2022-11-09 11:12:47

by Borislav Petkov

Subject: AMD uncore kmemleak

Hey,

Peter said you liked this stuff so here it is. :-)

On one of my test boxes (Carrizo laptop - that's an Excavator core) I
see the below. It happens with 6.1-rc1.

I booted latest -rc4 + tip/master and it is gone so something has
changed in the meantime or this particular version doesn't fire...

Anyway, we thought you should know in case you want to dig deeper.

Thx.

cat /sys/kernel/debug/kmemleak
unreferenced object 0xffff888101a3d3a0 (size 32):
comm "swapper/0", pid 1, jiffies 4294892455 (age 294.968s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000002f5fe849>] __kmalloc_node+0x5f/0x180
[<00000000066a1fb1>] amd_uncore_cpu_up_prepare+0xc7/0x240
[<000000005fa7fab5>] cpuhp_invoke_callback+0xf8/0x460
[<00000000d4b0952e>] cpuhp_issue_call+0x118/0x190
[<000000008cb1592a>] __cpuhp_setup_state_cpuslocked+0xee/0x2a0
[<0000000007534fa6>] __cpuhp_setup_state+0x76/0x100
[<00000000d8277f50>] amd_uncore_init+0x275/0x336
[<000000004af336cd>] do_one_initcall+0x44/0x200
[<0000000040cb5caf>] kernel_init_freeable+0x198/0x1e4
[<000000008b0e9fae>] kernel_init+0x16/0x120
[<00000000459bcaeb>] ret_from_fork+0x22/0x30
unreferenced object 0xffff888101a3d240 (size 32):
comm "swapper/0", pid 1, jiffies 4294892455 (age 294.968s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000002f5fe849>] __kmalloc_node+0x5f/0x180
[<00000000066a1fb1>] amd_uncore_cpu_up_prepare+0xc7/0x240
[<000000005fa7fab5>] cpuhp_invoke_callback+0xf8/0x460
[<00000000d4b0952e>] cpuhp_issue_call+0x118/0x190
[<000000008cb1592a>] __cpuhp_setup_state_cpuslocked+0xee/0x2a0
[<0000000007534fa6>] __cpuhp_setup_state+0x76/0x100
[<00000000d8277f50>] amd_uncore_init+0x275/0x336
[<000000004af336cd>] do_one_initcall+0x44/0x200
[<0000000040cb5caf>] kernel_init_freeable+0x198/0x1e4
[<000000008b0e9fae>] kernel_init+0x16/0x120
[<00000000459bcaeb>] ret_from_fork+0x22/0x30
unreferenced object 0xffff888101a3d4a0 (size 32):
comm "swapper/0", pid 1, jiffies 4294892455 (age 294.968s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000002f5fe849>] __kmalloc_node+0x5f/0x180
[<00000000066a1fb1>] amd_uncore_cpu_up_prepare+0xc7/0x240
[<000000005fa7fab5>] cpuhp_invoke_callback+0xf8/0x460
[<00000000d4b0952e>] cpuhp_issue_call+0x118/0x190
[<000000008cb1592a>] __cpuhp_setup_state_cpuslocked+0xee/0x2a0
[<0000000007534fa6>] __cpuhp_setup_state+0x76/0x100
[<00000000d8277f50>] amd_uncore_init+0x275/0x336
[<000000004af336cd>] do_one_initcall+0x44/0x200
[<0000000040cb5caf>] kernel_init_freeable+0x198/0x1e4
[<000000008b0e9fae>] kernel_init+0x16/0x120
[<00000000459bcaeb>] ret_from_fork+0x22/0x30

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2022-11-09 11:30:28

by Borislav Petkov

Subject: Re: AMD uncore kmemleak

On Wed, Nov 09, 2022 at 04:33:59PM +0530, Ravi Bangoria wrote:
> I guess this should fix it:
> https://lore.kernel.org/all/4fa9e5ac6d6e41fa889101e7af7e6ba372cfea52.1662613255.git.sandipan.das@amd.com

Cool, lemme run it.

Thx.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2022-11-09 11:52:47

by Ravi Bangoria

Subject: Re: AMD uncore kmemleak

Hi Boris,

On 09-Nov-22 4:22 PM, Borislav Petkov wrote:
> Hey,
>
> Peter said you liked this stuff so here it is. :-)
>
> On one of my test boxes (Carrizo laptop - that's an Excavator core) I
> see the below. It happens with 6.1-rc1.
>
> I booted latest -rc4 + tip/master and it is gone so something has
> changed in the meantime or this particular version doesn't fire...
>
> Anyway, we thought you should know in case you want to dig deeper.
>
> Thx.
>
> cat /sys/kernel/debug/kmemleak
> unreferenced object 0xffff888101a3d3a0 (size 32):
> comm "swapper/0", pid 1, jiffies 4294892455 (age 294.968s)
> hex dump (first 32 bytes):
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> backtrace:
> [<000000002f5fe849>] __kmalloc_node+0x5f/0x180
> [<00000000066a1fb1>] amd_uncore_cpu_up_prepare+0xc7/0x240
> [<000000005fa7fab5>] cpuhp_invoke_callback+0xf8/0x460
> [<00000000d4b0952e>] cpuhp_issue_call+0x118/0x190
> [<000000008cb1592a>] __cpuhp_setup_state_cpuslocked+0xee/0x2a0
> [<0000000007534fa6>] __cpuhp_setup_state+0x76/0x100
> [<00000000d8277f50>] amd_uncore_init+0x275/0x336
> [<000000004af336cd>] do_one_initcall+0x44/0x200
> [<0000000040cb5caf>] kernel_init_freeable+0x198/0x1e4
> [<000000008b0e9fae>] kernel_init+0x16/0x120
> [<00000000459bcaeb>] ret_from_fork+0x22/0x30
> unreferenced object 0xffff888101a3d240 (size 32):

I guess this should fix it:
https://lore.kernel.org/all/4fa9e5ac6d6e41fa889101e7af7e6ba372cfea52.1662613255.git.sandipan.das@amd.com

Thanks,
Ravi