From: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
To: Oliver Neukum <oneukum@suse.com>
Cc: linux-usb@vger.kernel.org
Subject: Re: circular submissions in cdc-wdm and how to break them on disconnect
Date: Sat, 23 Jan 2021 11:57:50 +0900 [thread overview]
Message-ID: <fc789f07-9b29-a86b-5527-ac6f5b3ef2dd@i-love.sakura.ne.jp> (raw)
In-Reply-To: <3dfe07c7ad08d4dfd7eac7bd54e6b821319abe90.camel@suse.com>
On 2021/01/22 0:30, Oliver Neukum wrote:
> Hi,
>
> you have moved kill_urbs() below
>
> cancel_work_sync(&desc->rxwork);
> cancel_work_sync(&desc->service_outs_intr);
>
> to close a race, as
>
> rv = usb_submit_urb(desc->response, GFP_KERNEL);
>
> in service_outstanding_interrupt() would submit the response URB,
> right?
Right. Shouldn't remaining
kill_urbs(desc);
cancel_work_sync(&desc->rxwork);
cancel_work_sync(&desc->service_outs_intr);
sequence in wdm_suspend() and wdm_pre_reset() be updated as well?
> Unfortunately we have in wdm_in_callback() the following code path
>
> if (desc->rerr) {
> /*
> * Since there was an error, userspace may decide to not read
> * any data after poll'ing.
> * We should respond to further attempts from the device to send
> * data, so that we can get unstuck.
> */
> schedule_work(&desc->service_outs_intr);
>
> It looks to me like we have a circular dependency here and this needs some
> change to break. What do you think about the attached patch?
I don't know how poisoning works. But why can't we simply use test_bit() on
WDM_SUSPENDING/WDM_RESETTING/WDM_DISCONNECTING flags, for schedule_work() in
wdm_in_callback() is called with desc->iuspin (which serializes setting of
these flags) held.
By the way, since someone might interpret "broken" as "out of order / not working",
I expect not using "This needs to be broken." in the commit message. There would be
some better idiom.
next prev parent reply other threads:[~2021-01-23 2:58 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-21 15:30 circular submissions in cdc-wdm and how to break them on disconnect Oliver Neukum
2021-01-23 2:57 ` Tetsuo Handa [this message]
2021-02-18 12:55 ` Oliver Neukum
2021-02-18 13:39 ` Tetsuo Handa
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=fc789f07-9b29-a86b-5527-ac6f5b3ef2dd@i-love.sakura.ne.jp \
--to=penguin-kernel@i-love.sakura.ne.jp \
--cc=linux-usb@vger.kernel.org \
--cc=oneukum@suse.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).