On 8/12/24 14:49, Guenter Roeck wrote:
Hi,
On 8/8/24 02:11, Greg Kroah-Hartman wrote:
This is the start of the stable review cycle for the 6.1.104 release. There are 86 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Sat, 10 Aug 2024 09:11:02 +0000. Anything received after that time might be too late.
The whole patch series can be found in one patch at: https://www.kernel.org/pub/linux/kernel/v6.x/stable-review/patch-6.1.104-rc2... or in the git tree and branch at: git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-6.1.y and the diffstat can be found below.
...
Naohiro Aota naohiro.aota@wdc.com btrfs: zoned: fix zone_unusable accounting on making block group read-write again
This patch results in a variety of problems with the parisc64 qemu emulation. Some examples from various test runs:
[ 10.527204] alg: akcipher: test 1 failed for rsa-generic, err=-22 [ 10.529743] alg: self-tests for rsa using rsa-generic failed (rc=-22) [ 10.529905] ------------[ cut here ]------------ [ 10.530276] alg: self-tests for rsa using rsa-generic failed (rc=-22) [ 10.530732] WARNING: CPU: 0 PID: 50 at crypto/testmgr.c:5907 alg_test+0x618/0x688
[ 7.420056] ================================================================== [ 7.420507] BUG: KFENCE: invalid read in walk_stackframe.isra.0+0xb4/0x138 [ 7.420507] [ 7.420827] Invalid read at 0x0000000042f07000: [ 7.421037] walk_stackframe.isra.0+0xb4/0x138 [ 7.421204] arch_stack_walk+0x38/0x50 [ 7.421338] stack_trace_save_regs+0x58/0x70 [ 7.421498] kfence_report_error+0x14c/0x730 [ 7.421649] kfence_handle_page_fault+0x2c8/0x2d0 [ 7.421996] handle_interruption+0x9b0/0xb58 [ 7.422168] intr_check_sig+0x0/0x3c
[ 8.891194] ============================================================================= [ 8.891558] BUG sgpool-32 (Tainted: G B W N): Wrong object count. Counter is 3 but counted were 18
[ 0.403174] ============================================================================= [ 0.403568] BUG audit_buffer (Not tainted): Wrong object count. Counter is 1 but counted were 34
[ 0.505914] ============================================================================= [ 0.506258] BUG skbuff_head_cache (Tainted: G B ): Freechain corrupt
[ 2.831636] ============================================================================= [ 2.832144] BUG skbuff_head_cache (Tainted: G B ): Left Redzone overwritten
Reverting it fixes the problem.
Bisect log is attached for reference.
I tried to repeat the test with v6.1.105-rc1, but that fails to compile for parisc64.
/home/groeck/src/linux-stable/include/linux/slab.h:228: warning: "ARCH_KMALLOC_MINALIGN" redefined 228 | #define ARCH_KMALLOC_MINALIGN ARCH_DMA_MINALIGN
This is due to commit 96423e23e05b ("parisc: fix a possible DMA corruption"). After reverting that patch I don't see the above problems anymore in v6.1.105-rc1, so it looks like the btrfs patches in v6.1.105-rc1 may have fixed it (or maybe there is another hidden bug in the parisc64 qemu emulation).
Oh, never mind. I did a reverse bisect on v6.1.105-rc1 (after fixing the build failure there), and it points to an unrelated commit. Looks like another qemu emulation bug. Sorry for the noise.
Guenter