While running selftests: memfd: run_hugetlbfs_test.sh on qemu_i386 and i386 the following invalid opcode was noticed on stable-rc 6.1 and 6.0.
This is always reproducible on stable-rc 6.1 and 6.0 with qemu_i386 and i386. Build, config and test log details provided in the below links [1].
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
# selftests: memfd: run_hugetlbfs_test.sh [ 111.866742] run_hugetlbfs_t (1023): drop_caches: 3 [ 111.892484] invalid opcode: 0000 [#1] PREEMPT SMP [ 111.893089] CPU: 2 PID: 1026 Comm: memfd_test Tainted: G N 6.1.4-rc1 #1 [ 111.894015] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014 [ 111.895048] EIP: hugetlb_file_setup.cold+0x0/0x33 [ 111.895597] Code: ff e9 07 f3 26 ff 0f 0b c7 04 24 c8 ee dd c9 e8 25 47 ff ff b8 ea ff ff ff e9 ac f8 26 ff 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b <0f> 0b 64 a1 98 48 3a ca c6 05 f3 42 22 ca 01 8b 90 d4 03 00 00 05 [ 111.897716] EAX: c4957128 EBX: 00000000 ECX: 00000020 EDX: 00000000 [ 111.898431] ESI: 80000004 EDI: 00000005 EBP: c552bf04 ESP: c552bee4 [ 111.899218] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 EFLAGS: 00010212 [ 111.899991] CR0: 80050033 CR2: 00000000 CR3: 0528b000 CR4: 003506d0 [ 111.900774] Call Trace: [ 111.901072] __ia32_sys_memfd_create+0x196/0x220 [ 111.901616] __do_fast_syscall_32+0x77/0xd0 [ 111.902119] do_fast_syscall_32+0x32/0x70 [ 111.902620] do_SYSENTER_32+0x15/0x20 [ 111.903134] entry_SYSENTER_32+0x98/0xf6 [ 111.903701] EIP: 0xb7ef1549 [ 111.904034] Code: 03 74 c0 01 10 05 03 74 b8 01 10 06 03 74 b4 01 10 07 03 74 b0 01 10 08 03 74 d8 01 00 00 00 00 00 51 52 55 89 e5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d 76 00 58 b8 77 00 00 00 cd 80 90 8d 76 [ 111.906124] EAX: ffffffda EBX: 0804b486 ECX: 80000004 EDX: 080493de [ 111.906902] ESI: b7ccf220 EDI: b7dc58e0 EBP: bfe56cf8 ESP: bfe56c8c [ 111.907637] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00000292 [ 111.908502] Modules linked in: sch_fq_codel fuse configfs [last unloaded: test_strscpy(N)] [ 111.909502] ---[ end trace 0000000000000000 ]--- [ 111.910064] EIP: hugetlb_file_setup.cold+0x0/0x33 [ 111.910638] Code: ff e9 07 f3 26 ff 0f 0b c7 04 24 c8 ee dd c9 e8 25 47 ff ff b8 ea ff ff ff e9 ac f8 26 ff 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b 0f 0b <0f> 0b 64 a1 98 48 3a ca c6 05 f3 42 22 ca 01 8b 90 d4 03 00 00 05 [ 111.912785] EAX: c4957128 EBX: 00000000 ECX: 00000020 EDX: 00000000 [ 111.913518] ESI: 80000004 EDI: 00000005 EBP: c552bf04 ESP: c552bee4 [ 111.914259] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 EFLAGS: 00010212 [ 111.915104] CR0: 80050033 CR2: 00000000 CR3: 0528b000 CR4: 003506d0 # ./run_hugetlbfs_test.sh: line 60: 1026 Segmentation fault ./memfd_test hugetlbfs # opening: ./mnt/memfd # fuse: DONE ok 3 selftests: memfd: run_hugetlbfs_test.sh
[1] https://lkft.validation.linaro.org/scheduler/job/6022150#L2079 https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-6.1.y/build/v6.1.3-... https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-6.1.y/build/v6.1.3-...
metadata: git_ref: linux-6.1.y git_repo: https://gitlab.com/Linaro/lkft/mirrors/stable/linux-stable-rc git_sha: a31425cbf493ef8bc7f7ce775a1028b1e0612f32 git_describe: v6.1.3-208-ga31425cbf493 kernel_version: 6.1.4-rc1 kernel-config: https://storage.tuxsuite.com/public/linaro/lkft/builds/2JrzvZc223pctlAxVhCIe... build-url: https://gitlab.com/Linaro/lkft/mirrors/stable/linux-stable-rc/-/pipelines/73... artifact-location: https://storage.tuxsuite.com/public/linaro/lkft/builds/2JrzvZc223pctlAxVhCIe... toolchain: gcc-11 vmlinux: https://storage.tuxsuite.com/public/linaro/lkft/builds/2JrzvZc223pctlAxVhCIe... System.map: https://storage.tuxsuite.com/public/linaro/lkft/builds/2JrzvZc223pctlAxVhCIe...
-- Linaro LKFT https://lkft.linaro.org
On 01/05/23 15:14, Naresh Kamboju wrote:
While running selftests: memfd: run_hugetlbfs_test.sh on qemu_i386 and i386 the following invalid opcode was noticed on stable-rc 6.1 and 6.0.
This is always reproducible on stable-rc 6.1 and 6.0 with qemu_i386 and i386. Build, config and test log details provided in the below links [1].
Hello Naresh,
I have tried to create this issue a few times without success. Since I do not have i386 HW, I am using qemu_i386. If I use the supplied config, my kernel does not boot. I then try to modify config options which I think are not relevant. By the time I get to a config that will boot, I can not recreate the issue. :(
Just curious if you have any suggestions? Or, Wondering if anyone else has suggestions on how to proceed?
[Resending because logs.txt attachment was not successful ] Hi Mike,
On Wed, 25 Jan 2023 at 06:04, Mike Kravetz mike.kravetz@oracle.com wrote:
On 01/05/23 15:14, Naresh Kamboju wrote:
While running selftests: memfd: run_hugetlbfs_test.sh on qemu_i386 and i386 the following invalid opcode was noticed on stable-rc 6.1 and 6.0.
This is always reproducible on stable-rc 6.1 and 6.0 with qemu_i386 and i386. Build, config and test log details provided in the below links [1].
Hello Naresh,
I have tried to create this issue a few times without success. Since I do not have i386 HW, I am using qemu_i386. If I use the supplied config, my kernel does not boot. I then try to modify config options which I think are not relevant. By the time I get to a config that will boot, I can not recreate the issue. :(
Just curious if you have any suggestions? Or, Wondering if anyone else has suggestions on how to proceed?
Please install tuxmake and run attached script to reproduce reported issues, $ pip3 install tuxmake $ ./memfd-crash-test-qemu-i386.sh
This script downloads kernel Image and rootfs and runs run_hugetlbfs_test.sh. If you have any questions please get back to me.
-- Mike Kravetz
linux-kselftest-mirror@lists.linaro.org