Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 25 Mar 2022 17:07:12 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     William McVicker <willmcvicker@google.com>
Cc:     Johannes Berg <johannes@sipsolutions.net>,
        linux-wireless@vger.kernel.org,
        Marek Szyprowski <m.szyprowski@samsung.com>,
        Kalle Valo <kvalo@codeaurora.org>,
        "David S. Miller" <davem@davemloft.net>, netdev@vger.kernel.org,
        Amitkumar Karwar <amitkarwar@gmail.com>,
        Ganapathi Bhat <ganapathi.bhat@nxp.com>,
        Xinming Hu <huxinming820@gmail.com>, kernel-team@android.com,
        Paolo Abeni <pabeni@redhat.com>
Subject: Re: [BUG] deadlock in nl80211_vendor_cmd
Message-ID: <20220325170712.69c2c8d3@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
In-Reply-To: <Yj5W7VEij0BGwFK0@google.com>
References: <5d5cf050-7de0-7bad-2407-276970222635@quicinc.com>
        <YjpGlRvcg72zNo8s@google.com>
        <dc556455-51a2-06e8-8ec5-b807c2901b7e@quicinc.com>
        <Yjzpo3TfZxtKPMAG@google.com>
        <19e12e6b5f04ba9e5b192001fbe31a3fc47d380a.camel@sipsolutions.net>
        <20220325094952.10c46350@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
        <f4f8a27dc07c1adaab470fde302ed841113e6b7f.camel@sipsolutions.net>
        <Yj4FFIXi//ivQC3X@google.com>
        <Yj4ntUejxaPhrM5b@google.com>
        <976e8cf697c7e5bc3a752e758a484b69a058710a.camel@sipsolutions.net>
        <Yj5W7VEij0BGwFK0@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri, 25 Mar 2022 23:57:33 +0000 William McVicker wrote:
> Instead of open coding it, we could just pass the internal_flags via struct
> genl_info to nl80211_vendor_cmds() and then handle the rtnl_unlock() there if
> the vendor command doesn't request it. I included the patch below in case
> there's any chance you would consider this for upstream. This would at least
> add backwards compatibility to the vendor ops API so that existing drivers that
> depend on the RTNL being held don't need to be fully refactored.

Sorry to step in, Johannes may be AFK already. There's no asterisk next
to the "we don't cater to out of tree code" rule, AFAIK.  We change
locking often, making a precedent like this would be ill-advised.