From: Jakub Kicinski <kubakici@wp.pl>
To: Daniel Wagner <daniel.wagner@bmw-carit.de>,
Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: "Luis R. Rodriguez" <mcgrof@kernel.org>, linux-kernel@vger.kernel.org
Subject: Re: 4.10-rc3, firmware loading via user space helper crashes if firmware not present
Date: Fri, 13 Jan 2017 19:40:50 -0800 [thread overview]
Message-ID: <20170113194050.48fa5e0b@laptop> (raw)
In-Reply-To: <20170113133258.1cc6bf39@laptop>
On Fri, 13 Jan 2017 13:32:58 -0800, Jakub Kicinski wrote:
> If one requests a FW which does not exist in the FS and the user space
> helper is used then fw_load_abort() will be called twice which leads to
> NULL-deref.
>
> It will be called once in firmware_loading_store() (handling the -1
> case) and then again in _request_firmware_load() because return value
> from fw_state_wait_timeout() was negative.
>
> I think this is introduced in by f52cc379423d ("firmware: refactor
> loading status").
>
> The simple fix would be to not "unlink" the buf by fw_load_abort() in
> firmware_loading_store() and always rely on firmware_loading_store().
>
> ------->8------------------------------------
>
> diff --git a/drivers/base/firmware_class.c
> b/drivers/base/firmware_class.c index 4497d263209f..89eb9de81145 100644
> --- a/drivers/base/firmware_class.c
> +++ b/drivers/base/firmware_class.c
> @@ -766,7 +770,7 @@ static ssize_t firmware_loading_store(struct device
> *dev, dev_err(dev, "%s: unexpected value (%d)\n", __func__, loading);
> /* fallthrough */
> case -1:
> - fw_load_abort(fw_priv);
> + fw_state_aborted(&fw_buf->fw_st);
> break;
> }
> out:
>
> -------8<------------------------------------
>
> Or should we fix up the ret code handling in __fw_state_wait_common()?
I got this backwards, I blamed the wrong commit, it's the 5d47ec02c37e
("firmware: Correct handling of fw_state_wait() return value") in fact.
It made the assumption that all errors returned from
fw_state_wait_timeout() require calling the abort, while in fact abort
should only be called if user mode helper didn't do anything (either
it timed out or someone hit ^C).
I'm leaning towards this:
------->8------------------------------------
diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c
index 4497d263209f..ce142e6b2c72 100644
--- a/drivers/base/firmware_class.c
+++ b/drivers/base/firmware_class.c
@@ -1020,7 +1020,7 @@ static int _request_firmware_load(struct firmware_priv *fw_priv,
}
retval = fw_state_wait_timeout(&buf->fw_st, timeout);
- if (retval < 0) {
+ if (retval == -ETIMEDOUT || retval == -ERESTARTSYS) {
mutex_lock(&fw_lock);
fw_load_abort(fw_priv);
mutex_unlock(&fw_lock);
-------8<------------------------------------
Unless advised otherwise I will submit this officially on Monday :)
prev parent reply other threads:[~2017-01-14 3:41 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-13 21:32 4.10-rc3, firmware loading via user space helper crashes if firmware not present Jakub Kicinski
2017-01-14 3:40 ` Jakub Kicinski [this message]
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=20170113194050.48fa5e0b@laptop \
--to=kubakici@wp.pl \
--cc=bjorn.andersson@linaro.org \
--cc=daniel.wagner@bmw-carit.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mcgrof@kernel.org \
/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).