* [linus:master] [mm/codetag] 51f43d5d82: WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set
@ 2024-12-11 15:08 kernel test robot
2024-12-12 2:12 ` David Wang
2024-12-12 4:01 ` [PATCH] mm/codetag: clear tags before swap David Wang
0 siblings, 2 replies; 10+ messages in thread
From: kernel test robot @ 2024-12-11 15:08 UTC (permalink / raw)
To: David Wang
Cc: oe-lkp, lkp, linux-kernel, Andrew Morton, Suren Baghdasaryan,
Yu Zhao, Kent Overstreet, linux-mm, oliver.sang
Hello,
kernel test robot noticed "WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set" on:
commit: 51f43d5d82ed2ba3f9a3f9a2390c52f28e42af32 ("mm/codetag: swap tags when migrate pages")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
[test failed on linus/master fac04efc5c793dccbd07e2d59af9f90b7fc0dca4]
[test failed on linux-next/master ebe1b11614e079c5e366ce9bd3c8f44ca0fbcc1b]
in testcase: trinity
version:
with following parameters:
runtime: 600s
config: i386-randconfig-015-20241208
compiler: gcc-11
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
(please refer to attached dmesg/kmsg for entire log/backtrace)
+-----------------------------------------------------------+------------+------------+
| | 914eec5e98 | 51f43d5d82 |
+-----------------------------------------------------------+------------+------------+
| WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set | 0 | 5 |
| EIP:__alloc_tag_ref_set | 0 | 5 |
+-----------------------------------------------------------+------------+------------+
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
[ 626.127550][ T24] ------------[ cut here ]------------
[ 626.128516][ T24] alloc_tag was not cleared (got tag for mm/shmem.c:1794)
[ 626.136360][ T24] WARNING: CPU: 0 PID: 24 at include/linux/alloc_tag.h:138 __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.137957][ T24] Modules linked in: serio_raw rtc_cmos bochs drm_client_lib drm_shmem_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm_kms_helper fb drm drm_panel_orientation_quirks backlight intel_agp intel_gtt
[ 626.140833][ T24] CPU: 0 UID: 0 PID: 24 Comm: kcompactd0 Tainted: G T 6.13.0-rc1-00015-g51f43d5d82ed #1
[ 626.142284][ T24] Tainted: [T]=RANDSTRUCT
[ 626.142870][ T24] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[ 626.144610][ T24] EIP: __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.145517][ T24] Code: 53 89 c3 51 8b 00 85 c0 74 25 80 3d a4 8a 35 c2 00 75 1c ff 70 04 c6 05 a4 8a 35 c2 01 ff 70 10 68 24 bc c0 c1 e8 7d bf cd ff <0f> 0b 83 c4 0c 85 f6 75 25 a0 a3 8a 35 c2 84 c0 75 22 68 6d bc c0
All code
========
0: 53 push %rbx
1: 89 c3 mov %eax,%ebx
3: 51 push %rcx
4: 8b 00 mov (%rax),%eax
6: 85 c0 test %eax,%eax
8: 74 25 je 0x2f
a: 80 3d a4 8a 35 c2 00 cmpb $0x0,-0x3dca755c(%rip) # 0xffffffffc2358ab5
11: 75 1c jne 0x2f
13: ff 70 04 push 0x4(%rax)
16: c6 05 a4 8a 35 c2 01 movb $0x1,-0x3dca755c(%rip) # 0xffffffffc2358ac1
1d: ff 70 10 push 0x10(%rax)
20: 68 24 bc c0 c1 push $0xffffffffc1c0bc24
25: e8 7d bf cd ff call 0xffffffffffcdbfa7
2a:* 0f 0b ud2 <-- trapping instruction
2c: 83 c4 0c add $0xc,%esp
2f: 85 f6 test %esi,%esi
31: 75 25 jne 0x58
33: a0 a3 8a 35 c2 84 c0 movabs 0x2275c084c2358aa3,%al
3a: 75 22
3c: 68 .byte 0x68
3d: 6d insl (%dx),%es:(%rdi)
3e: bc .byte 0xbc
3f: c0 .byte 0xc0
Code starting with the faulting instruction
===========================================
0: 0f 0b ud2
2: 83 c4 0c add $0xc,%esp
5: 85 f6 test %esi,%esi
7: 75 25 jne 0x2e
9: a0 a3 8a 35 c2 84 c0 movabs 0x2275c084c2358aa3,%al
10: 75 22
12: 68 .byte 0x68
13: 6d insl (%dx),%es:(%rdi)
14: bc .byte 0xbc
15: c0 .byte 0xc0
[ 626.148477][ T24] EAX: 00000037 EBX: c37e1cd0 ECX: 00000000 EDX: 00000002
[ 626.149410][ T24] ESI: c2364ee0 EDI: e8359650 EBP: c37e1cbc ESP: c37e1ca4
[ 626.150473][ T24] DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010246
[ 626.151777][ T24] CR0: 80050033 CR2: 08ad2b00 CR3: 2e405000 CR4: 000406d0
[ 626.152802][ T24] Call Trace:
[ 626.153277][ T24] ? show_regs (arch/x86/kernel/dumpstack.c:478)
[ 626.153821][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.154506][ T24] ? __warn (kernel/panic.c:748)
[ 626.155033][ T24] ? report_bug (lib/bug.c:201 lib/bug.c:219)
[ 626.155645][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.156293][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
[ 626.156881][ T24] ? handle_bug (arch/x86/kernel/traps.c:285)
[ 626.157503][ T24] ? exc_invalid_op (arch/x86/kernel/traps.c:309 (discriminator 1))
[ 626.158156][ T24] ? handle_exception (arch/x86/entry/entry_32.S:1055)
[ 626.159050][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
[ 626.159866][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.160591][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
[ 626.161157][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
[ 626.161831][ T24] ? pgalloc_tag_get (include/linux/pgalloc_tag.h:220)
[ 626.162455][ T24] pgalloc_tag_swap (lib/alloc_tag.c:214)
[ 626.163139][ T24] folio_migrate_flags (mm/migrate.c:750)
[ 626.163986][ T24] __migrate_folio+0x8c/0x96
[ 626.164956][ T24] ? __migrate_folio+0x96/0x96
[ 626.165986][ T24] migrate_folio (mm/migrate.c:800 (discriminator 2))
[ 626.166790][ T24] move_to_new_folio (mm/migrate.c:1060)
[ 626.167675][ T24] migrate_pages_batch (mm/migrate.c:1369 mm/migrate.c:1899)
[ 626.168566][ T24] ? list_add (arch/x86/kernel/cpu/resctrl/rdtgroup.c:2015 (discriminator 2))
[ 626.169303][ T24] migrate_pages (mm/migrate.c:1971 mm/migrate.c:2074)
[ 626.169948][ T24] ? list_add (arch/x86/kernel/cpu/resctrl/rdtgroup.c:2015 (discriminator 2))
[ 626.170574][ T24] ? fragmentation_score_node (mm/compaction.c:1879)
[ 626.171569][ T24] compact_zone (mm/compaction.c:2641)
[ 626.172363][ T24] compact_node (mm/compaction.c:2912)
[ 626.173081][ T24] kcompactd (mm/compaction.c:3209)
[ 626.173630][ T24] ? lockdep_assert_rq_held (kernel/sched/sched.h:1731)
[ 626.174318][ T24] kthread (kernel/kthread.c:391)
[ 626.174972][ T24] ? kcompactd_do_work (mm/compaction.c:3155)
[ 626.175708][ T24] ? list_del_init (kernel/signal.c:466)
[ 626.176296][ T24] ret_from_fork (arch/x86/kernel/process.c:153)
[ 626.176833][ T24] ? list_del_init (kernel/signal.c:466)
[ 626.177415][ T24] ret_from_fork_asm (arch/x86/entry/entry_32.S:737)
[ 626.177999][ T24] entry_INT80_32 (arch/x86/entry/entry_32.S:945)
[ 626.178652][ T24] irq event stamp: 18033
[ 626.179292][ T24] hardirqs last enabled at (18043): __up_console_sem (arch/x86/include/asm/irqflags.h:26 (discriminator 3) arch/x86/include/asm/irqflags.h:87 (discriminator 3) arch/x86/include/asm/irqflags.h:147 (discriminator 3) kernel/printk/printk.c:344 (discriminator 3))
[ 626.180568][ T24] hardirqs last disabled at (18052): __up_console_sem (kernel/printk/printk.c:342 (discriminator 3))
[ 626.181908][ T24] softirqs last enabled at (17870): handle_softirqs (kernel/softirq.c:401 kernel/softirq.c:582)
[ 626.183280][ T24] softirqs last disabled at (17865): __do_softirq (kernel/softirq.c:589)
[ 626.184539][ T24] ---[ end trace 0000000000000000 ]---
[ 626.460849][ T3069] trinity-c1 invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=500
[ 626.462854][ T3069] CPU: 0 UID: 65534 PID: 3069 Comm: trinity-c1 Tainted: G W T 6.13.0-rc1-00015-g51f43d5d82ed #1
[ 626.464654][ T3069] Tainted: [W]=WARN, [T]=RANDSTRUCT
[ 626.465456][ T3069] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[ 626.467024][ T3069] Call Trace:
[ 626.467563][ T3069] dump_stack_lvl (lib/dump_stack.c:122 (discriminator 4))
[ 626.468317][ T3069] dump_stack (lib/dump_stack.c:130)
[ 626.468989][ T3069] dump_header (mm/oom_kill.c:73 mm/oom_kill.c:462)
[ 626.469701][ T3069] oom_kill_process (mm/oom_kill.c:444 mm/oom_kill.c:1035)
[ 626.470501][ T3069] out_of_memory (mm/oom_kill.c:1174)
[ 626.471229][ T3069] __alloc_pages_slowpath+0x584/0x63d
[ 626.472253][ T3069] __alloc_pages_noprof (mm/page_alloc.c:4764)
[ 626.473076][ T3069] __folio_alloc_noprof (mm/internal.h:709 mm/page_alloc.c:4785)
[ 626.473927][ T3069] shmem_alloc_folio+0x21/0x48
[ 626.474891][ T3069] shmem_alloc_and_add_folio+0x22/0x138
[ 626.475933][ T3069] shmem_get_folio_gfp+0x164/0x358
[ 626.476877][ T3069] ? kunmap_local_indexed (mm/highmem.c:630 (discriminator 3))
[ 626.477692][ T3069] shmem_get_folio (mm/shmem.c:2463)
[ 626.478404][ T3069] shmem_write_begin (mm/shmem.c:3118)
[ 626.479168][ T3069] generic_perform_write (mm/filemap.c:4057)
[ 626.479993][ T3069] shmem_file_write_iter (mm/shmem.c:3293)
[ 626.480869][ T3069] iter_file_splice_write (fs/splice.c:744)
[ 626.481773][ T3069] ? splice_from_pipe_next (fs/splice.c:669)
[ 626.482677][ T3069] do_splice_from (fs/splice.c:941)
[ 626.483432][ T3069] direct_splice_actor (fs/splice.c:1164)
[ 626.484239][ T3069] splice_direct_to_actor (fs/splice.c:1109)
[ 626.485142][ T3069] ? file_end_write (fs/read_write.c:843)
[ 626.485925][ T3069] do_splice_direct_actor (fs/splice.c:1208)
[ 626.486772][ T3069] ? pipe_buf_confirm (fs/pipe.c:85)
[ 626.487522][ T3069] do_splice_direct (fs/splice.c:1235)
[ 626.488317][ T3069] ? file_end_write (fs/read_write.c:843)
[ 626.489088][ T3069] do_sendfile (fs/read_write.c:1363)
[ 626.489832][ T3069] __ia32_sys_sendfile (fs/read_write.c:1402 fs/read_write.c:1391 fs/read_write.c:1391)
[ 626.490659][ T3069] ia32_sys_call (kbuild/obj/consumer/i386-randconfig-015-20241208/./arch/x86/include/generated/asm/syscalls_32.h:188)
[ 626.491455][ T3069] do_int80_syscall_32 (arch/x86/entry/common.c:165 arch/x86/entry/common.c:339)
[ 626.492222][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
[ 626.493011][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
[ 626.493775][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
[ 626.494583][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
[ 626.495431][ T3069] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4470)
[ 626.496269][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
[ 626.497098][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
[ 626.497903][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
[ 626.498733][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
[ 626.499603][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
[ 626.500462][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
[ 626.501217][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
[ 626.502080][ T3069] ? find_held_lock+0x22/0x5f
[ 626.503012][ T3069] ? __lock_release+0xb0/0x150
[ 626.503862][ T3069] ? rcu_read_unlock (include/linux/rcupdate.h:347 include/linux/rcupdate.h:880)
[ 626.504608][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
[ 626.505449][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
[ 626.506281][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
[ 626.507051][ T3069] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4470)
[ 626.507856][ T3069] ? syscall_exit_to_user_mode (kernel/entry/common.c:221)
[ 626.508762][ T3069] ? do_int80_syscall_32 (arch/x86/entry/common.c:343)
[ 626.509576][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
[ 626.510425][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
[ 626.511252][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
[ 626.512110][ T3069] ? find_held_lock+0x22/0x5f
The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20241211/202412112227.df61ebb-lkp@intel.com
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re:[linus:master] [mm/codetag] 51f43d5d82: WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set
2024-12-11 15:08 [linus:master] [mm/codetag] 51f43d5d82: WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set kernel test robot
@ 2024-12-12 2:12 ` David Wang
2024-12-12 4:01 ` [PATCH] mm/codetag: clear tags before swap David Wang
1 sibling, 0 replies; 10+ messages in thread
From: David Wang @ 2024-12-12 2:12 UTC (permalink / raw)
To: kernel test robot, Suren Baghdasaryan
Cc: oe-lkp, lkp, linux-kernel, Andrew Morton, Yu Zhao,
Kent Overstreet, linux-mm
Hi,
Thanks for reporting this.
This happens when CONFIG_MEM_ALLOC_PROFILING_DEBUG=y, I will send out a patch later.
Thanks~
David
At 2024-12-11 23:08:46, "kernel test robot" <oliver.sang@intel.com> wrote:
>
>
>Hello,
>
>kernel test robot noticed "WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set" on:
>
>commit: 51f43d5d82ed2ba3f9a3f9a2390c52f28e42af32 ("mm/codetag: swap tags when migrate pages")
>https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
>
>[test failed on linus/master fac04efc5c793dccbd07e2d59af9f90b7fc0dca4]
>[test failed on linux-next/master ebe1b11614e079c5e366ce9bd3c8f44ca0fbcc1b]
>
>in testcase: trinity
>version:
>with following parameters:
>
> runtime: 600s
>
>
>
>config: i386-randconfig-015-20241208
>compiler: gcc-11
>test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
>
>(please refer to attached dmesg/kmsg for entire log/backtrace)
>
>
>+-----------------------------------------------------------+------------+------------+
>| | 914eec5e98 | 51f43d5d82 |
>+-----------------------------------------------------------+------------+------------+
>| WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set | 0 | 5 |
>| EIP:__alloc_tag_ref_set | 0 | 5 |
>+-----------------------------------------------------------+------------+------------+
>
>
>If you fix the issue in a separate patch/commit (i.e. not just a new version of
>the same patch/commit), kindly add following tags
>| Reported-by: kernel test robot <oliver.sang@intel.com>
>| Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
>
>
>[ 626.127550][ T24] ------------[ cut here ]------------
>[ 626.128516][ T24] alloc_tag was not cleared (got tag for mm/shmem.c:1794)
>[ 626.136360][ T24] WARNING: CPU: 0 PID: 24 at include/linux/alloc_tag.h:138 __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.137957][ T24] Modules linked in: serio_raw rtc_cmos bochs drm_client_lib drm_shmem_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm_kms_helper fb drm drm_panel_orientation_quirks backlight intel_agp intel_gtt
>[ 626.140833][ T24] CPU: 0 UID: 0 PID: 24 Comm: kcompactd0 Tainted: G T 6.13.0-rc1-00015-g51f43d5d82ed #1
>[ 626.142284][ T24] Tainted: [T]=RANDSTRUCT
>[ 626.142870][ T24] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
>[ 626.144610][ T24] EIP: __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.145517][ T24] Code: 53 89 c3 51 8b 00 85 c0 74 25 80 3d a4 8a 35 c2 00 75 1c ff 70 04 c6 05 a4 8a 35 c2 01 ff 70 10 68 24 bc c0 c1 e8 7d bf cd ff <0f> 0b 83 c4 0c 85 f6 75 25 a0 a3 8a 35 c2 84 c0 75 22 68 6d bc c0
>All code
>========
> 0: 53 push %rbx
> 1: 89 c3 mov %eax,%ebx
> 3: 51 push %rcx
> 4: 8b 00 mov (%rax),%eax
> 6: 85 c0 test %eax,%eax
> 8: 74 25 je 0x2f
> a: 80 3d a4 8a 35 c2 00 cmpb $0x0,-0x3dca755c(%rip) # 0xffffffffc2358ab5
> 11: 75 1c jne 0x2f
> 13: ff 70 04 push 0x4(%rax)
> 16: c6 05 a4 8a 35 c2 01 movb $0x1,-0x3dca755c(%rip) # 0xffffffffc2358ac1
> 1d: ff 70 10 push 0x10(%rax)
> 20: 68 24 bc c0 c1 push $0xffffffffc1c0bc24
> 25: e8 7d bf cd ff call 0xffffffffffcdbfa7
> 2a:* 0f 0b ud2 <-- trapping instruction
> 2c: 83 c4 0c add $0xc,%esp
> 2f: 85 f6 test %esi,%esi
> 31: 75 25 jne 0x58
> 33: a0 a3 8a 35 c2 84 c0 movabs 0x2275c084c2358aa3,%al
> 3a: 75 22
> 3c: 68 .byte 0x68
> 3d: 6d insl (%dx),%es:(%rdi)
> 3e: bc .byte 0xbc
> 3f: c0 .byte 0xc0
>
>Code starting with the faulting instruction
>===========================================
> 0: 0f 0b ud2
> 2: 83 c4 0c add $0xc,%esp
> 5: 85 f6 test %esi,%esi
> 7: 75 25 jne 0x2e
> 9: a0 a3 8a 35 c2 84 c0 movabs 0x2275c084c2358aa3,%al
> 10: 75 22
> 12: 68 .byte 0x68
> 13: 6d insl (%dx),%es:(%rdi)
> 14: bc .byte 0xbc
> 15: c0 .byte 0xc0
>[ 626.148477][ T24] EAX: 00000037 EBX: c37e1cd0 ECX: 00000000 EDX: 00000002
>[ 626.149410][ T24] ESI: c2364ee0 EDI: e8359650 EBP: c37e1cbc ESP: c37e1ca4
>[ 626.150473][ T24] DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 EFLAGS: 00010246
>[ 626.151777][ T24] CR0: 80050033 CR2: 08ad2b00 CR3: 2e405000 CR4: 000406d0
>[ 626.152802][ T24] Call Trace:
>[ 626.153277][ T24] ? show_regs (arch/x86/kernel/dumpstack.c:478)
>[ 626.153821][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.154506][ T24] ? __warn (kernel/panic.c:748)
>[ 626.155033][ T24] ? report_bug (lib/bug.c:201 lib/bug.c:219)
>[ 626.155645][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.156293][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
>[ 626.156881][ T24] ? handle_bug (arch/x86/kernel/traps.c:285)
>[ 626.157503][ T24] ? exc_invalid_op (arch/x86/kernel/traps.c:309 (discriminator 1))
>[ 626.158156][ T24] ? handle_exception (arch/x86/entry/entry_32.S:1055)
>[ 626.159050][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
>[ 626.159866][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.160591][ T24] ? exc_overflow (arch/x86/kernel/traps.c:301)
>[ 626.161157][ T24] ? __alloc_tag_ref_set (include/linux/alloc_tag.h:138 include/linux/alloc_tag.h:157)
>[ 626.161831][ T24] ? pgalloc_tag_get (include/linux/pgalloc_tag.h:220)
>[ 626.162455][ T24] pgalloc_tag_swap (lib/alloc_tag.c:214)
>[ 626.163139][ T24] folio_migrate_flags (mm/migrate.c:750)
>[ 626.163986][ T24] __migrate_folio+0x8c/0x96
>[ 626.164956][ T24] ? __migrate_folio+0x96/0x96
>[ 626.165986][ T24] migrate_folio (mm/migrate.c:800 (discriminator 2))
>[ 626.166790][ T24] move_to_new_folio (mm/migrate.c:1060)
>[ 626.167675][ T24] migrate_pages_batch (mm/migrate.c:1369 mm/migrate.c:1899)
>[ 626.168566][ T24] ? list_add (arch/x86/kernel/cpu/resctrl/rdtgroup.c:2015 (discriminator 2))
>[ 626.169303][ T24] migrate_pages (mm/migrate.c:1971 mm/migrate.c:2074)
>[ 626.169948][ T24] ? list_add (arch/x86/kernel/cpu/resctrl/rdtgroup.c:2015 (discriminator 2))
>[ 626.170574][ T24] ? fragmentation_score_node (mm/compaction.c:1879)
>[ 626.171569][ T24] compact_zone (mm/compaction.c:2641)
>[ 626.172363][ T24] compact_node (mm/compaction.c:2912)
>[ 626.173081][ T24] kcompactd (mm/compaction.c:3209)
>[ 626.173630][ T24] ? lockdep_assert_rq_held (kernel/sched/sched.h:1731)
>[ 626.174318][ T24] kthread (kernel/kthread.c:391)
>[ 626.174972][ T24] ? kcompactd_do_work (mm/compaction.c:3155)
>[ 626.175708][ T24] ? list_del_init (kernel/signal.c:466)
>[ 626.176296][ T24] ret_from_fork (arch/x86/kernel/process.c:153)
>[ 626.176833][ T24] ? list_del_init (kernel/signal.c:466)
>[ 626.177415][ T24] ret_from_fork_asm (arch/x86/entry/entry_32.S:737)
>[ 626.177999][ T24] entry_INT80_32 (arch/x86/entry/entry_32.S:945)
>[ 626.178652][ T24] irq event stamp: 18033
>[ 626.179292][ T24] hardirqs last enabled at (18043): __up_console_sem (arch/x86/include/asm/irqflags.h:26 (discriminator 3) arch/x86/include/asm/irqflags.h:87 (discriminator 3) arch/x86/include/asm/irqflags.h:147 (discriminator 3) kernel/printk/printk.c:344 (discriminator 3))
>[ 626.180568][ T24] hardirqs last disabled at (18052): __up_console_sem (kernel/printk/printk.c:342 (discriminator 3))
>[ 626.181908][ T24] softirqs last enabled at (17870): handle_softirqs (kernel/softirq.c:401 kernel/softirq.c:582)
>[ 626.183280][ T24] softirqs last disabled at (17865): __do_softirq (kernel/softirq.c:589)
>[ 626.184539][ T24] ---[ end trace 0000000000000000 ]---
>[ 626.460849][ T3069] trinity-c1 invoked oom-killer: gfp_mask=0x140cca(GFP_HIGHUSER_MOVABLE|__GFP_COMP), order=0, oom_score_adj=500
>[ 626.462854][ T3069] CPU: 0 UID: 65534 PID: 3069 Comm: trinity-c1 Tainted: G W T 6.13.0-rc1-00015-g51f43d5d82ed #1
>[ 626.464654][ T3069] Tainted: [W]=WARN, [T]=RANDSTRUCT
>[ 626.465456][ T3069] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
>[ 626.467024][ T3069] Call Trace:
>[ 626.467563][ T3069] dump_stack_lvl (lib/dump_stack.c:122 (discriminator 4))
>[ 626.468317][ T3069] dump_stack (lib/dump_stack.c:130)
>[ 626.468989][ T3069] dump_header (mm/oom_kill.c:73 mm/oom_kill.c:462)
>[ 626.469701][ T3069] oom_kill_process (mm/oom_kill.c:444 mm/oom_kill.c:1035)
>[ 626.470501][ T3069] out_of_memory (mm/oom_kill.c:1174)
>[ 626.471229][ T3069] __alloc_pages_slowpath+0x584/0x63d
>[ 626.472253][ T3069] __alloc_pages_noprof (mm/page_alloc.c:4764)
>[ 626.473076][ T3069] __folio_alloc_noprof (mm/internal.h:709 mm/page_alloc.c:4785)
>[ 626.473927][ T3069] shmem_alloc_folio+0x21/0x48
>[ 626.474891][ T3069] shmem_alloc_and_add_folio+0x22/0x138
>[ 626.475933][ T3069] shmem_get_folio_gfp+0x164/0x358
>[ 626.476877][ T3069] ? kunmap_local_indexed (mm/highmem.c:630 (discriminator 3))
>[ 626.477692][ T3069] shmem_get_folio (mm/shmem.c:2463)
>[ 626.478404][ T3069] shmem_write_begin (mm/shmem.c:3118)
>[ 626.479168][ T3069] generic_perform_write (mm/filemap.c:4057)
>[ 626.479993][ T3069] shmem_file_write_iter (mm/shmem.c:3293)
>[ 626.480869][ T3069] iter_file_splice_write (fs/splice.c:744)
>[ 626.481773][ T3069] ? splice_from_pipe_next (fs/splice.c:669)
>[ 626.482677][ T3069] do_splice_from (fs/splice.c:941)
>[ 626.483432][ T3069] direct_splice_actor (fs/splice.c:1164)
>[ 626.484239][ T3069] splice_direct_to_actor (fs/splice.c:1109)
>[ 626.485142][ T3069] ? file_end_write (fs/read_write.c:843)
>[ 626.485925][ T3069] do_splice_direct_actor (fs/splice.c:1208)
>[ 626.486772][ T3069] ? pipe_buf_confirm (fs/pipe.c:85)
>[ 626.487522][ T3069] do_splice_direct (fs/splice.c:1235)
>[ 626.488317][ T3069] ? file_end_write (fs/read_write.c:843)
>[ 626.489088][ T3069] do_sendfile (fs/read_write.c:1363)
>[ 626.489832][ T3069] __ia32_sys_sendfile (fs/read_write.c:1402 fs/read_write.c:1391 fs/read_write.c:1391)
>[ 626.490659][ T3069] ia32_sys_call (kbuild/obj/consumer/i386-randconfig-015-20241208/./arch/x86/include/generated/asm/syscalls_32.h:188)
>[ 626.491455][ T3069] do_int80_syscall_32 (arch/x86/entry/common.c:165 arch/x86/entry/common.c:339)
>[ 626.492222][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
>[ 626.493011][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
>[ 626.493775][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
>[ 626.494583][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
>[ 626.495431][ T3069] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4470)
>[ 626.496269][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
>[ 626.497098][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
>[ 626.497903][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
>[ 626.498733][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
>[ 626.499603][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
>[ 626.500462][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
>[ 626.501217][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
>[ 626.502080][ T3069] ? find_held_lock+0x22/0x5f
>[ 626.503012][ T3069] ? __lock_release+0xb0/0x150
>[ 626.503862][ T3069] ? rcu_read_unlock (include/linux/rcupdate.h:347 include/linux/rcupdate.h:880)
>[ 626.504608][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
>[ 626.505449][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
>[ 626.506281][ T3069] ? __this_cpu_preempt_check (lib/smp_processor_id.c:67)
>[ 626.507051][ T3069] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4470)
>[ 626.507856][ T3069] ? syscall_exit_to_user_mode (kernel/entry/common.c:221)
>[ 626.508762][ T3069] ? do_int80_syscall_32 (arch/x86/entry/common.c:343)
>[ 626.509576][ T3069] ? kvm_sched_clock_read (arch/x86/kernel/kvmclock.c:91)
>[ 626.510425][ T3069] ? sched_clock_noinstr (arch/x86/kernel/tsc.c:269)
>[ 626.511252][ T3069] ? local_clock_noinstr (kernel/sched/clock.c:301)
>[ 626.512110][ T3069] ? find_held_lock+0x22/0x5f
>
>
>The kernel config and materials to reproduce are available at:
>https://download.01.org/0day-ci/archive/20241211/202412112227.df61ebb-lkp@intel.com
>
>
>
>--
>0-DAY CI Kernel Test Service
>https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH] mm/codetag: clear tags before swap
2024-12-11 15:08 [linus:master] [mm/codetag] 51f43d5d82: WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set kernel test robot
2024-12-12 2:12 ` David Wang
@ 2024-12-12 4:01 ` David Wang
2024-12-12 7:09 ` Suren Baghdasaryan
1 sibling, 1 reply; 10+ messages in thread
From: David Wang @ 2024-12-12 4:01 UTC (permalink / raw)
To: surenb, kent.overstreet
Cc: yuzhao, oliver.sang, akpm, linux-kernel, linux-mm, lkp, oe-lkp,
David Wang
When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
triggered when calling __alloc_tag_ref_set() during swap:
alloc_tag was not cleared (got tag for mm/filemap.c:1951)
WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
Clear code tags before swap can fix the warning. And this patch also fix
a potential invalid address dereference in alloc_tag_add_check() when
CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
which is defined as ((void *)1).
Signed-off-by: David Wang <00107082@163.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
---
include/linux/alloc_tag.h | 2 +-
lib/alloc_tag.c | 4 ++++
2 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
index 7c0786bdf9af..cba024bf2db3 100644
--- a/include/linux/alloc_tag.h
+++ b/include/linux/alloc_tag.h
@@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
#ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
{
- WARN_ONCE(ref && ref->ct,
+ WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
"alloc_tag was not cleared (got tag for %s:%u)\n",
ref->ct->filename, ref->ct->lineno);
diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
index 35f7560a309a..cc5fda9901c2 100644
--- a/lib/alloc_tag.c
+++ b/lib/alloc_tag.c
@@ -209,6 +209,10 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
return;
}
+ /* clear tags before swap */
+ set_codetag_empty(&ref_old);
+ set_codetag_empty(&ref_new);
+
/* swap tags */
__alloc_tag_ref_set(&ref_old, tag_new);
update_page_tag_ref(handle_old, &ref_old);
--
2.39.2
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] mm/codetag: clear tags before swap
2024-12-12 4:01 ` [PATCH] mm/codetag: clear tags before swap David Wang
@ 2024-12-12 7:09 ` Suren Baghdasaryan
2024-12-12 8:17 ` David Wang
2024-12-12 8:29 ` [PATCH v2] " David Wang
0 siblings, 2 replies; 10+ messages in thread
From: Suren Baghdasaryan @ 2024-12-12 7:09 UTC (permalink / raw)
To: David Wang
Cc: kent.overstreet, yuzhao, oliver.sang, akpm, linux-kernel,
linux-mm, lkp, oe-lkp
On Wed, Dec 11, 2024 at 8:03 PM David Wang <00107082@163.com> wrote:
>
> When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
> triggered when calling __alloc_tag_ref_set() during swap:
>
> alloc_tag was not cleared (got tag for mm/filemap.c:1951)
> WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
>
> Clear code tags before swap can fix the warning. And this patch also fix
> a potential invalid address dereference in alloc_tag_add_check() when
> CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
> which is defined as ((void *)1).
^^^
Good catch!
>
> Signed-off-by: David Wang <00107082@163.com>
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
> ---
> include/linux/alloc_tag.h | 2 +-
> lib/alloc_tag.c | 4 ++++
> 2 files changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
> index 7c0786bdf9af..cba024bf2db3 100644
> --- a/include/linux/alloc_tag.h
> +++ b/include/linux/alloc_tag.h
> @@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
> #ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
> static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
> {
> - WARN_ONCE(ref && ref->ct,
> + WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
> "alloc_tag was not cleared (got tag for %s:%u)\n",
> ref->ct->filename, ref->ct->lineno);
>
> diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
> index 35f7560a309a..cc5fda9901c2 100644
> --- a/lib/alloc_tag.c
> +++ b/lib/alloc_tag.c
> @@ -209,6 +209,10 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
> return;
> }
>
> + /* clear tags before swap */
The above comment states what we already know from the code but does
not explain why we do this. Better to describe the reason and not what
we do. Something like:
/*
* Clear tag references to avoid debug warning when using
* __alloc_tag_ref_set() with non-empty reference.
*/
> + set_codetag_empty(&ref_old);
> + set_codetag_empty(&ref_new);
> +
> /* swap tags */
> __alloc_tag_ref_set(&ref_old, tag_new);
> update_page_tag_ref(handle_old, &ref_old);
> --
> 2.39.2
>
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH] mm/codetag: clear tags before swap
2024-12-12 7:09 ` Suren Baghdasaryan
@ 2024-12-12 8:17 ` David Wang
2024-12-12 8:29 ` [PATCH v2] " David Wang
1 sibling, 0 replies; 10+ messages in thread
From: David Wang @ 2024-12-12 8:17 UTC (permalink / raw)
To: Suren Baghdasaryan
Cc: kent.overstreet, yuzhao, oliver.sang, akpm, linux-kernel,
linux-mm, lkp, oe-lkp
At 2024-12-12 15:09:59, "Suren Baghdasaryan" <surenb@google.com> wrote:
>On Wed, Dec 11, 2024 at 8:03 PM David Wang <00107082@163.com> wrote:
>>
>> When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
>> triggered when calling __alloc_tag_ref_set() during swap:
>>
>> alloc_tag was not cleared (got tag for mm/filemap.c:1951)
>> WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
>>
>> Clear code tags before swap can fix the warning. And this patch also fix
>> a potential invalid address dereference in alloc_tag_add_check() when
>> CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
>> which is defined as ((void *)1).
>^^^
>Good catch!
>
>>
>> Signed-off-by: David Wang <00107082@163.com>
>> Reported-by: kernel test robot <oliver.sang@intel.com>
>> Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
>> ---
>> include/linux/alloc_tag.h | 2 +-
>> lib/alloc_tag.c | 4 ++++
>> 2 files changed, 5 insertions(+), 1 deletion(-)
>>
>> diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
>> index 7c0786bdf9af..cba024bf2db3 100644
>> --- a/include/linux/alloc_tag.h
>> +++ b/include/linux/alloc_tag.h
>> @@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
>> #ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
>> static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
>> {
>> - WARN_ONCE(ref && ref->ct,
>> + WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
>> "alloc_tag was not cleared (got tag for %s:%u)\n",
>> ref->ct->filename, ref->ct->lineno);
>>
>> diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
>> index 35f7560a309a..cc5fda9901c2 100644
>> --- a/lib/alloc_tag.c
>> +++ b/lib/alloc_tag.c
>> @@ -209,6 +209,10 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
>> return;
>> }
>>
>> + /* clear tags before swap */
>
>The above comment states what we already know from the code but does
>not explain why we do this. Better to describe the reason and not what
>we do. Something like:
>
>/*
> * Clear tag references to avoid debug warning when using
> * __alloc_tag_ref_set() with non-empty reference.
> */
>
Copy that~!
Thanks!
David
>> + set_codetag_empty(&ref_old);
>> + set_codetag_empty(&ref_new);
>> +
>> /* swap tags */
>> __alloc_tag_ref_set(&ref_old, tag_new);
>> update_page_tag_ref(handle_old, &ref_old);
>> --
>> 2.39.2
>>
>>
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v2] mm/codetag: clear tags before swap
2024-12-12 7:09 ` Suren Baghdasaryan
2024-12-12 8:17 ` David Wang
@ 2024-12-12 8:29 ` David Wang
2024-12-12 23:18 ` Suren Baghdasaryan
1 sibling, 1 reply; 10+ messages in thread
From: David Wang @ 2024-12-12 8:29 UTC (permalink / raw)
To: surenb
Cc: akpm, kent.overstreet, linux-kernel, linux-mm, lkp, oe-lkp,
oliver.sang, yuzhao, David Wang
When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
triggered when calling __alloc_tag_ref_set() during swap:
alloc_tag was not cleared (got tag for mm/filemap.c:1951)
WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
Clear code tags before swap can fix the warning. And this patch also fix
a potential invalid address dereference in alloc_tag_add_check() when
CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
which is defined as ((void *)1).
Signed-off-by: David Wang <00107082@163.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
Suggested-by: Suren Baghdasaryan <surenb@google.com>
---
include/linux/alloc_tag.h | 2 +-
lib/alloc_tag.c | 7 +++++++
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
index 7c0786bdf9af..cba024bf2db3 100644
--- a/include/linux/alloc_tag.h
+++ b/include/linux/alloc_tag.h
@@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
#ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
{
- WARN_ONCE(ref && ref->ct,
+ WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
"alloc_tag was not cleared (got tag for %s:%u)\n",
ref->ct->filename, ref->ct->lineno);
diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
index 35f7560a309a..3a0413462e9f 100644
--- a/lib/alloc_tag.c
+++ b/lib/alloc_tag.c
@@ -209,6 +209,13 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
return;
}
+ /*
+ * Clear tag references to avoid debug warning when using
+ * __alloc_tag_ref_set() with non-empty reference.
+ */
+ set_codetag_empty(&ref_old);
+ set_codetag_empty(&ref_new);
+
/* swap tags */
__alloc_tag_ref_set(&ref_old, tag_new);
update_page_tag_ref(handle_old, &ref_old);
--
2.39.2
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v2] mm/codetag: clear tags before swap
2024-12-12 8:29 ` [PATCH v2] " David Wang
@ 2024-12-12 23:18 ` Suren Baghdasaryan
2024-12-13 1:33 ` [PATCH v3] " David Wang
0 siblings, 1 reply; 10+ messages in thread
From: Suren Baghdasaryan @ 2024-12-12 23:18 UTC (permalink / raw)
To: David Wang
Cc: akpm, kent.overstreet, linux-kernel, linux-mm, lkp, oe-lkp,
oliver.sang, yuzhao
On Thu, Dec 12, 2024 at 12:29 AM David Wang <00107082@163.com> wrote:
>
> When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
> triggered when calling __alloc_tag_ref_set() during swap:
>
> alloc_tag was not cleared (got tag for mm/filemap.c:1951)
> WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
>
> Clear code tags before swap can fix the warning. And this patch also fix
> a potential invalid address dereference in alloc_tag_add_check() when
> CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
> which is defined as ((void *)1).
>
> Signed-off-by: David Wang <00107082@163.com>
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
> Suggested-by: Suren Baghdasaryan <surenb@google.com>
I didn't really suggest much in this patch, so please replace above
Suggested-by with:
Acked-by: Suren Baghdasaryan <surenb@google.com>
Thanks for fixing this!
> ---
> include/linux/alloc_tag.h | 2 +-
> lib/alloc_tag.c | 7 +++++++
> 2 files changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
> index 7c0786bdf9af..cba024bf2db3 100644
> --- a/include/linux/alloc_tag.h
> +++ b/include/linux/alloc_tag.h
> @@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
> #ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
> static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
> {
> - WARN_ONCE(ref && ref->ct,
> + WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
> "alloc_tag was not cleared (got tag for %s:%u)\n",
> ref->ct->filename, ref->ct->lineno);
>
> diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
> index 35f7560a309a..3a0413462e9f 100644
> --- a/lib/alloc_tag.c
> +++ b/lib/alloc_tag.c
> @@ -209,6 +209,13 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
> return;
> }
>
> + /*
> + * Clear tag references to avoid debug warning when using
> + * __alloc_tag_ref_set() with non-empty reference.
> + */
> + set_codetag_empty(&ref_old);
> + set_codetag_empty(&ref_new);
> +
> /* swap tags */
> __alloc_tag_ref_set(&ref_old, tag_new);
> update_page_tag_ref(handle_old, &ref_old);
> --
> 2.39.2
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v3] mm/codetag: clear tags before swap
2024-12-12 23:18 ` Suren Baghdasaryan
@ 2024-12-13 1:33 ` David Wang
2024-12-13 4:12 ` Andrew Morton
0 siblings, 1 reply; 10+ messages in thread
From: David Wang @ 2024-12-13 1:33 UTC (permalink / raw)
To: surenb
Cc: akpm, kent.overstreet, linux-kernel, linux-mm, lkp, oe-lkp,
oliver.sang, yuzhao, David Wang
When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
triggered when calling __alloc_tag_ref_set() during swap:
alloc_tag was not cleared (got tag for mm/filemap.c:1951)
WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
Clear code tags before swap can fix the warning. And this patch also fix
a potential invalid address dereference in alloc_tag_add_check() when
CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
which is defined as ((void *)1).
Signed-off-by: David Wang <00107082@163.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
Acked-by: Suren Baghdasaryan <surenb@google.com>
---
include/linux/alloc_tag.h | 2 +-
lib/alloc_tag.c | 7 +++++++
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/include/linux/alloc_tag.h b/include/linux/alloc_tag.h
index 7c0786bdf9af..cba024bf2db3 100644
--- a/include/linux/alloc_tag.h
+++ b/include/linux/alloc_tag.h
@@ -135,7 +135,7 @@ static inline struct alloc_tag_counters alloc_tag_read(struct alloc_tag *tag)
#ifdef CONFIG_MEM_ALLOC_PROFILING_DEBUG
static inline void alloc_tag_add_check(union codetag_ref *ref, struct alloc_tag *tag)
{
- WARN_ONCE(ref && ref->ct,
+ WARN_ONCE(ref && ref->ct && !is_codetag_empty(ref),
"alloc_tag was not cleared (got tag for %s:%u)\n",
ref->ct->filename, ref->ct->lineno);
diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
index 35f7560a309a..3a0413462e9f 100644
--- a/lib/alloc_tag.c
+++ b/lib/alloc_tag.c
@@ -209,6 +209,13 @@ void pgalloc_tag_swap(struct folio *new, struct folio *old)
return;
}
+ /*
+ * Clear tag references to avoid debug warning when using
+ * __alloc_tag_ref_set() with non-empty reference.
+ */
+ set_codetag_empty(&ref_old);
+ set_codetag_empty(&ref_new);
+
/* swap tags */
__alloc_tag_ref_set(&ref_old, tag_new);
update_page_tag_ref(handle_old, &ref_old);
--
2.39.2
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3] mm/codetag: clear tags before swap
2024-12-13 1:33 ` [PATCH v3] " David Wang
@ 2024-12-13 4:12 ` Andrew Morton
2024-12-13 4:22 ` Suren Baghdasaryan
0 siblings, 1 reply; 10+ messages in thread
From: Andrew Morton @ 2024-12-13 4:12 UTC (permalink / raw)
To: David Wang
Cc: surenb, kent.overstreet, linux-kernel, linux-mm, lkp, oe-lkp,
oliver.sang, yuzhao
On Fri, 13 Dec 2024 09:33:32 +0800 David Wang <00107082@163.com> wrote:
> When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
> triggered when calling __alloc_tag_ref_set() during swap:
>
> alloc_tag was not cleared (got tag for mm/filemap.c:1951)
> WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
>
> Clear code tags before swap can fix the warning. And this patch also fix
> a potential invalid address dereference in alloc_tag_add_check() when
> CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
> which is defined as ((void *)1).
>
> Signed-off-by: David Wang <00107082@163.com>
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
This points at
51f43d5d82ed ("mm/codetag: swap tags when migrate pages"), which had
Fixes: e0a955bf7f61 ("mm/codetag: add pgalloc_tag_copy()")
e0a955bf7f61 ("mm/codetag: add pgalloc_tag_copy()") had
Fixes: dcfe378c81f7 ("lib: introduce support for page allocation tagging")
Cc: <stable@vger.kernel.org>
And I'm thinking that this fix should have
Fixes: 51f43d5d82ed ("mm/codetag: swap tags when migrate pages")
Cc: <stable@vger.kernel.org>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3] mm/codetag: clear tags before swap
2024-12-13 4:12 ` Andrew Morton
@ 2024-12-13 4:22 ` Suren Baghdasaryan
0 siblings, 0 replies; 10+ messages in thread
From: Suren Baghdasaryan @ 2024-12-13 4:22 UTC (permalink / raw)
To: Andrew Morton
Cc: David Wang, kent.overstreet, linux-kernel, linux-mm, lkp, oe-lkp,
oliver.sang, yuzhao
On Thu, Dec 12, 2024 at 8:12 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Fri, 13 Dec 2024 09:33:32 +0800 David Wang <00107082@163.com> wrote:
>
> > When CONFIG_MEM_ALLOC_PROFILING_DEBUG is set, kernel WARN would be
> > triggered when calling __alloc_tag_ref_set() during swap:
> >
> > alloc_tag was not cleared (got tag for mm/filemap.c:1951)
> > WARNING: CPU: 0 PID: 816 at ./include/linux/alloc_tag.h...
> >
> > Clear code tags before swap can fix the warning. And this patch also fix
> > a potential invalid address dereference in alloc_tag_add_check() when
> > CONFIG_MEM_ALLOC_PROFILING_DEBUG is set and ref->ct is CODETAG_EMPTY,
> > which is defined as ((void *)1).
> >
> > Signed-off-by: David Wang <00107082@163.com>
> > Reported-by: kernel test robot <oliver.sang@intel.com>
> > Closes: https://lore.kernel.org/oe-lkp/202412112227.df61ebb-lkp@intel.com
>
> This points at
>
> 51f43d5d82ed ("mm/codetag: swap tags when migrate pages"), which had
> Fixes: e0a955bf7f61 ("mm/codetag: add pgalloc_tag_copy()")
>
> e0a955bf7f61 ("mm/codetag: add pgalloc_tag_copy()") had
> Fixes: dcfe378c81f7 ("lib: introduce support for page allocation tagging")
> Cc: <stable@vger.kernel.org>
>
> And I'm thinking that this fix should have
> Fixes: 51f43d5d82ed ("mm/codetag: swap tags when migrate pages")
> Cc: <stable@vger.kernel.org>
Yes, that is correct. Sorry for missing that.
>
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2024-12-13 4:22 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-12-11 15:08 [linus:master] [mm/codetag] 51f43d5d82: WARNING:at_include/linux/alloc_tag.h:#__alloc_tag_ref_set kernel test robot
2024-12-12 2:12 ` David Wang
2024-12-12 4:01 ` [PATCH] mm/codetag: clear tags before swap David Wang
2024-12-12 7:09 ` Suren Baghdasaryan
2024-12-12 8:17 ` David Wang
2024-12-12 8:29 ` [PATCH v2] " David Wang
2024-12-12 23:18 ` Suren Baghdasaryan
2024-12-13 1:33 ` [PATCH v3] " David Wang
2024-12-13 4:12 ` Andrew Morton
2024-12-13 4:22 ` Suren Baghdasaryan
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox