Re: [PATCH RFT v2] x86: move cacheinfo sysfs to generic cacheinfo infrastructure

[Borislav Petkov <bp@suse.de>]

On Mon, Feb 23, 2015 at 06:14:25PM +0000, Sudeep Holla wrote:
>  - Rebased on v4.0-rc1
>  - Fixed lockdep warning reported by Borislav

You probably have fixed the lockdep splat but not the NULL pointer
dereference which was there in the first mail I sent you. I'd suggest
you find an AMD box with an L3 and do some testing yourself before
sending out another version.

Also, when testing, do a snapshot of the sysfs cache hierarchy by doing:

grep . -EriIn /sys/devices/system/cpu/cpu?/cache/*

before your changes and after and check for differences.

We can't allow ourselves any differences because this is an API.

HTH.

[    2.981740] BUG: unable to handle kernel NULL pointer dereference at 00000000000000f8
[    2.989918] IP: [<ffffffff810141c0>] _populate_cache_leaves+0x440/0x460
[    2.996724] PGD 0 
[    2.998928] Oops: 0002 [#1] PREEMPT SMP 
[    3.003147] Modules linked in:
[    3.006395] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.0.0-rc1+ #1
[    3.012801] Hardware name: MICRO-STAR INTERNATIONAL CO.,LTD MS-7599/870-C45 (MS-7599), BIOS V1.15 03/04/2011
[    3.022871] task: ffff880132138000 ti: ffff880132180000 task.ti: ffff880132180000
[    3.030598] RIP: 0010:[<ffffffff810141c0>]  [<ffffffff810141c0>] _populate_cache_leaves+0x440/0x460
[    3.039942] RSP: 0000:ffff880132183c88  EFLAGS: 00010086
[    3.045395] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00000000ffffffff
[    3.052671] RDX: 0000000000000000 RSI: 0000000000000006 RDI: 0000000000000000
[    3.059945] RBP: ffff880132183d28 R08: ffff880136c0a0c0 R09: 0000000000000001
[    3.067221] R10: 0000000000000000 R11: 000000000000001c R12: 0000000000000000
[    3.074495] R13: 000000000000a0c0 R14: ffffffff81ac4f98 R15: 0000000000000001
[    3.081770] FS:  0000000000000000(0000) GS:ffff880136c00000(0000) knlGS:0000000000000000
[    3.090108] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[    3.095996] CR2: 00000000000000f8 CR3: 0000000001a0b000 CR4: 00000000000006f0
[    3.103270] Stack:
[    3.105426]  ffff880132183de4 000000000000a000 000000000000a000 ffff880136c0fbc0
[    3.113272]  0000000000000090 ffff8801322c16b0 0000000000000006 ffff8801322c1720
[    3.121120]  000000000000a0c0 0000000300000000 0bc0003f14000163 00000000000007ff
[    3.128959] Call Trace:
[    3.131548]  [<ffffffff81013d80>] ? cpuid4_cache_lookup_regs+0x3e0/0x3e0
[    3.138390]  [<ffffffff810fd666>] generic_exec_single+0x136/0x1a0
[    3.144625]  [<ffffffff81013d80>] ? cpuid4_cache_lookup_regs+0x3e0/0x3e0
[    3.151466]  [<ffffffff810fd733>] smp_call_function_single+0x63/0xd0
[    3.157960]  [<ffffffff811ab560>] ? __kmalloc+0x160/0x200
[    3.163500]  [<ffffffff81014b72>] populate_cache_leaves+0x22/0x30
[    3.169734]  [<ffffffff813c8feb>] detect_cache_attributes+0x7b/0x1a0
[    3.176230]  [<ffffffff810002e0>] ? do_one_initcall+0x90/0x1f0
[    3.182204]  [<ffffffff81b17076>] ? container_dev_init+0x2f/0x2f
[    3.188352]  [<ffffffff81b170b8>] cacheinfo_sysfs_init+0x42/0x93
[    3.194498]  [<ffffffff81b17076>] ? container_dev_init+0x2f/0x2f
[    3.200649]  [<ffffffff810002f0>] do_one_initcall+0xa0/0x1f0
[    3.206448]  [<ffffffff8109ab00>] ? parse_args+0x150/0x420
[    3.212076]  [<ffffffff81ae5fae>] kernel_init_freeable+0x115/0x19d
[    3.218397]  [<ffffffff8157683f>] ? ret_from_fork+0xf/0xb0
[    3.224025]  [<ffffffff81568b40>] ? rest_init+0xd0/0xd0
[    3.229393]  [<ffffffff81568b4e>] kernel_init+0xe/0xf0
[    3.234674]  [<ffffffff815768ac>] ret_from_fork+0x7c/0xb0
[    3.240214]  [<ffffffff81568b40>] ? rest_init+0xd0/0xd0
[    3.245579] Code: 3c dd 40 4c ac 81 ff c2 48 63 d2 be 06 00 00 00 e8 a6 7c 2b 00 3b 05 24 1f ab 00 89 c2 7d a4 89 c0 
[    3.268135] RIP  [<ffffffff810141c0>] _populate_cache_leaves+0x440/0x460
[    3.275024]  RSP <ffff880132183c88>
[    3.278652] CR2: 00000000000000f8
[    3.282112] ---[ end trace 17992ce30d6a630e ]---
[    3.286872] Kernel panic - not syncing: Fatal exception
[    3.292302] Kernel Offset: disabled
[    3.295934] ---[ end Kernel panic - not syncing: Fatal exception

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--