[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231128172053.GA43811@kernel.org>
Date: Tue, 28 Nov 2023 17:20:53 +0000
From: Simon Horman <horms@...nel.org>
To: Subbaraya Sundeep <sbhatta@...vell.com>
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org, kuba@...nel.org,
davem@...emloft.net, pabeni@...hat.com, edumazet@...gle.com,
sgoutham@...vell.com, gakula@...vell.com, hkelam@...vell.com
Subject: Re: [PATCH net] octeontx2-pf: Add missing mutex lock in
otx2_get_pauseparam
On Sat, Nov 25, 2023 at 10:05:41PM +0530, Subbaraya Sundeep wrote:
> All the mailbox messages sent to AF needs to be guarded
> by mutex lock. Add the missing lock in otx2_get_pauseparam
> function.
>
> Fixes: 75f36270990c ("octeontx2-pf: Support to enable/disable pause frames via ethtool")
> Signed-off-by: Subbaraya Sundeep <sbhatta@...vell.com>
Hi,
I am wondering if the call to otx2_nix_config_bp()
in otx2_dcbnl_ieee_setpfc() also needs to be protected by mbox.lock.
And although not strictly related to this patch, while looking over this, I
noticed that in otx2_init_hw_resources() it appears that &mbox->lock may be
unlocked twice in some error paths.
e.g.
/* Init Auras and pools used by NIX RQ, for free buffer ptrs */
err = otx2_rq_aura_pool_init(pf);
if (err) {
mutex_unlock(&mbox->lock);
goto err_free_nix_lf;
}
...
err_free_nix_lf:
mutex_lock(&mbox->lock);
...
...
Powered by blists - more mailing lists