Received-SPF: pass (google.com: domain of linux-kernel+bounces-200801-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Tue, 4 Jun 2024 10:20:31 -0400
From: Alan Stern <stern@rowland.harvard.edu>
To: Marcello Sylvester Bauer <marcello.bauer@9elements.com>
Cc: syzbot <syzbot+c793a7eca38803212c61@syzkaller.appspotmail.com>,
  bp@alien8.de, dave.hansen@linux.intel.com, gregkh@linuxfoundation.org,
  hpa@zytor.com, linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org,
  mingo@redhat.com, syzkaller-bugs@googlegroups.com, tglx@linutronix.de,
  x86@kernel.org, Matthias Stoeckl <matthias.stoeckl@secunet.com>,
  Uwe Kleine-Koenig <u.kleine-koenig@pengutronix.de>
Subject: Re: [syzbot] [usb?] INFO: rcu detected stall in schedule_timeout (6)
Message-ID: <3e1b188f-2a68-461c-9b7c-a1d85bf8eb31@rowland.harvard.edu>
References: <000000000000f8112e0618995e6e@google.com>
 <46cbe0a914065917ea1024636e864a1e2c982145.camel@9elements.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <46cbe0a914065917ea1024636e864a1e2c982145.camel@9elements.com>

On Tue, Jun 04, 2024 at 02:05:12PM +0200, Marcello Sylvester Bauer wrote:
> Greetings,
> 
> I'm currently investigating this regression to properly fix it. My
> patch only replaces the corresponding timer API calls without actually
> changing the code. I'm trying to get it to work properly with the
> hrtimer API.
> 
> Any hints on how to accomplish this are welcome.

Start by figuring out what isn't getting executed.  That is, which USB 
(or usb_request) transfer isn't completing.  Once you know that, see 
what timer callbacks do occur (if any) and figure out why they don't 
cause the transfer to complete.

If there are no timer callbacks at all when the hang occurs, make a 
record (printk if nothing else), from the start of the test, of each 
callback and each timer restart, so you can definitively prove to the 
hrtimer maintainers that there's a restart with no later callback.  Then 
it will be up to them to figure out what's going wrong.

Alan Stern