On 8/21/23 6:47 AM, Saravanan Vajravel wrote:
> Ok. I will work on the resource leak issue on surprise removal and push
> the patch. Until then, we can revert Commit: 699826f4e30a. I will find
> some time this week to repro the resource leak issue.
>
> Thanks & Regards,
> Saravanan Vajravel
> +91-80-46116256
>
>
> -----Original Message-----
> From: Leon Romanovsky <leon@xxxxxxxxxx>
> Sent: Sunday, August 20, 2023 11:04 PM
> To: Sagi Grimberg <sagi@xxxxxxxxxxx>
> Cc: Dennis Dalessandro <dennis.dalessandro@xxxxxxxxxxxxxxxxxxxx>;
> saravanan.vajravel@xxxxxxxxxxxx; OFED mailing list
> <linux-rdma@xxxxxxxxxxxxxxx>; Ehab Ababneh
> <ehab.ababneh@xxxxxxxxxxxxxxxxxxxx>; Selvin Xavier
> <selvin.xavier@xxxxxxxxxxxx>
> Subject: Re: isert patch leaving resources behind
>
> On Sun, Aug 20, 2023 at 05:46:12PM +0300, Sagi Grimberg wrote:
>>
>>>>> Commit: 699826f4e30a ("IB/isert: Fix incorrect release of isert
>>>>> connection") is causing problems on OPA when we try to unload
>>>>> the driver after doing iSCI testing. Reverting this commit causes
> the problem to go away. Any ideas?
>>>>
>>>>
>>>> Saravanan, can you please post kernel logs as you wrote "When a
>>>> bunch of iSER target is cleared, this issue can lead to
>>>> use-after-free memory issue as isert conn is twice released" in the
> reverted commit?
>>>
>>> So what is the resolution here?
>>
>> We need saravanan to address the reported resource leak upon
>> DEVICE_REMOVAL. If this stalls, I suggest we revert the offending
>> commit and add it again without any leaks on surprise hca removals.
>
> Thanks, I'll keep an eye on it, if we don't get patch by EOW, I will
> revert it
Thanks for the help on this folks!
-Denny
