NAK: [PATCH SRU Vivid] Revert "[SCSI] libiscsi: Reduce locking contention in fast path"

[Stefan Bader]

On 14.01.2016 20:59, Tim Gardner wrote:
> On 01/14/2016 12:10 PM, Tim Gardner wrote:
>> Lets hold off on this one until I get a few more details from IBM.
>>
> 
> IBM is recommending that we revert 659743b02c411075b26601725947b21df0bb29c8
> ('[SCSI] libiscsi: Reduce locking contention in fast path')
> 
> git describe --contains 659743b02c411075b26601725947b21df0bb29c8
> v3.15-rc1~136^2~61
> 
> All kernels from Vivid forward are affected.
> 
> I have some reservations. What do y'all think ?

Feeling uneasy about it as well. On one hand the worst case (if properly
catching cases using the new locking which came after the introduction) would be
to fall back to a (not quantified) worse performance.
On the other side the upstream maintainer has made no move to revert it upstream
either. So hard to say whether there possibly is a actual fix in preparation.
The revert would be an awful big churn as well... :/

-Stefan
> 
> rtg


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20160115/7c25df75/attachment.sig>