diff mbox series

[RFC,PATCHv2] mm: use stack_depot for recording kmemleak's backtrace

Message ID 1665561689-29498-1-git-send-email-zhaoyang.huang@unisoc.com (mailing list archive)
State New, archived
Headers show
Series [RFC,PATCHv2] mm: use stack_depot for recording kmemleak's backtrace | expand

Commit Message

zhaoyang.huang Oct. 12, 2022, 8:01 a.m. UTC
From: Zhaoyang Huang <zhaoyang.huang@unisoc.com>

Using stack_depot to record kmemleak's backtrace which has been implemented
on slub for reducing redundant information.

Signed-off-by: Zhaoyang Huang <zhaoyang.huang@unisoc.com>
---
changes of v2: fix bugs of stack_depot_init related issue
---
---
 mm/kmemleak.c | 45 +++++++++++++++++++++++++++++++++++----------
 1 file changed, 35 insertions(+), 10 deletions(-)

Comments

Zhaoyang Huang Oct. 13, 2022, 8:38 a.m. UTC | #1
@Vlastimil Could you please have an eye on this series of robot
reports which are caused by stack_depot_init related issues. The
problem arises from a very early access of stack_depot_save/init by
kmemleak within setup_arch which happens even before
stack_depot_early_init and zone related data ready. I would like to
suggest adding a criteria at the entrance check of stack_depot_save
which help the stackdepot API more aggregation and the caller free to
call

[    0.062350][    T0]  ? stack_depot_init.cold+0x5/0xbd
[    0.063072][    T0]  ? set_track_prepare+0x6e/0x80
[    0.063957][    T0]  ?
__raw_callee_save___native_queued_spin_unlock+0x11/0x22
[    0.064952][    T0]  ? write_comp_data+0x2a/0x80
[    0.065623][    T0]  ? strncpy+0x2f/0x60
[    0.066205][    T0]  ? __create_object+0x10c/0x3c0
[    0.066904][    T0]  ? kmemleak_alloc_phys+0x6f/0x80
[    0.067561][    T0]  ? memblock_alloc_range_nid+0x274/0x28f
[    0.068396][    T0]  ? memblock_phys_alloc_range+0xa4/0xb3
[    0.069200][    T0]  ? reserve_real_mode+0x87/0xd7
[    0.069895][    T0]  ? setup_arch+0x6a9/0x995
[    0.070526][    T0]  ? start_kernel+0x7c/0x854
[    0.071195][    T0]  ? load_ucode_bsp+0x1bb/0x1c6
[    0.071875][    T0]  ? secondary_startup_64_no_verify+0xe0/0xeb
[    0.072682][    T0]  </TASK>

On Thu, Oct 13, 2022 at 1:58 PM kernel test robot <oliver.sang@intel.com> wrote:
>
>
> Hi zhaoyang.huang,
>
> seems this is the fix based on our report
>     "[mm]  0e949320db: BUG:kernel_NULL_pointer_dereference,address"
> at
>     https://lore.kernel.org/all/202210121406.d4ebc9bc-oliver.sang@intel.com/
> but now it seems have new issue. report FYI
>
>
> Greeting,
>
> FYI, we noticed the following commit (built with gcc-11):
>
> commit: db8d280d38efb061ad1a57ce060cbb917a4cf503 ("[RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace")
> url: https://github.com/intel-lab-lkp/linux/commits/zhaoyang-huang/mm-use-stack_depot-for-recording-kmemleak-s-backtrace/20221012-160458
> base: https://git.kernel.org/cgit/linux/kernel/git/akpm/mm.git mm-everything
> patch link: https://lore.kernel.org/linux-mm/1665561689-29498-1-git-send-email-zhaoyang.huang@unisoc.com
> patch subject: [RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace
>
> in testcase: boot
>
> on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
>
> caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
>
>
> +-------------------------------+------------+------------+
> |                               | 95f1b43741 | db8d280d38 |
> +-------------------------------+------------+------------+
> | boot_successes                | 20         | 0          |
> | boot_failures                 | 0          | 18         |
> | PANIC:early_exception         | 0          | 18         |
> | RIP:nr_free_zone_pages        | 0          | 18         |
> | BUG:kernel_hang_in_boot_stage | 0          | 18         |
> +-------------------------------+------------+------------+
>
>
> If you fix the issue, kindly add following tag
> | Reported-by: kernel test robot <oliver.sang@intel.com>
> | Link: https://lore.kernel.org/r/202210131309.fe5427b-oliver.sang@intel.com
>
>
> [    0.029254][    T0] Scan for SMP in [mem 0x00000000-0x000003ff]
> [    0.030178][    T0] Scan for SMP in [mem 0x0009fc00-0x0009ffff]
> [    0.031080][    T0] Scan for SMP in [mem 0x000f0000-0x000fffff]
> [    0.043370][    T0] found SMP MP-table at [mem 0x000f5ba0-0x000f5baf]
> [    0.044301][    T0]   mpc: f5bb0-f5c80
> PANIC: early exception 0x0e IP 10:ffffffff8149c282 error 0 cr2 0x1e08
> [    0.045770][    T0] CPU: 0 PID: 0 Comm: swapper Not tainted 6.0.0-rc3-00711-gdb8d280d38ef #5
> [    0.046970][    T0] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
> [ 0.048356][ T0] RIP: 0010:nr_free_zone_pages (kbuild/src/x86_64-2/include/linux/mmzone.h:1478 kbuild/src/x86_64-2/include/linux/mmzone.h:1504 kbuild/src/x86_64-2/mm/page_alloc.c:5886)
> [ 0.049158][ T0] Code: e9 a3 5c 0b 00 0f 1f 00 e8 9b cd be ff 65 8b 05 14 b8 b8 7e 48 98 41 54 48 8b 04 c5 e0 e4 bc 83 53 89 fb 4c 8d 80 00 1e 00 00 <3b> b8 08 1e 00 00 72 5b 49 8b 10 45 31 e4 48 85 d2 75 05 eb 35 49
> All code
> ========
>    0:   e9 a3 5c 0b 00          jmpq   0xb5ca8
>    5:   0f 1f 00                nopl   (%rax)
>    8:   e8 9b cd be ff          callq  0xffffffffffbecda8
>    d:   65 8b 05 14 b8 b8 7e    mov    %gs:0x7eb8b814(%rip),%eax        # 0x7eb8b828
>   14:   48 98                   cltq
>   16:   41 54                   push   %r12
>   18:   48 8b 04 c5 e0 e4 bc    mov    -0x7c431b20(,%rax,8),%rax
>   1f:   83
>   20:   53                      push   %rbx
>   21:   89 fb                   mov    %edi,%ebx
>   23:   4c 8d 80 00 1e 00 00    lea    0x1e00(%rax),%r8
>   2a:*  3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi                <-- trapping instruction
>   30:   72 5b                   jb     0x8d
>   32:   49 8b 10                mov    (%r8),%rdx
>   35:   45 31 e4                xor    %r12d,%r12d
>   38:   48 85 d2                test   %rdx,%rdx
>   3b:   75 05                   jne    0x42
>   3d:   eb 35                   jmp    0x74
>   3f:   49                      rex.WB
>
> Code starting with the faulting instruction
> ===========================================
>    0:   3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi
>    6:   72 5b                   jb     0x63
>    8:   49 8b 10                mov    (%r8),%rdx
>    b:   45 31 e4                xor    %r12d,%r12d
>    e:   48 85 d2                test   %rdx,%rdx
>   11:   75 05                   jne    0x18
>   13:   eb 35                   jmp    0x4a
>   15:   49                      rex.WB
> [    0.051803][    T0] RSP: 0000:ffffffff83603d18 EFLAGS: 00010046 ORIG_RAX: 0000000000000000
> [    0.052881][    T0] RAX: 0000000000000000 RBX: 0000000000000002 RCX: 0001ffff84b317e0
> [    0.053932][    T0] RDX: 0000000000000485 RSI: 0001ffffffffffff RDI: 0000000000000002
> [    0.055007][    T0] RBP: ffffffff84b077d0 R08: 0000000000001e00 R09: 0000000000000000
> [    0.056054][    T0] R10: ffffffff81b16d30 R11: 0001ffff84b317e8 R12: 0000000000000001
> [    0.057088][    T0] R13: ffffffff84b077d8 R14: 0000000000098000 R15: 0000000000007000
> [    0.058160][    T0] FS:  0000000000000000(0000) GS:ffffffff842c9000(0000) knlGS:0000000000000000
> [    0.059440][    T0] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [    0.060381][    T0] CR2: 0000000000001e08 CR3: 00000000043b6000 CR4: 00000000000406a0
> [    0.061516][    T0] Call Trace:
> [    0.061969][    T0]  <TASK>
> [ 0.062350][ T0] ? stack_depot_init.cold (kbuild/src/x86_64-2/lib/stackdepot.c:258)
> [ 0.063072][ T0] ? set_track_prepare (kbuild/src/x86_64-2/mm/slub.c:752)
> [ 0.063957][ T0] ? __raw_callee_save___native_queued_spin_unlock (??:?)
> [ 0.064952][ T0] ? write_comp_data (kbuild/src/x86_64-2/kernel/kcov.c:236)
>
>
> To reproduce:
>
>         # build kernel
>         cd linux
>         cp config-6.0.0-rc3-00711-gdb8d280d38ef .config
>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 olddefconfig prepare modules_prepare bzImage modules
>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 INSTALL_MOD_PATH=<mod-install-dir> modules_install
>         cd <mod-install-dir>
>         find lib/ | cpio -o -H newc --quiet | gzip > modules.cgz
>
>
>         git clone https://github.com/intel/lkp-tests.git
>         cd lkp-tests
>         bin/lkp qemu -k <bzImage> -m modules.cgz job-script # job-script is attached in this email
>
>         # if come across any failure that blocks the test,
>         # please remove ~/.lkp and /lkp dir to run from a clean state.
>
>
>
> --
> 0-DAY CI Kernel Test Service
> https://01.org/lkp
>
>
Vlastimil Babka Oct. 13, 2022, 8:51 a.m. UTC | #2
On 10/13/22 10:38, Zhaoyang Huang wrote:
> @Vlastimil Could you please have an eye on this series of robot
> reports which are caused by stack_depot_init related issues. The
> problem arises from a very early access of stack_depot_save/init by
> kmemleak within setup_arch which happens even before
> stack_depot_early_init and zone related data ready. I would like to
> suggest adding a criteria at the entrance check of stack_depot_save
> which help the stackdepot API more aggregation and the caller free to
> call

From the report it seems to me the problem is in your set_track_prepare():

+	if (!kmemleak_initialized)
+		stack_depot_init();

stack_depot_init() should be called only once and from well defined places,
this looks like neither condition is met.

> [    0.062350][    T0]  ? stack_depot_init.cold+0x5/0xbd
> [    0.063072][    T0]  ? set_track_prepare+0x6e/0x80
> [    0.063957][    T0]  ?
> __raw_callee_save___native_queued_spin_unlock+0x11/0x22
> [    0.064952][    T0]  ? write_comp_data+0x2a/0x80
> [    0.065623][    T0]  ? strncpy+0x2f/0x60
> [    0.066205][    T0]  ? __create_object+0x10c/0x3c0
> [    0.066904][    T0]  ? kmemleak_alloc_phys+0x6f/0x80
> [    0.067561][    T0]  ? memblock_alloc_range_nid+0x274/0x28f
> [    0.068396][    T0]  ? memblock_phys_alloc_range+0xa4/0xb3
> [    0.069200][    T0]  ? reserve_real_mode+0x87/0xd7
> [    0.069895][    T0]  ? setup_arch+0x6a9/0x995
> [    0.070526][    T0]  ? start_kernel+0x7c/0x854
> [    0.071195][    T0]  ? load_ucode_bsp+0x1bb/0x1c6
> [    0.071875][    T0]  ? secondary_startup_64_no_verify+0xe0/0xeb
> [    0.072682][    T0]  </TASK>
> 
> On Thu, Oct 13, 2022 at 1:58 PM kernel test robot <oliver.sang@intel.com> wrote:
>>
>>
>> Hi zhaoyang.huang,
>>
>> seems this is the fix based on our report
>>     "[mm]  0e949320db: BUG:kernel_NULL_pointer_dereference,address"
>> at
>>     https://lore.kernel.org/all/202210121406.d4ebc9bc-oliver.sang@intel.com/
>> but now it seems have new issue. report FYI
>>
>>
>> Greeting,
>>
>> FYI, we noticed the following commit (built with gcc-11):
>>
>> commit: db8d280d38efb061ad1a57ce060cbb917a4cf503 ("[RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace")
>> url: https://github.com/intel-lab-lkp/linux/commits/zhaoyang-huang/mm-use-stack_depot-for-recording-kmemleak-s-backtrace/20221012-160458
>> base: https://git.kernel.org/cgit/linux/kernel/git/akpm/mm.git mm-everything
>> patch link: https://lore.kernel.org/linux-mm/1665561689-29498-1-git-send-email-zhaoyang.huang@unisoc.com
>> patch subject: [RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace
>>
>> in testcase: boot
>>
>> on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
>>
>> caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
>>
>>
>> +-------------------------------+------------+------------+
>> |                               | 95f1b43741 | db8d280d38 |
>> +-------------------------------+------------+------------+
>> | boot_successes                | 20         | 0          |
>> | boot_failures                 | 0          | 18         |
>> | PANIC:early_exception         | 0          | 18         |
>> | RIP:nr_free_zone_pages        | 0          | 18         |
>> | BUG:kernel_hang_in_boot_stage | 0          | 18         |
>> +-------------------------------+------------+------------+
>>
>>
>> If you fix the issue, kindly add following tag
>> | Reported-by: kernel test robot <oliver.sang@intel.com>
>> | Link: https://lore.kernel.org/r/202210131309.fe5427b-oliver.sang@intel.com
>>
>>
>> [    0.029254][    T0] Scan for SMP in [mem 0x00000000-0x000003ff]
>> [    0.030178][    T0] Scan for SMP in [mem 0x0009fc00-0x0009ffff]
>> [    0.031080][    T0] Scan for SMP in [mem 0x000f0000-0x000fffff]
>> [    0.043370][    T0] found SMP MP-table at [mem 0x000f5ba0-0x000f5baf]
>> [    0.044301][    T0]   mpc: f5bb0-f5c80
>> PANIC: early exception 0x0e IP 10:ffffffff8149c282 error 0 cr2 0x1e08
>> [    0.045770][    T0] CPU: 0 PID: 0 Comm: swapper Not tainted 6.0.0-rc3-00711-gdb8d280d38ef #5
>> [    0.046970][    T0] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
>> [ 0.048356][ T0] RIP: 0010:nr_free_zone_pages (kbuild/src/x86_64-2/include/linux/mmzone.h:1478 kbuild/src/x86_64-2/include/linux/mmzone.h:1504 kbuild/src/x86_64-2/mm/page_alloc.c:5886)
>> [ 0.049158][ T0] Code: e9 a3 5c 0b 00 0f 1f 00 e8 9b cd be ff 65 8b 05 14 b8 b8 7e 48 98 41 54 48 8b 04 c5 e0 e4 bc 83 53 89 fb 4c 8d 80 00 1e 00 00 <3b> b8 08 1e 00 00 72 5b 49 8b 10 45 31 e4 48 85 d2 75 05 eb 35 49
>> All code
>> ========
>>    0:   e9 a3 5c 0b 00          jmpq   0xb5ca8
>>    5:   0f 1f 00                nopl   (%rax)
>>    8:   e8 9b cd be ff          callq  0xffffffffffbecda8
>>    d:   65 8b 05 14 b8 b8 7e    mov    %gs:0x7eb8b814(%rip),%eax        # 0x7eb8b828
>>   14:   48 98                   cltq
>>   16:   41 54                   push   %r12
>>   18:   48 8b 04 c5 e0 e4 bc    mov    -0x7c431b20(,%rax,8),%rax
>>   1f:   83
>>   20:   53                      push   %rbx
>>   21:   89 fb                   mov    %edi,%ebx
>>   23:   4c 8d 80 00 1e 00 00    lea    0x1e00(%rax),%r8
>>   2a:*  3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi                <-- trapping instruction
>>   30:   72 5b                   jb     0x8d
>>   32:   49 8b 10                mov    (%r8),%rdx
>>   35:   45 31 e4                xor    %r12d,%r12d
>>   38:   48 85 d2                test   %rdx,%rdx
>>   3b:   75 05                   jne    0x42
>>   3d:   eb 35                   jmp    0x74
>>   3f:   49                      rex.WB
>>
>> Code starting with the faulting instruction
>> ===========================================
>>    0:   3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi
>>    6:   72 5b                   jb     0x63
>>    8:   49 8b 10                mov    (%r8),%rdx
>>    b:   45 31 e4                xor    %r12d,%r12d
>>    e:   48 85 d2                test   %rdx,%rdx
>>   11:   75 05                   jne    0x18
>>   13:   eb 35                   jmp    0x4a
>>   15:   49                      rex.WB
>> [    0.051803][    T0] RSP: 0000:ffffffff83603d18 EFLAGS: 00010046 ORIG_RAX: 0000000000000000
>> [    0.052881][    T0] RAX: 0000000000000000 RBX: 0000000000000002 RCX: 0001ffff84b317e0
>> [    0.053932][    T0] RDX: 0000000000000485 RSI: 0001ffffffffffff RDI: 0000000000000002
>> [    0.055007][    T0] RBP: ffffffff84b077d0 R08: 0000000000001e00 R09: 0000000000000000
>> [    0.056054][    T0] R10: ffffffff81b16d30 R11: 0001ffff84b317e8 R12: 0000000000000001
>> [    0.057088][    T0] R13: ffffffff84b077d8 R14: 0000000000098000 R15: 0000000000007000
>> [    0.058160][    T0] FS:  0000000000000000(0000) GS:ffffffff842c9000(0000) knlGS:0000000000000000
>> [    0.059440][    T0] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [    0.060381][    T0] CR2: 0000000000001e08 CR3: 00000000043b6000 CR4: 00000000000406a0
>> [    0.061516][    T0] Call Trace:
>> [    0.061969][    T0]  <TASK>
>> [ 0.062350][ T0] ? stack_depot_init.cold (kbuild/src/x86_64-2/lib/stackdepot.c:258)
>> [ 0.063072][ T0] ? set_track_prepare (kbuild/src/x86_64-2/mm/slub.c:752)
>> [ 0.063957][ T0] ? __raw_callee_save___native_queued_spin_unlock (??:?)
>> [ 0.064952][ T0] ? write_comp_data (kbuild/src/x86_64-2/kernel/kcov.c:236)
>>
>>
>> To reproduce:
>>
>>         # build kernel
>>         cd linux
>>         cp config-6.0.0-rc3-00711-gdb8d280d38ef .config
>>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 olddefconfig prepare modules_prepare bzImage modules
>>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 INSTALL_MOD_PATH=<mod-install-dir> modules_install
>>         cd <mod-install-dir>
>>         find lib/ | cpio -o -H newc --quiet | gzip > modules.cgz
>>
>>
>>         git clone https://github.com/intel/lkp-tests.git
>>         cd lkp-tests
>>         bin/lkp qemu -k <bzImage> -m modules.cgz job-script # job-script is attached in this email
>>
>>         # if come across any failure that blocks the test,
>>         # please remove ~/.lkp and /lkp dir to run from a clean state.
>>
>>
>>
>> --
>> 0-DAY CI Kernel Test Service
>> https://01.org/lkp
>>
>>
Zhaoyang Huang Oct. 13, 2022, 8:59 a.m. UTC | #3
On Thu, Oct 13, 2022 at 4:51 PM Vlastimil Babka <vbabka@suse.cz> wrote:
>
> On 10/13/22 10:38, Zhaoyang Huang wrote:
> > @Vlastimil Could you please have an eye on this series of robot
> > reports which are caused by stack_depot_init related issues. The
> > problem arises from a very early access of stack_depot_save/init by
> > kmemleak within setup_arch which happens even before
> > stack_depot_early_init and zone related data ready. I would like to
> > suggest adding a criteria at the entrance check of stack_depot_save
> > which help the stackdepot API more aggregation and the caller free to
> > call
>
> From the report it seems to me the problem is in your set_track_prepare():
>
> +       if (!kmemleak_initialized)
> +               stack_depot_init();
>
> stack_depot_init() should be called only once and from well defined places,
> this looks like neither condition is met.
yes, but this is not the key point for this series problem as kmemleak
wants to access stack_depot_save during setup_arch and lead to NULL
stack_table access. To my understanding, there is no way to provide a
valid stack_table so far until stack_depot_early_init called. kmemleak
could take another action if it get a vacant trace_handle from
stack_depot_save.
>
> > [    0.062350][    T0]  ? stack_depot_init.cold+0x5/0xbd
> > [    0.063072][    T0]  ? set_track_prepare+0x6e/0x80
> > [    0.063957][    T0]  ?
> > __raw_callee_save___native_queued_spin_unlock+0x11/0x22
> > [    0.064952][    T0]  ? write_comp_data+0x2a/0x80
> > [    0.065623][    T0]  ? strncpy+0x2f/0x60
> > [    0.066205][    T0]  ? __create_object+0x10c/0x3c0
> > [    0.066904][    T0]  ? kmemleak_alloc_phys+0x6f/0x80
> > [    0.067561][    T0]  ? memblock_alloc_range_nid+0x274/0x28f
> > [    0.068396][    T0]  ? memblock_phys_alloc_range+0xa4/0xb3
> > [    0.069200][    T0]  ? reserve_real_mode+0x87/0xd7
> > [    0.069895][    T0]  ? setup_arch+0x6a9/0x995
> > [    0.070526][    T0]  ? start_kernel+0x7c/0x854
> > [    0.071195][    T0]  ? load_ucode_bsp+0x1bb/0x1c6
> > [    0.071875][    T0]  ? secondary_startup_64_no_verify+0xe0/0xeb
> > [    0.072682][    T0]  </TASK>
> >
> > On Thu, Oct 13, 2022 at 1:58 PM kernel test robot <oliver.sang@intel.com> wrote:
> >>
> >>
> >> Hi zhaoyang.huang,
> >>
> >> seems this is the fix based on our report
> >>     "[mm]  0e949320db: BUG:kernel_NULL_pointer_dereference,address"
> >> at
> >>     https://lore.kernel.org/all/202210121406.d4ebc9bc-oliver.sang@intel.com/
> >> but now it seems have new issue. report FYI
> >>
> >>
> >> Greeting,
> >>
> >> FYI, we noticed the following commit (built with gcc-11):
> >>
> >> commit: db8d280d38efb061ad1a57ce060cbb917a4cf503 ("[RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace")
> >> url: https://github.com/intel-lab-lkp/linux/commits/zhaoyang-huang/mm-use-stack_depot-for-recording-kmemleak-s-backtrace/20221012-160458
> >> base: https://git.kernel.org/cgit/linux/kernel/git/akpm/mm.git mm-everything
> >> patch link: https://lore.kernel.org/linux-mm/1665561689-29498-1-git-send-email-zhaoyang.huang@unisoc.com
> >> patch subject: [RFC PATCHv2] mm: use stack_depot for recording kmemleak's backtrace
> >>
> >> in testcase: boot
> >>
> >> on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
> >>
> >> caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
> >>
> >>
> >> +-------------------------------+------------+------------+
> >> |                               | 95f1b43741 | db8d280d38 |
> >> +-------------------------------+------------+------------+
> >> | boot_successes                | 20         | 0          |
> >> | boot_failures                 | 0          | 18         |
> >> | PANIC:early_exception         | 0          | 18         |
> >> | RIP:nr_free_zone_pages        | 0          | 18         |
> >> | BUG:kernel_hang_in_boot_stage | 0          | 18         |
> >> +-------------------------------+------------+------------+
> >>
> >>
> >> If you fix the issue, kindly add following tag
> >> | Reported-by: kernel test robot <oliver.sang@intel.com>
> >> | Link: https://lore.kernel.org/r/202210131309.fe5427b-oliver.sang@intel.com
> >>
> >>
> >> [    0.029254][    T0] Scan for SMP in [mem 0x00000000-0x000003ff]
> >> [    0.030178][    T0] Scan for SMP in [mem 0x0009fc00-0x0009ffff]
> >> [    0.031080][    T0] Scan for SMP in [mem 0x000f0000-0x000fffff]
> >> [    0.043370][    T0] found SMP MP-table at [mem 0x000f5ba0-0x000f5baf]
> >> [    0.044301][    T0]   mpc: f5bb0-f5c80
> >> PANIC: early exception 0x0e IP 10:ffffffff8149c282 error 0 cr2 0x1e08
> >> [    0.045770][    T0] CPU: 0 PID: 0 Comm: swapper Not tainted 6.0.0-rc3-00711-gdb8d280d38ef #5
> >> [    0.046970][    T0] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.0-debian-1.16.0-4 04/01/2014
> >> [ 0.048356][ T0] RIP: 0010:nr_free_zone_pages (kbuild/src/x86_64-2/include/linux/mmzone.h:1478 kbuild/src/x86_64-2/include/linux/mmzone.h:1504 kbuild/src/x86_64-2/mm/page_alloc.c:5886)
> >> [ 0.049158][ T0] Code: e9 a3 5c 0b 00 0f 1f 00 e8 9b cd be ff 65 8b 05 14 b8 b8 7e 48 98 41 54 48 8b 04 c5 e0 e4 bc 83 53 89 fb 4c 8d 80 00 1e 00 00 <3b> b8 08 1e 00 00 72 5b 49 8b 10 45 31 e4 48 85 d2 75 05 eb 35 49
> >> All code
> >> ========
> >>    0:   e9 a3 5c 0b 00          jmpq   0xb5ca8
> >>    5:   0f 1f 00                nopl   (%rax)
> >>    8:   e8 9b cd be ff          callq  0xffffffffffbecda8
> >>    d:   65 8b 05 14 b8 b8 7e    mov    %gs:0x7eb8b814(%rip),%eax        # 0x7eb8b828
> >>   14:   48 98                   cltq
> >>   16:   41 54                   push   %r12
> >>   18:   48 8b 04 c5 e0 e4 bc    mov    -0x7c431b20(,%rax,8),%rax
> >>   1f:   83
> >>   20:   53                      push   %rbx
> >>   21:   89 fb                   mov    %edi,%ebx
> >>   23:   4c 8d 80 00 1e 00 00    lea    0x1e00(%rax),%r8
> >>   2a:*  3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi                <-- trapping instruction
> >>   30:   72 5b                   jb     0x8d
> >>   32:   49 8b 10                mov    (%r8),%rdx
> >>   35:   45 31 e4                xor    %r12d,%r12d
> >>   38:   48 85 d2                test   %rdx,%rdx
> >>   3b:   75 05                   jne    0x42
> >>   3d:   eb 35                   jmp    0x74
> >>   3f:   49                      rex.WB
> >>
> >> Code starting with the faulting instruction
> >> ===========================================
> >>    0:   3b b8 08 1e 00 00       cmp    0x1e08(%rax),%edi
> >>    6:   72 5b                   jb     0x63
> >>    8:   49 8b 10                mov    (%r8),%rdx
> >>    b:   45 31 e4                xor    %r12d,%r12d
> >>    e:   48 85 d2                test   %rdx,%rdx
> >>   11:   75 05                   jne    0x18
> >>   13:   eb 35                   jmp    0x4a
> >>   15:   49                      rex.WB
> >> [    0.051803][    T0] RSP: 0000:ffffffff83603d18 EFLAGS: 00010046 ORIG_RAX: 0000000000000000
> >> [    0.052881][    T0] RAX: 0000000000000000 RBX: 0000000000000002 RCX: 0001ffff84b317e0
> >> [    0.053932][    T0] RDX: 0000000000000485 RSI: 0001ffffffffffff RDI: 0000000000000002
> >> [    0.055007][    T0] RBP: ffffffff84b077d0 R08: 0000000000001e00 R09: 0000000000000000
> >> [    0.056054][    T0] R10: ffffffff81b16d30 R11: 0001ffff84b317e8 R12: 0000000000000001
> >> [    0.057088][    T0] R13: ffffffff84b077d8 R14: 0000000000098000 R15: 0000000000007000
> >> [    0.058160][    T0] FS:  0000000000000000(0000) GS:ffffffff842c9000(0000) knlGS:0000000000000000
> >> [    0.059440][    T0] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >> [    0.060381][    T0] CR2: 0000000000001e08 CR3: 00000000043b6000 CR4: 00000000000406a0
> >> [    0.061516][    T0] Call Trace:
> >> [    0.061969][    T0]  <TASK>
> >> [ 0.062350][ T0] ? stack_depot_init.cold (kbuild/src/x86_64-2/lib/stackdepot.c:258)
> >> [ 0.063072][ T0] ? set_track_prepare (kbuild/src/x86_64-2/mm/slub.c:752)
> >> [ 0.063957][ T0] ? __raw_callee_save___native_queued_spin_unlock (??:?)
> >> [ 0.064952][ T0] ? write_comp_data (kbuild/src/x86_64-2/kernel/kcov.c:236)
> >>
> >>
> >> To reproduce:
> >>
> >>         # build kernel
> >>         cd linux
> >>         cp config-6.0.0-rc3-00711-gdb8d280d38ef .config
> >>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 olddefconfig prepare modules_prepare bzImage modules
> >>         make HOSTCC=gcc-11 CC=gcc-11 ARCH=x86_64 INSTALL_MOD_PATH=<mod-install-dir> modules_install
> >>         cd <mod-install-dir>
> >>         find lib/ | cpio -o -H newc --quiet | gzip > modules.cgz
> >>
> >>
> >>         git clone https://github.com/intel/lkp-tests.git
> >>         cd lkp-tests
> >>         bin/lkp qemu -k <bzImage> -m modules.cgz job-script # job-script is attached in this email
> >>
> >>         # if come across any failure that blocks the test,
> >>         # please remove ~/.lkp and /lkp dir to run from a clean state.
> >>
> >>
> >>
> >> --
> >> 0-DAY CI Kernel Test Service
> >> https://01.org/lkp
> >>
> >>
>
diff mbox series

Patch

diff --git a/mm/kmemleak.c b/mm/kmemleak.c
index 1eddc01..0f78774 100644
--- a/mm/kmemleak.c
+++ b/mm/kmemleak.c
@@ -79,6 +79,7 @@ 
 #include <linux/mutex.h>
 #include <linux/rcupdate.h>
 #include <linux/stacktrace.h>
+#include <linux/stackdepot.h>
 #include <linux/cache.h>
 #include <linux/percpu.h>
 #include <linux/memblock.h>
@@ -159,8 +160,7 @@  struct kmemleak_object {
 	u32 checksum;
 	/* memory ranges to be scanned inside an object (empty for all) */
 	struct hlist_head area_list;
-	unsigned long trace[MAX_TRACE];
-	unsigned int trace_len;
+	depot_stack_handle_t trace_handle;
 	unsigned long jiffies;		/* creation timestamp */
 	pid_t pid;			/* pid of the current task */
 	char comm[TASK_COMM_LEN];	/* executable name */
@@ -346,8 +346,11 @@  static void print_unreferenced(struct seq_file *seq,
 			       struct kmemleak_object *object)
 {
 	int i;
+	unsigned long *entries;
+	unsigned int nr_entries;
 	unsigned int msecs_age = jiffies_to_msecs(jiffies - object->jiffies);
 
+	nr_entries = stack_depot_fetch(object->trace_handle, &entries);
 	warn_or_seq_printf(seq, "unreferenced object 0x%08lx (size %zu):\n",
 		   object->pointer, object->size);
 	warn_or_seq_printf(seq, "  comm \"%s\", pid %d, jiffies %lu (age %d.%03ds)\n",
@@ -356,10 +359,10 @@  static void print_unreferenced(struct seq_file *seq,
 	hex_dump_object(seq, object);
 	warn_or_seq_printf(seq, "  backtrace:\n");
 
-	for (i = 0; i < object->trace_len; i++) {
-		void *ptr = (void *)object->trace[i];
-		warn_or_seq_printf(seq, "    [<%p>] %pS\n", ptr, ptr);
-	}
+       for (i = 0; i < nr_entries; i++) {
+               void *ptr = (void *)entries[i];
+               warn_or_seq_printf(seq, "    [<%p>] %pS\n", ptr, ptr);
+       }
 }
 
 /*
@@ -378,7 +381,8 @@  static void dump_object_info(struct kmemleak_object *object)
 	pr_notice("  flags = 0x%x\n", object->flags);
 	pr_notice("  checksum = %u\n", object->checksum);
 	pr_notice("  backtrace:\n");
-	stack_trace_print(object->trace, object->trace_len, 4);
+	if(object->trace_handle)
+		stack_depot_print(object->trace_handle);
 }
 
 /*
@@ -591,6 +595,27 @@  static struct kmemleak_object *find_and_remove_object(unsigned long ptr, int ali
 	return object;
 }
 
+#ifdef CONFIG_STACKDEPOT
+static noinline depot_stack_handle_t set_track_prepare(void)
+{
+	depot_stack_handle_t trace_handle;
+	unsigned long entries[MAX_TRACE];
+	unsigned int nr_entries;
+
+	if (!kmemleak_initialized)
+		stack_depot_init();
+	nr_entries = stack_trace_save(entries, ARRAY_SIZE(entries), 3);
+	trace_handle = stack_depot_save(entries, nr_entries, GFP_NOWAIT);
+
+	return trace_handle;
+}
+#else
+static inline depot_stack_handle_t set_track_prepare(void)
+{
+	return 0;
+}
+#endif
+
 /*
  * Save stack trace to the given array of MAX_TRACE size.
  */
@@ -654,7 +679,7 @@  static struct kmemleak_object *__create_object(unsigned long ptr, size_t size,
 	}
 
 	/* kernel backtrace */
-	object->trace_len = __save_stack_trace(object->trace);
+	object->trace_handle = set_track_prepare();
 
 	raw_spin_lock_irqsave(&kmemleak_lock, flags);
 
@@ -694,7 +719,6 @@  static struct kmemleak_object *__create_object(unsigned long ptr, size_t size,
 	rb_link_node(&object->rb_node, rb_parent, link);
 	rb_insert_color(&object->rb_node, is_phys ? &object_phys_tree_root :
 					  &object_tree_root);
-
 	list_add_tail_rcu(&object->object_list, &object_list);
 out:
 	raw_spin_unlock_irqrestore(&kmemleak_lock, flags);
@@ -1094,7 +1118,7 @@  void __ref kmemleak_update_trace(const void *ptr)
 	}
 
 	raw_spin_lock_irqsave(&object->lock, flags);
-	object->trace_len = __save_stack_trace(object->trace);
+	object->trace_handle = set_track_prepare();
 	raw_spin_unlock_irqrestore(&object->lock, flags);
 
 	put_object(object);
@@ -2064,6 +2088,7 @@  void __init kmemleak_init(void)
 	if (kmemleak_error)
 		return;
 
+	stack_depot_init();
 	jiffies_min_age = msecs_to_jiffies(MSECS_MIN_AGE);
 	jiffies_scan_wait = msecs_to_jiffies(SECS_SCAN_WAIT * 1000);