[PATCH] Revert "scsi: qla2xxx: Fix memcpy() field-spanning write issue"

[John Meneghini]

On 9/25/25 9:38 AM, Gustavo A. R. Silva wrote: 
> On 9/25/25 15:07, John Meneghini wrote:
>> This reverts commit 6f4b10226b6b1e7d1ff3cdb006cf0f6da6eed71e.
>>
>> We've been testing this patch and it turns out there is a significant
>> bug here. This leaks memory and causes a driver hang.
>>
>> Link:
>> https://lore.kernel.org/linux-scsi/yq1zfajqpec.fsf@ca-mkp.ca.oracle.com/
> 
> Thanks for the report. I wonder if you have any logs or something I could
> look at to figure out what's going on.


We have a fix already.  Chris and Bryan figured it out.

> Bryan,
> 
> Could you please share how this patch[1] was tested?

Bryan, please reply with bug fix patch you emailed me yesterday as an RFC patch.

Gustavo, this patch is being tested as a part of our FPIN LI changes. To run this code you need a Brocade switch and a whole lot of hardware.

You can see a example test plan here: https://bugzilla.kernel.org/attachment.cgi?id=308368&action=view

I am about to submit a version 10 patch series for these changes and I will include a new/fixed version of your patch in that series.

/John

> Thanks
> -Gustavo
> 
> [1] https://lore.kernel.org/linux-scsi/20250813200744.17975-10-bgurney@redhat.com/
>