Kernel 5.13

xeno74 · Post by **xeno74** » Tue Apr 27, 2021 5:02 pm

Hi All,

The Linux 5.13 merge window is open.

Happy testing!

New:

Download: linux-image-5.13-alpha1-X1000_X5000.tar.gz

Please test the kernels.

Thanks,
Christian

kilaueabart · Post by **kilaueabart** » Tue Apr 27, 2021 8:25 pm

Hate to get my computer "-dirty" again, and 513 sounds kind of unlucky, but I'm giving it a go and so far so good.

xeno74 · Post by **xeno74** » Wed Apr 28, 2021 4:12 am

kilaueabart wrote: Tue Apr 27, 2021 8:25 pm Hate to get my computer "-dirty" again, and 513 sounds kind of unlucky, but I'm giving it a go and so far so good.

Thank you for testing!

xeno74 · Post by **xeno74** » Fri Apr 30, 2021 1:22 pm

Hi All,

I compiled the latest Git kernel today. Unfortunately this kernel doesn't boot on the X5000 and X1000 anymore. It boots in a virtual e5500 QEMU machine.

I bisected today.

Code: Select all
```
git bisect start
```

Code: Select all

git bisect good 9f4ad9e425a1d3b6a34617b8ea226d56a119a717

(Good: Linux 5.12)

Code: Select all
```
git bisect bad 9d31d2338950293ec19d9b095fbaa9030899dcb4
```
(Bad: Merge tag 'net-next-5.13' -- 2021-04-29 11:57:23)
git bisect good
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect bad
git bisect good
git bisect good
git bisect bad
git bisect bad

Result: 434fb1e7444a2efc3a4ebd950c7f771ebfcffa31 is the first bad commit.

Code: Select all

Author: Gustavo A. R. Silva <[email protected]>
Date:   Wed Feb 10 17:49:27 2021 -0600

    drm/radeon/nislands_smc.h: Replace one-element array with flexible-array member in struct NISLANDS_SMC_SWSTATE
    
    There is a regular need in the kernel to provide a way to declare having
    a dynamically sized set of trailing elements in a structure. Kernel code
    should always use “flexible array members”[1] for these cases. The older
    style of one-element or zero-length arrays should no longer be used[2].
    
    Use flexible-array member in struct NISLANDS_SMC_SWSTATE, instead of
    one-element array.
    
    Also, this helps with the ongoing efforts to enable -Warray-bounds by
    fixing the following warnings:
    
    drivers/gpu/drm/radeon/ni_dpm.c: In function ‘ni_convert_power_state_to_smc’:
    drivers/gpu/drm/radeon/ni_dpm.c:2521:20: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2521 |   smc_state->levels[i].dpm2.MaxPS =
          |   ~~~~~~~~~~~~~~~~~^~~
    drivers/gpu/drm/radeon/ni_dpm.c:2523:20: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2523 |   smc_state->levels[i].dpm2.NearTDPDec = NISLANDS_DPM2_NEAR_TDP_DEC;
          |   ~~~~~~~~~~~~~~~~~^~~
    drivers/gpu/drm/radeon/ni_dpm.c:2524:20: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2524 |   smc_state->levels[i].dpm2.AboveSafeInc = NISLANDS_DPM2_ABOVE_SAFE_INC;
          |   ~~~~~~~~~~~~~~~~~^~~
    drivers/gpu/drm/radeon/ni_dpm.c:2525:20: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2525 |   smc_state->levels[i].dpm2.BelowSafeInc = NISLANDS_DPM2_BELOW_SAFE_INC;
          |   ~~~~~~~~~~~~~~~~~^~~
    drivers/gpu/drm/radeon/ni_dpm.c:2526:35: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2526 |   smc_state->levels[i].stateFlags |=
          |                                   ^~
    drivers/gpu/drm/radeon/ni_dpm.c:2526:35: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2526 |   smc_state->levels[i].stateFlags |=
          |   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~
     2527 |    ((i != (state->performance_level_count - 1)) && power_boost_limit) ?
          |    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     2528 |    PPSMC_STATEFLAG_POWERBOOST : 0;
          |    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    drivers/gpu/drm/radeon/ni_dpm.c:2442:20: warning: array subscript 1 is above array bounds of ‘NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’ {aka ‘struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL[1]’} [-Warray-bounds]
     2442 |   smc_state->levels[i + 1].aT = cpu_to_be32(a_t);
    
    [1] https://en.wikipedia.org/wiki/Flexible_array_member
    [2] https://www.kernel.org/doc/html/v5.9/process/deprecated.html#zero-length-and-one-element-arrays
    
    Link: https://github.com/KSPP/linux/issues/79
    Link: https://github.com/KSPP/linux/issues/109
    Build-tested-by: kernel test robot <[email protected]>
    Link: https://lore.kernel.org/lkml/6023ed54.BfIY+9Uz81I6nq19%[email protected]/
    Signed-off-by: Gustavo A. R. Silva <[email protected]>
    Signed-off-by: Alex Deucher <[email protected]>

:040000 040000 825941c1622038715917db88900263261697ff43 44e16f7b0c4f28a6ca1fb7cc71a6d6a7a9db668f M	drivers

I was able to revert the first bad commit.

Code: Select all

git revert 434fb1e7444a2efc3a4ebd950c7f771ebfcffa31

Output:

Code: Select all

[master 437344379819] Revert "drm/radeon/nislands_smc.h: Replace one-element array with flexible-array member in struct NISLANDS_SMC_SWSTATE"
 1 file changed, 5 insertions(+), 5 deletions(-)

After a new compiling, the kernel boots without any problems.

I created a patch for reverting the bad commit:

Code: Select all

diff -rupN a/drivers/gpu/drm/radeon/nislands_smc.h b/drivers/gpu/drm/radeon/nislands_smc.h
--- a/drivers/gpu/drm/radeon/nislands_smc.h     2021-04-30 12:53:11.426435827 +0200
+++ b/drivers/gpu/drm/radeon/nislands_smc.h     2021-04-30 12:52:35.298525162 +0200
@@ -134,11 +134,11 @@ typedef struct NISLANDS_SMC_HW_PERFORMAN

 struct NISLANDS_SMC_SWSTATE
 {
-       uint8_t                             flags;
-       uint8_t                             levelCount;
-       uint8_t                             padding2;
-       uint8_t                             padding3;
-       NISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[];
+    uint8_t                             flags;
+    uint8_t                             levelCount;
+    uint8_t                             padding2;
+    uint8_t                             padding3;
+    NISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[1];
 };

 typedef struct NISLANDS_SMC_SWSTATE NISLANDS_SMC_SWSTATE;

The kernel compiles and boots with this patch on my X1000 and X5000.

Cheers,
Christian

xeno74 · Post by **xeno74** » Fri Apr 30, 2021 4:02 pm

Hi All,

I released the second alpha of kernel 5.13 for the X1000 and X5000 today.

New:

Download: linux-image-5.13-alpha2-X1000_X5000.tar.gz

Please test the kernels.

Thanks,
Christian

xeno74 · Post by **xeno74** » Fri Apr 30, 2021 4:35 pm

The same commit is also available for SI graphics cards.

Linux Git

Please also test the alpha2 with Radeon SI graphics cards. It could be, that your AmigaOne doesn't boot. I only patched the alpha2 for NI Radeons.

kilaueabart · Post by **kilaueabart** » Fri Apr 30, 2021 7:35 pm

I think maybe I have an SI card, but too senile to remember for sure, or how to find out. In any event, 5.13-a2 seems to be working as it should.

xeno74 · Post by **xeno74** » Sat May 01, 2021 11:17 am

kilaueabart wrote: Fri Apr 30, 2021 7:35 pm I think maybe I have an SI card, but too senile to remember for sure, or how to find out. In any event, 5.13-a2 seems to be working as it should.

Thanks a lot for testing the alpha2! You can figure out your graphics card with lspci.

@Skateman
Does the alpha2 boots on your X5000?

acefnq · Post by **acefnq** » Sat May 01, 2021 4:54 pm

5.13a2 boots and runs fine for me.

Thanks

ace

xeno74 · Post by **xeno74** » Sat May 01, 2021 5:58 pm

acefnq wrote: Sat May 01, 2021 4:54 pm 5.13a2 boots and runs fine for me.

Thanks

ace

Hi Ace,

Great! Thank you for testing!

Which graphics card is installed in your X5000?

Thanks,
Christian

Hyperion Entertainment Message Boards

Kernel 5.13

Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13

Re: Kernel 5.13