);
-      current->state = TASK_INTERRUPTIBLE;
+      __set_current_state(TASK_INTERRUPTIBLE);
       mask = file->f_op->poll(file, &table.pt);
       if (!(mask & (POLLIN | POLLERR | POLLHUP))) {
          vm->vmhost->vcpuSemaTask[vcpuid] = current;
          schedule_timeout(timeoutms * HZ / 1000);  // convert to Hz
          vm->vmhost->vcpuSemaTask[vcpuid] = NULL;
       }
-      current->state = TASK_RUNNING;
+      __set_current_state(TASK_RUNNING);
       poll_freewait(&table);
    }
 
@@ -2388,9 +2470,11 @@ HostIF_SemaphoreWait(VMDriver *vm,   //
     * the code to happily deal with a pipe or an eventfd. We only care about
     * reading no bytes (EAGAIN - non blocking fd) or sizeof(uint64).
     */
-
-   res = file->f_op->read(file, (char *) &value, sizeof value, &file->f_pos);
-
+#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 14, 0)
+   res = kernel_read(file, file->f_pos, (char *)&value, sizeof value);
+#else
+   res = kernel_read(file, &value, sizeof value, &file->f_pos);
+#endif
    if (res == sizeof value) {
       res = MX_WAITNORMAL;
    } else {
@@ -2399,7 +2483,6 @@ HostIF_SemaphoreWait(VMDriver *vm,   //
       }
    }
 
-   set_fs(old_fs);
    fput(file);
 
    /*
@@ -2452,7 +2542,7 @@
    FOR_EACH_VCPU_IN_SET(vcs, vcpuid) {
       struct task_struct *t = vm->vmhost->vcpuSemaTask[vcpuid];
       vm->vmhost->vcpuSemaTask[vcpuid] = NULL;
-      if (t && (t->state & TASK_INTERRUPTIBLE)) {
+      if (t && (get_task_state(t) & TASK_INTERRUPTIBLE)) {
          wake_up_process(t);
       }
    } ROF_EACH_VCPU_IN_SET();
@@ -2482,8 +2565,8 @@ HostIF_SemaphoreForceWakeup(VMDriver *vm
 int
 HostIF_SemaphoreSignal(uint64 *args)  // IN:
 {
+   struct eventfd_ctx *eventfd;
    struct file *file;
-   mm_segment_t old_fs;
    int res;
    int signalFD = args[1];
    uint64 value = 1;  // make an eventfd happy should it be there
@@ -2493,22 +2576,36 @@ HostIF_SemaphoreSignal(uint64 *args)  //
       return MX_WAITERROR;
    }
 
-   old_fs = get_fs();
-   set_fs(get_ds());
+   /*
+    * If it's eventfd, use specific eventfd interface as kernel writes
+    * to eventfd may not be allowed in kernel 5.10 and later.
+    */
+   eventfd = eventfd_ctx_fileget(file);
+   if (!IS_ERR(eventfd)) {
+#if COMPAT_LINUX_VERSION_CHECK_LT(6, 8, 0)
+      eventfd_signal(eventfd, 1);
+#else
+      eventfd_signal(eventfd);
+#endif
+      fput(file);
+      return MX_WAITNORMAL;
+   }
 
    /*
     * Always write sizeof(uint64) bytes. This works fine for eventfd and
     * pipes. The data written is formatted to make an eventfd happy should
     * it be present.
     */
-
-   res = file->f_op->write(file, (char *) &value, sizeof value, &file->f_pos);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 14, 0)
+   res = kernel_write(file, (char *)&value, sizeof value, file->f_pos);
+#else
+   res = kernel_write(file, &value, sizeof value, &file->f_pos);
+#endif
 
    if (res == sizeof value) {
       res = MX_WAITNORMAL;
    }
 
-   set_fs(old_fs);
    fput(file);
 
    /*
@@ -3027,12 +3120,9 @@ static int
 HostIFFastClockThread(void *unused)  // IN:
 {
    int res;
-   mm_segment_t oldFS;
    unsigned int rate = 0;
    unsigned int prevRate = 0;
 
-   oldFS = get_fs();
-   set_fs(KERNEL_DS);
    allow_signal(SIGKILL);
 
    while ((rate = linuxState.fastClockRate) > MIN_RATE) {
@@ -3055,8 +3145,6 @@ HostIFFastClockThread(void *unused)  //
    }
 
  out:
-   set_fs(oldFS);
-
    /*
     * Do not exit thread until we are told to do so.
     */
@@ -3154,7 +3242,6 @@ HostIF_SetFastClockRate(unsigned int rat
       }
    } else {
       if (linuxState.fastClockThread) {
-         force_sig(SIGKILL, linuxState.fastClockThread);
          kthread_stop(linuxState.fastClockThread);
 
          linuxState.fastClockThread = NULL;
@@ -3200,7 +3287,12 @@ HostIF_MapUserMem(VA addr,
 
    ASSERT(handle);
 
-   if (!access_ok(VERIFY_WRITE, p, size)) {
+#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 0, 0)
+   if (!access_ok(VERIFY_WRITE, p, size))
+#else
+   if (!access_ok(p, size))
+#endif
+   {
       printk(KERN_ERR "%s: Couldn't verify write to uva 0x%p with size %"
              FMTSZ"u\n", __func__, p, size);
 
From 4c2a103fd2d71f2084f1fe7ceacb816b9832ffa2 Mon Sep 17 00:00:00 2001
From: Michal Kubecek <mkubecek@suse.cz>
Date: Sun, 22 Oct 2023 23:24:05 +0200
Subject: [PATCH] vmmon: use get_user_pages to get page PFN

As a side effect of mainline commit 0d940a9b270b ("mm/pgtable: allow
pte_offset_map[_lock]() to fail") in 6.5-rc1, __pte_offset_map(), called by
pte_offset_map(), is no longer exported. WMware developers decided to hack
around this by replacing pte_offset_map() by pte_offset_kernel() which does
not seem to be a good idea and apparently may trigger warn checks in RCU
code on some systems as mentioned in the discussion on issue #223.
Therefore let's use the same solution as we had for 17.0.2 and older
versions as it does not show these problems.

Based on an upstream IRC discussion and the hva_to_pfn_*() family of
functions in KVM code, what PgtblVa2MPNLocked() does seems to be an
incomplete and partial open coded logic of get_user_pages() and as it is
only used to get PFN from a virtual address, it can be easily implemented
using get_user_pages() family.

Without knowledge what exactly are the PFNs used for in VMware, it is hard
to guess the right flags, these seem to work and have been tested by
multiple users over last few weeks.

We could likely use get_user_pages() also on older kernels and it might be
actually cleaner and more reliable as existing open coded implementation
does not seem to handle some corner cases but without knowledge of VMware
internals, it will be safer to stick to existing code where possible.
---
 vmmon-only/include/pgtbl.h | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/vmmon-only/include/pgtbl.h b/vmmon-only/include/pgtbl.h
index 3f43c62..7eaa49a 100644
--- a/vmmon-only/include/pgtbl.h
+++ b/vmmon-only/include/pgtbl.h
@@ -25,6 +25,7 @@
 #include "compat_pgtable.h"
 #include "compat_spinlock.h"
 #include "compat_page.h"
+#include "compat_version.h"
 
 
 /*
@@ -45,6 +46,7 @@
  *-----------------------------------------------------------------------------
  */
 
+#if COMPAT_LINUX_VERSION_CHECK_LT(6, 5, 0) // only used by PgtblVa2MPN() below
 static INLINE MPN
 PgtblVa2MPNLocked(struct mm_struct *mm, // IN: Mm structure of a process
                   VA addr)              // IN: Address in the virtual address
@@ -110,6 +112,7 @@ PgtblVa2MPNLocked(struct mm_struct *mm, // IN: Mm structure of a process
    }
    return mpn;
 }
+#endif
 
 
 /*
@@ -129,6 +132,8 @@ PgtblVa2MPNLocked(struct mm_struct *mm, // IN: Mm structure of a process
  *-----------------------------------------------------------------------------
  */
 
+#if COMPAT_LINUX_VERSION_CHECK_LT(6, 5, 0)
+
 static INLINE MPN
 PgtblVa2MPN(VA addr)  // IN
 {
@@ -143,4 +148,24 @@ PgtblVa2MPN(VA addr)  // IN
    return mpn;
 }
 
+#else /* COMPAT_LINUX_VERSION_CHECK_LT(6, 5, 0) */
+
+static INLINE MPN
+PgtblVa2MPN(VA addr)  // IN
+{
+   struct page *page;
+   int npages;
+   MPN mpn;
+
+   npages = get_user_pages_unlocked(addr, 1, &page, FOLL_HWPOISON);
+   if (npages != 1)
+	   return INVALID_MPN;
+   mpn = page_to_pfn(page);
+   put_page(page);
+
+   return mpn;
+}
+
+#endif /* COMPAT_LINUX_VERSION_CHECK_LT(6, 5, 0) */
+
 #endif /* __PGTBL_H__ */
From add7a6d8b99565fdfa79098315b0a69fd244eda3 Mon Sep 17 00:00:00 2001
From: Michal Kubecek <mkubecek@suse.cz>
Date: Fri, 12 Jan 2024 08:30:33 +0100
Subject: [PATCH] modules: fix build with -Wmissing-prototypes

Mainline commit 0fcb70851fbf ("Makefile.extrawarn: turn on
missing-prototypes globally") in 6.8-rc1 enables -Wmissing-prototypes
globally, revealing a lot of unclean code and also some actual problems.
This is also the case in vmmon and vmnet modules.

Most of them are addressed by making functions used only within one file
static. A special case are Vmx86_MapPage() and Vmx86_UnmapPage() which were
defined since Workstation 14.0.0 but they are not actually used anywhere
until 15.0.0 so that dropping them seems to be the best option.

The missing prototype of random_get_entropy_fallback() is handled by
including <linux/timex.h> rather than <asm/timex.h>.

Finally, there are four functions in vmnet module which are actually used
in multiple files but instead of proper declarations, their prototype is
duplicated in vmnet-only/driver.c, risking that the two copies won't match
(which actually happened in one case). The cleanest solution would be
creating separate header files for them (bridge.h, netif.h, userif.h and
vnetUserListener.h) and including them in the respective source file and
driver.c. As the developers already handle similar cases by simply putting
the declarations into vnetInt.h, let us do the same to keep things simple.
---
 vmmon-only/bootstrap/monLoaderVmmon.c |  1 +
 vmmon-only/common/vmx86.c             | 50 +--------------------------
 vmmon-only/linux/driver.c             |  4 +--
 vmmon-only/linux/hostif.c             |  6 ++--
 vmnet-only/bridge.c                   |  2 +-
 vmnet-only/driver.c                   | 16 ++-------
 vmnet-only/vnetInt.h                  |  7 ++++
 7 files changed, 17 insertions(+), 69 deletions(-)

diff --git a/vmmon-only/bootstrap/monLoaderVmmon.c b/vmmon-only/bootstrap/monLoaderVmmon.c
index 6879e617..55bcd2c3 100644
--- a/vmmon-only/bootstrap/monLoaderVmmon.c
+++ b/vmmon-only/bootstrap/monLoaderVmmon.c
@@ -37,6 +37,7 @@
 #include "x86paging_common.h"
 #include "x86paging_64.h"
 #include "vmx86.h"
+#include "monLoader.h"
 #include "monLoaderLog.h"
 
 typedef struct MonLoaderEnvContext {
diff --git a/vmmon-only/common/vmx86.c b/vmmon-only/common/vmx86.c
index f6b0d6fc..69455206 100644
--- a/vmmon-only/common/vmx86.c
+++ b/vmmon-only/common/vmx86.c
@@ -58,7 +58,7 @@
 #include "x86svm.h"
 #include "x86cpuid_asm.h"
 #if defined(linux)
-#include <asm/timex.h>
+#include <linux/timex.h>
 #endif
 #include "x86perfctr.h"
 #include "x86vtinstr.h"
@@ -3066,51 +3066,3 @@ Vmx86_GetPageRoot(VMDriver *vm,  // IN:
    *mpn = vm->ptRootMpns[vcpuid];
    return TRUE;
 }
-
-
-/*
- *----------------------------------------------------------------------
- *
- * Vmx86_MapPage --
- *
- *      Maps the specified MPN into the host kernel address space.
- *      returns the VPN of the mapping.
- *
- * Results:
- *      The VPN in the kernel address space of the new mapping, or 0 if
- *      the mapping failed.
- *
- * Side effects:
- *      None.
- *
- *----------------------------------------------------------------------
- */
-
-VPN
-Vmx86_MapPage(MPN mpn) // IN:
-{
-   return HostIF_MapPage(mpn);
-}
-
-
-/*
- *----------------------------------------------------------------------
- *
- * Vmx86_UnmapPage --
- *
- *      Unmaps the specified VPN from the host kernel address space.
- *
- * Results:
- *      None.
- *
- * Side effects:
- *      None.
- *
- *----------------------------------------------------------------------
- */
-
-void
-Vmx86_UnmapPage(VPN vpn) // IN:
-{
-   HostIF_UnmapPage(vpn);
-}
diff --git a/vmmon-only/linux/driver.c b/vmmon-only/linux/driver.c
index f015eb0b..337b1cf8 100644
--- a/vmmon-only/linux/driver.c
+++ b/vmmon-only/linux/driver.c
@@ -273,7 +273,7 @@ LinuxDriverInitTSCkHz(void)
  *----------------------------------------------------------------------
  */
 
-int
+static int
 LinuxDriverInit(void)
 {
    int retval;
@@ -361,7 +361,7 @@ LinuxDriverInit(void)
  *----------------------------------------------------------------------
  */
 
-void
+static void
 LinuxDriverExit(void)
 {
    /*
diff --git a/vmmon-only/linux/hostif.c b/vmmon-only/linux/hostif.c
index cac16735..f33c442a 100644
--- a/vmmon-only/linux/hostif.c
+++ b/vmmon-only/linux/hostif.c
@@ -2818,7 +2818,7 @@ HostIF_CallOnEachCPU(void (*func)(void*), // IN: function to call
  *-----------------------------------------------------------------------------
  */
 
-Bool
+static Bool
 HostIFCheckTrackedMPN(VMDriver *vm, // IN: The VM instance
                       MPN mpn)      // IN: The MPN
 {
@@ -2946,7 +2946,7 @@ HostIF_ReadPhysical(VMDriver *vm,      // IN: The VM instance
  *----------------------------------------------------------------------
  */
 
-int
+static int
 HostIFWritePhysicalWork(MA ma,             // MA to be written to
                         VA64 addr,         // src data to write
                         Bool kernelBuffer, // is the buffer in kernel space?
@@ -3085,7 +3085,7 @@ HostIF_GetCurrentPCPU(void)
  *----------------------------------------------------------------------
  */
 
-int
+static int
 HostIFStartTimer(Bool rateChanged,  //IN: Did rate change?
                  unsigned int rate) //IN: current clock rate
 {