[SRU][F:linux-bluefield][PATCH 0/1] UBUNTU: SAUCE: Revert "UBUNTU: SAUCE: mlxbf_gige: add interrupt counts to "ethtool -S""

David Thompson davthompson at nvidia.com
Fri May 20 14:44:09 UTC 2022


Hello Tim,

This commit should be reverted to align the Ubuntu kernel with upstream code.
The comments from upstream reviewers was that driver-managed interrupt counts are
unnecessary, since the core logic already tracks this and exposes it via /proc/interrupts

Since the commit "add interrupt counts" was only in Ubuntu and not in upstream,
this commit needs to be reverted, and instead replaced with the upstream commit
f4826443f4d6 (see attached email showing notice of commit being applied to net-next)

Does that approach make sense?

- Dave


> -----Original Message-----
> From: Tim Gardner <tim.gardner at canonical.com>
> Sent: Friday, May 20, 2022 8:28 AM
> To: David Thompson <davthompson at nvidia.com>; kernel-
> team at lists.ubuntu.com
> Cc: Meriton Tuli <meriton at nvidia.com>; Khoa Vo <khoav at nvidia.com>
> Subject: Re: [SRU][F:linux-bluefield][PATCH 0/1] UBUNTU: SAUCE: Revert
> "UBUNTU: SAUCE: mlxbf_gige: add interrupt counts to "ethtool -S""
> 
> Why revert this now ? A quick look upstream shows that interrupt statistics do
> not appear to be kept. Won't this break user space ?
> 
> rtg
> 
> On 5/19/22 15:51, David Thompson wrote:
> > BugLink: https://bugs.launchpad.net/bugs/1974241
> >
> > SRU Justification:
> >
> > [Impact]
> >
> > There is a mlxbf_gige driver commit (ed0dd97e64a5) which pertains to
> > "UBUNTU: SAUCE: mlxbf_gige: add interrupt counts to "ethtool -S".
> > This logic is modified in upstream Linux, so this SAUCE commit will
> > need to be reverted.
> >
> > [Fix]
> >
> > The fix is to revert this driver commit.
> >
> > [Test Case]
> >
> > With this commit reverted the driver should function as before (e.g.
> > probe succeeds, oob_net0 link up, etc) EXCEPT that the output of
> > "ethtool -S" will no longer show the interrupt counts.
> >
> > [Regression Potential]
> >
> > The amount of driver logic being reverted is quite small, so no
> > potential for creating a regression.
> >
> > David Thompson (1):
> >    UBUNTU: SAUCE: Revert "UBUNTU: SAUCE: mlxbf_gige: add interrupt
> counts
> >      to "ethtool -S""
> >
> >   .../mellanox/mlxbf_gige/mlxbf_gige_ethtool.c       | 14 +++++---------
> >   1 file changed, 5 insertions(+), 9 deletions(-)
> >
> 
> --
> -----------
> Tim Gardner
> Canonical, Inc
-------------- next part --------------
An embedded message was scrubbed...
From: "patchwork-bot+netdevbpf at kernel.org"
	<patchwork-bot+netdevbpf at kernel.org>
Subject: Re: [PATCH net-next v2] mlxbf_gige: remove driver-managed interrupt counts
Date: Thu, 12 May 2022 23:40:13 +0000
Size: 10258
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20220520/779c42e3/attachment.mht>


More information about the kernel-team mailing list