[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Y71H5ub5K+9GlGTm@unreal>
Date: Tue, 10 Jan 2023 13:11:34 +0200
From: Leon Romanovsky <leon@...nel.org>
To: Paolo Abeni <pabeni@...hat.com>
Cc: Hariprasad Kelam <hkelam@...vell.com>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, kuba@...nel.org, davem@...emloft.net,
edumazet@...gle.com, sgoutham@...vell.com, gakula@...vell.com,
sbhatta@...vell.com
Subject: Re: [net PATCH] octeontx2-pf: Fix resource leakage in VF driver
unbind
On Tue, Jan 10, 2023 at 11:43:28AM +0100, Paolo Abeni wrote:
> Hello,
>
> On Tue, 2023-01-10 at 12:29 +0200, Leon Romanovsky wrote:
> > On Tue, Jan 10, 2023 at 10:20:15AM +0000, patchwork-bot+netdevbpf@...nel.org wrote:
> > > Hello:
> > >
> > > This patch was applied to netdev/net.git (master)
> > > by Paolo Abeni <pabeni@...hat.com>:
> > >
> > > On Mon, 9 Jan 2023 11:43:25 +0530 you wrote:
> > > > resources allocated like mcam entries to support the Ntuple feature
> > > > and hash tables for the tc feature are not getting freed in driver
> > > > unbind. This patch fixes the issue.
> > > >
> > > > Fixes: 2da489432747 ("octeontx2-pf: devlink params support to set mcam entry count")
> > > > Signed-off-by: Hariprasad Kelam <hkelam@...vell.com>
> > > > Signed-off-by: Sunil Kovvuri Goutham <sgoutham@...vell.com>
> > > >
> > > > [...]
> > >
> > > Here is the summary with links:
> > > - [net] octeontx2-pf: Fix resource leakage in VF driver unbind
> > > https://git.kernel.org/netdev/net/c/53da7aec3298
> >
> > I don't think that this patch should be applied.
> >
> > It looks like wrong Fixes to me and I don't see clearly how structures
> > were allocated on VF which were cleared in this patch.
>
> My understanding is that the resource allocation happens via:
>
> otx2_dl_mcam_count_set()
>
> which is registered as the devlink parameter write operation on the vf
> by the fixes commit - the patch looks legit to me.
>
> Did I miss something?
No, you are right. I'm not sure if I would be able to see that OTX2_FLAG_MCAM_ENTRIES_ALLOC
flag without your hint.
Thanks
Powered by blists - more mailing lists