Message-ID: <CYYPR11MB84296F0E205598814A792099BD2B2@CYYPR11MB8429.namprd11.prod.outlook.com>
Date: Tue, 12 Mar 2024 15:16:43 +0000
From: "Pucha, HimasekharX Reddy" <himasekharx.reddy.pucha@...el.com>
To: "Brandeburg, Jesse" <jesse.brandeburg@...el.com>, "pmenzel@...gen.mpg.de"
	<pmenzel@...gen.mpg.de>, "intel-wired-lan@...ts.osuosl.org"
	<intel-wired-lan@...ts.osuosl.org>
CC: "netdev@...r.kernel.org" <netdev@...r.kernel.org>, "Kitszel, Przemyslaw"
	<przemyslaw.kitszel@...el.com>, Eric Dumazet <edumazet@...gle.com>, "Nguyen,
 Anthony L" <anthony.l.nguyen@...el.com>, "horms@...nel.org"
	<horms@...nel.org>, Michal Swiatkowski <michal.swiatkowski@...ux.intel.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, "David S.
 Miller" <davem@...emloft.net>, "Elliott, Rob" <elliott@....com>
Subject: RE: [Intel-wired-lan] [PATCH iwl-net v2] ice: fix memory corruption
 bug with suspend and rebuild

> -----Original Message-----
> From: Intel-wired-lan <intel-wired-lan-bounces@...osl.org> On Behalf Of Jesse Brandeburg
> Sent: Wednesday, March 6, 2024 4:32 AM
> To: pmenzel@...gen.mpg.de; intel-wired-lan@...ts.osuosl.org
> Cc: netdev@...r.kernel.org; Kitszel, Przemyslaw <przemyslaw.kitszel@...el.com>; Eric Dumazet <edumazet@...gle.com>; Nguyen, Anthony L <anthony.l.nguyen@...el.com>; horms@...nel.org; Michal Swiatkowski <michal.swiatkowski@...ux.intel.com>; Jakub Kicinski <kuba@...nel.org>; Paolo Abeni <pabeni@...hat.com>; David S. Miller <davem@...emloft.net>; Elliott, Rob <elliott@....com>
> Subject: [Intel-wired-lan] [PATCH iwl-net v2] ice: fix memory corruption bug with suspend and rebuild
>
> The ice driver would previously panic after suspend. This is caused from the driver *only* calling the ice_vsi_free_q_vectors() function by itself, when it is suspending. Since commit b3e7b3a6ee92 ("ice: prevent NULL pointer deref during reload") the driver has zeroed out num_q_vectors, and only restored it in ice_vsi_cfg_def().
>
> This further causes the ice_rebuild() function to allocate a zero length buffer, after which num_q_vectors is updated, and then the new value of num_q_vectors is used to index into the zero length buffer, which corrupts memory.
>
> The fix entails making sure all the code referencing num_q_vectors only does so after it has been reset via ice_vsi_cfg_def().
>
> I didn't perform a full bisect, but I was able to test against 6.1.77 kernel and that ice driver works fine for suspend/resume with no panic, so sometime since then, this problem was introduced.
>
> Also clean up an un-needed init of a local variable in the function being modified.
>
> PANIC from 6.8.0-rc1:
>
> [1026674.915596] PM: suspend exit
> [1026675.664697] ice 0000:17:00.1: PTP reset successful [1026675.664707] ice 0000:17:00.1: 2755 msecs passed between update to cached PHC time [1026675.667660] ice 0000:b1:00.0: PTP reset successful [1026675.675944] ice 0000:b1:00.0: 2832 msecs passed between update to cached PHC time [1026677.137733] ixgbe 0000:31:00.0 ens787: NIC Link is Up 1 Gbps, Flow Control: None [1026677.190201] BUG: kernel NULL pointer dereference, address: 0000000000000010 [1026677.192753] ice 0000:17:00.0: PTP reset successful [1026677.192764] ice 0000:17:00.0: 4548 msecs passed between update to cached PHC time [1026677.197928] #PF: supervisor read access in kernel mode [1026677.197933] #PF: error_code(0x0000) - not-present page [1026677.197937] PGD 1557a7067 P4D 0 [1026677.212133] ice 0000:b1:00.1: PTP reset successful [1026677.212143] ice 0000:b1:00.1: 4344 msecs passed between update to cached PHC time [1026677.212575] > [1026677.243142] Oops: 0000 [#1] PREEMPT SMP NOPTI
> [1026677.247918] CPU: 23 PID: 42790 Comm: kworker/23:0 Kdump: loaded Tainted: G        W          6.8.0-rc1+ #1
> [1026677.257989] Hardware name: Intel Corporation M50CYP2SBSTD/M50CYP2SBSTD, BIOS SE5C620.86B.01.01.0005.2202160810 02/16/2022 [1026677.269367] Workqueue: ice ice_service_task [ice] [1026677.274592] RIP: 0010:ice_vsi_rebuild_set_coalesce+0x130/0x1e0 [ice] [1026677.281421] Code: 0f 84 3a ff ff ff 41 0f b7 74 ec 02 66 89 b0 22 02 00 00 81 e6 ff 1f 00 00 e8 ec fd ff ff e9 35 ff ff ff 48 8b 43 30 49 63 ed <41> 0f b7 34 24 41 83 c5 01 48 8b 3c e8 66 89 b7 aa 02 00 00 81 e6 [1026677.300877] RSP: 0018:ff3be62a6399bcc0 EFLAGS: 00010202 [1026677.306556] RAX: ff28691e28980828 RBX: ff28691e41099828 RCX: 0000000000188000 [1026677.314148] RDX: 0000000000000000 RSI: 0000000000000010 RDI: ff28691e41099828 [1026677.321730] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 [1026677.329311] R10: 0000000000000007 R11: ffffffffffffffc0 R12: 0000000000000010 [1026677.336896] R13: 0000000000000000 R14: 0000000000000000 R15: ff28691e0eaa81a0 [1026677.344472] FS:  0000000000000000(0000) GS:ff28693cbffc0000(0000) knlGS:0000000000000000 [1026677.353000] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [1026677.359195] CR2: 0000000000000010 CR3: 0000000128df4001 CR4: 0000000000771ef0 [1026677.366779] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [1026677.374369] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [1026677.381952] PKRU: 55555554 [1026677.385116] Call Trace:
> [1026677.388023]  <TASK>
> [1026677.390589]  ? __die+0x20/0x70
> [1026677.394105]  ? page_fault_oops+0x82/0x160 [1026677.398576]  ? do_user_addr_fault+0x65/0x6a0 [1026677.403307]  ? exc_page_fault+0x6a/0x150 [1026677.407694]  ? asm_exc_page_fault+0x22/0x30 [1026677.412349]  ? ice_vsi_rebuild_set_coalesce+0x130/0x1e0 [ice] [1026677.418614]  ice_vsi_rebuild+0x34b/0x3c0 [ice] [1026677.423583]  ice_vsi_rebuild_by_type+0x76/0x180 [ice] [1026677.429147]  ice_rebuild+0x18b/0x520 [ice] [1026677.433746]  ? delay_tsc+0x8f/0xc0 [1026677.437630]  ice_do_reset+0xa3/0x190 [ice] [1026677.442231]  ice_service_task+0x26/0x440 [ice] [1026677.447180]  process_one_work+0x174/0x340 [1026677.451669]  worker_thread+0x27e/0x390 [1026677.455890]  ? __pfx_worker_thread+0x10/0x10 [1026677.460627]  kthread+0xee/0x120 [1026677.464235]  ? __pfx_kthread+0x10/0x10 [1026677.468445]  ret_from_fork+0x2d/0x50 [1026677.472476]  ? __pfx_kthread+0x10/0x10 [1026677.476671]  ret_from_fork_asm+0x1b/0x30 [1026677.481050]  </TASK>
> 
> Fixes: b3e7b3a6ee92 ("ice: prevent NULL pointer deref during reload")
> Reported-by: Robert Elliott <elliott@....com>
> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@...el.com>
> ---
> v2: fix uninitialized coalesce pointer on the exit path by moving the kfree to the later goto (simon), reword commit message (paul)
> ---
>  drivers/net/ethernet/intel/ice/ice_lib.c | 18 +++++++++---------
>  1 file changed, 9 insertions(+), 9 deletions(-)
>

Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@...el.com> (A Contingent worker at Intel)

Powered by blists - more mailing lists