linux-kernel - Re: [PATCH 12/12] wifi: mwifiex: drop asynchronous init waiting code

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] [day] [month] [year] [list]

Message-ID: <ZuADwpidC7wG7J6q@gaggiata.pivistrello.it>
Date: Tue, 10 Sep 2024 10:30:58 +0200
From: Francesco Dolcini <francesco@...cini.it>
To: Sascha Hauer <s.hauer@...gutronix.de>
Cc: Francesco Dolcini <francesco@...cini.it>,
	Brian Norris <briannorris@...omium.org>,
	Kalle Valo <kvalo@...nel.org>, linux-wireless@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 12/12] wifi: mwifiex: drop asynchronous init waiting code

On Mon, Sep 09, 2024 at 10:14:56PM +0200, Sascha Hauer wrote:
> On Mon, Sep 09, 2024 at 07:14:11PM +0200, Francesco Dolcini wrote:
> > On Mon, Aug 26, 2024 at 01:01:33PM +0200, Sascha Hauer wrote:
> > > Historically all commands sent to the mwifiex driver have been
> > > asynchronous. For this reason there is code that waits for the
> > > last initialization command to complete before going on. Nowadays the
> > > commands can be sent synchronously, meaning that they are completed
> > > when the command call returns. This makes all the waiting code
> > > unnecessary. It is removed in this patch.
> > 
> > I am not sure to understand this. Is the code to have asynchronous command gone
> > or it is just not used anymore? In the code here you remove waiting for the
> > firmware init to be complete, but from the patch is not clear why this is not
> > needed anymore.  Maybe a specific commit you can reference in which such
> > support was removed?
> 
> Commands can still be sent asynchronously by passing sync=false to
> mwifiex_send_cmd(), but this is no longer done in the initialization

Understood. So this is just not unused code since quite some time.
To me the change looks ok, but I would appreciate if someone else can have
another look.


> | commit 7bff9c974e1a70819c30c37d8ec0d84d456f8237
> | Author: Stone Piao <piaoyun@...vell.com>
> | Date:   Tue Sep 25 20:23:39 2012 -0700
> | 
> |     mwifiex: send firmware initialization commands synchronously
> | 
> |     The driver will send some commands to firmware during the
> |     initialization. Currently these commands are sent asynchronously,
> |     which means that we firstly insert all of them to a pre-allocated
> |     command queue, and then start to process them one by one. The
> |     command queue will soon be exhausted if we keep adding new
> |     initialization commands.
> | 
> |     This issue can be resolved by sending initialization commands
> |     synchronously because each command is consumed and the buffer is
> |     recycled before queuing next command.
> | 
> |     Signed-off-by: Stone Piao <piaoyun@...vell.com>
> |     Signed-off-by: Bing Zhao <bzhao@...vell.com>
> |     Signed-off-by: John W. Linville <linville@...driver.com>
> 
> I'll mention this commit in the commit message next round.

Perfect, thanks.