Arch Linux User Repository

I made a fork of this repo and fixed some bugs. Now this package builds from my fork.

Everything seems to work fine on 5.9.14-arch1-1.

There's a weird bug with device being down at start, I'm in the middle of investigation rn, but there's a workaround:

# systemctl edit NetworkManager

[Service]  
ExecStartPre=ip link set dev wlp4s0 up # replace wlp4s0 with your interface id

I made a fork of this repo and fixed some bugs. Now this package builds from my fork.

Everything seems to work fine on 5.9.14-arch1-1.

There's a weird bug with device being down at start, I'm in the middle of investigation rn, but there's a workaround:

# systemctl edit NetworkManager

[Service]  
ExecStartPre=ip link set dev wlp4s0 up # replace wlp4s0 with your interface id

I patched the GitHub repo below and got it to work with kernel >= 5 (since that repo already has fixes for kernel 4.15 timer changes), and my patch was approved. I would recommend anyone else trying to set up the rt5592sta driver use that repo instead of this package until further notice.

I am unable to install this package. Is anyone still maintaining this package?

First issue: parsing PKGBUILD

==> ERROR: conflicts contains invalid characters: ','

Easy to fix - remove the commas from the array on line 23 of the PKGBUILD

Second issue: during package install

make -j8 KERNELRELEASE=5.2.2-arch1-1-ARCH && cp os/linux/rt5592sta.ko ./rt5592sta.ko......(bad exit status: 2)
Error! Bad return status for module build on kernel: 5.2.2-arch1-1-ARCH (x86_64)
Consult /var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/make.log for more information.

The relevant error is:

/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/sta_ioctl.c: In function ‘rt28xx_sta_ioctl’:
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/sta_ioctl.c:2584:72: error: macro "access_ok" passed 3 arguments, but takes just 2
 2584 |     if ( access_ok(VERIFY_WRITE, wrqin->u.data.pointer, sizeof(privtab)) != TRUE)
      |                                                                        ^
./arch/x86/include/asm/uaccess.h:91: note: macro "access_ok" defined here
   91 | #define access_ok(addr, size)     \
      | 
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/sta_ioctl.c:2584:10: note: each undeclared identifier is reported only once for each function it appears in
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/sta_ioctl.c:2597:73: error: macro "access_ok" passed 3 arguments, but takes just 2
 2597 |    if(access_ok(VERIFY_READ, wrqin->u.data.pointer, wrqin->u.data.length) != TRUE)
      |                                                                         ^
./arch/x86/include/asm/uaccess.h:91: note: macro "access_ok" defined here
   91 | #define access_ok(addr, size)     \
      |

There is a simple fix to remove the first parameter from ACCESS_OK (see this zfs issue). But we have additional errors:

/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/rt_linux.c: In function ‘__RTMP_OS_Init_Timer’:
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/rt_linux.c:122:3: error: implicit declaration of function ‘init_timer’; did you mean ‘init_timers’? [-Werror=implicit-function-declaration]
  122 |   init_timer(pTimer);
      |   ^~~~~~~~~~
      |   init_timers
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/rt_linux.c:123:9: error: ‘OS_NDIS_MINIPORT_TIMER’ {aka ‘struct timer_list’} has no member named ‘data’
  123 |   pTimer->data = (unsigned long)data;
      |         ^~
/var/lib/dkms/rt5592sta_linux_patched-dkms/2.6.0.0/build/os/linux/../../os/linux/rt_linux.c:124:20: error: assignment to ‘void (*)(struct timer_list *)’ from incompatible pointer type ‘TIMER_FUNCTION’ {aka ‘void (*)(long unsigned int)’} [-Werror=incompatible-pointer-types]
  124 |   pTimer->function = function;
      |                    ^

It looks like the timer interface changed in kernel 4.15 and there are some similar fixes here n another realtek driver.

Looks like another repo with this driver has some fixes for these issues too.

@tomsik68 even tried downgrading to 4.7.1-1-ARCH but still cannot install https://pastebin.com/TNi0cJ5S

@Bmueller thanks for notifying me. I was able to come up with a patch to fix the building issue. I also tested the driver under kernel 4.7.1-1-ARCH with my PCI card and it seems to work fine again.

Package fails to build on last few kernels- currently trying to install on 4.7.0-1-ARCH (x86_64) and getting "DBGPRINT(RT_DEBUG_ERROR, ("AES_Key_Wrap: allocate %d bytes memory failure.\n", sizeof(UINT8)*PlainTextLength));" followed by several other error messages, all seeming to pertain to memory allocation failure of some sort. Error log posted here -->(http://pastebin.com/0GWh7btA), anyone have any idea what's going on?

Changelog for pkgrel 3: + Split up depends array without using IF to check for architecture

Changelog for package release 2: + Added compatibility with 32-bit systems + Fixed bug where a file required by driver wasn't installed