netdev - stable backport of 3 xen multiqueue network dev fixes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 21 Jan 2016 11:58:10 +0000
From:	Ian Campbell <ian.campbell@...rix.com>
To:	<netdev@...r.kernel.org>, David Miller <davem@...emloft.net>
CC:	xen-devel <xen-devel@...ts.xen.org>, Wei Liu <wei.liu2@...rix.com>,
	"Boris Ostrovsky" <boris.ostrovsky@...cle.com>,
	David Vrabel <david.vrabel@...rix.com>
Subject: stable backport of 3 xen multiqueue network dev fixes

Hi Dave,

Please could you queue these three:

32a8440 xen-netfront: respect user provided max_queues
4c82ac3 xen-netback: respect user provided max_queues
ca88ea1 xen-netfront: update num_queues to real created

for stable backports to at least 4.1. Multiqueue was added in v3.16 so I
think they could all go to LTS that far back.

Users are tripping over issues in the field with exhausting the number of
Xen grant entries available to guests when creating large numbers of queues
(which happens with guests with >=32 VCPUs).

ca88ea1 is a fix which causes the failure to allocate grants for all of the
desired queues gracefully, rather than crashing or failing to pass any
traffic.

32a8440 and 4c82ac3 fix the handling of the module parameters on both ends
so that users can choose to have fewer queues than p/vcpus if they like,
which would have been an effective workaround for the issue solved by
ca88ea1 if that hadn't been broken.

Thanks,
Ian.