linux-kernel - Re: [PATCH v5.1-rc] iwlwifi: make locking in iwl_mvm_tx

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

Message-ID: <nycvar.YFH.7.76.1904151328270.9803@cbobk.fhfr.pm>
Date:   Mon, 15 Apr 2019 13:33:47 +0200 (CEST)
From:   Jiri Kosina <jikos@...nel.org>
To:     Johannes Berg <johannes@...solutions.net>
cc:     Emmanuel Grumbach <emmanuel.grumbach@...el.com>,
        Luca Coelho <luciano.coelho@...el.com>,
        Intel Linux Wireless <linuxwifi@...el.com>,
        Kalle Valo <kvalo@...eaurora.org>,
        "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5.1-rc] iwlwifi: make locking in iwl_mvm_tx_mpdu()
 BH-safe

On Mon, 15 Apr 2019, Johannes Berg wrote:

> (FWIW, would be good to cc linux-wireless, but no need for linux-kernel
> I guess :-) )

Hm, no idea where I lost that ML from the get_maintainer output, sorry for 
that.

> > iwl_mvm_sta->lock can be taken from BH, and therefore BH must be 
> > disabled if taking it from other contexts.
> > 
> > This fixes the lockdep warning below.
> 
> I don't think this is the right fix.
> 
> The only problem is this path:

Agreed, that seems to be the only culprit.

> >      SOFTIRQ-ON-W at:
> >                        _raw_spin_lock+0x35/0x50
> >                        iwl_mvm_tx_mpdu+0xae/0x5f0 [iwlmvm]
> >                        iwl_mvm_tx_skb+0x200/0x470 [iwlmvm]
> >                        iwl_mvm_mac_itxq_xmit+0xd5/0x1f0 [iwlmvm]
> >                        iwl_mvm_add_new_dqa_stream_wk+0x34c/0x910 [iwlmvm]
> >                        process_one_work+0x1f0/0x5b0
> 
> which I believe we just made use local_bh_disable() since there are
> other reasons as well for that, IIRC.
> 
> Yes, here's the fix:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers.git/commit/?id=f5ae2f932e2f8f4f79796f44832ae8fca26f188a
> 
> It's on the way upstream.

If there are other reasons why disable BH for the whole function (are 
there?), then this bigger hammer works as well of course.

	Tested-by: Jiri Kosina <jkosina@...e.cz>

> Thanks for the patch/report though!

Thanks,

-- 
Jiri Kosina
SUSE Labs