Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <1556339208-7722-1-git-send-email-hofrat@osadl.org>
In-Reply-To: <1556339208-7722-1-git-send-email-hofrat@osadl.org>
From:   Sven Van Asbroeck <thesven73@gmail.com>
Date:   Sat, 27 Apr 2019 02:17:42 -0400
Message-ID: <CAGngYiXBrNu7Jrk535=L89-mOSNx-6EfqL=1EwwG49Sf44R0Vg@mail.gmail.com>
Subject: Re: [PATCH RFC] staging: fieldbus: anybus-s: use proper type for wait_for_completion_timeout
To:     Nicholas Mc Guire <hofrat@osadl.org>
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        devel@driverdev.osuosl.org,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hello Nicholas, thank you for your contribution, I really appreciate it !
See inline comments below.

On Sat, Apr 27, 2019 at 12:32 AM Nicholas Mc Guire <hofrat@osadl.org> wrote:
>
> wait_for_completion_timeout() returns unsigned long (0 on timeout or
> remaining jiffies) not int.

Nice catch !

> thus there is no negative case to check for
> here.

The current code can only break if wait_for_completion_timeout()
returns an unsigned long large enough to make the "int ret" turn
negative. This could result in probe() returning a nonsensical error
value, even though the wait did not time out.

Fortunately that currently cannot happen, because TIMEOUT
(2*HZ) won't overflow an int.

That being said, this return value type mismatch should definitely
be fixed up.

>
> Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
> ---
>
> Problem located with experimental API conformance checking cocci script
>
> Q: It is not really clear if the proposed fix is the right thing here or if
>    this should not be using wait_for_completion_interruptible - which would
>    return -ERESTARTSYS on interruption. Someone that knows the details of
>    this driver would need to check what condition should initiate the
>    goto err_reset; which was actually unreachable in the current code.

It's used in probe(), so no need for an interruptible wait, AFAIK.
It can stay as-is.

>
> Patch was compile-tested with. x86_64_defconfig + FIELDBUS_DEV=m,
> HMS_ANYBUSS_BUS=m
> (some unrelated sparse warnings (cast to restricted __be16))

That sounds interesting too. Could you provide more details?

<snip>

> -       ret = wait_for_completion_timeout(&cd->card_boot, TIMEOUT);
> -       if (ret == 0)
> +       time_left = wait_for_completion_timeout(&cd->card_boot, TIMEOUT);
> +       if (time_left == 0)
>                 ret = -ETIMEDOUT;
> -       if (ret < 0)
> -               goto err_reset;

NAK. This does not jump to err_reset on timeout.

May I suggest the following instead ? (manual formatting)

- ret = wait_for_completion_timeout(&cd->card_boot, TIMEOUT);
- if (ret == 0)
-         ret = -ETIMEDOUT;
- if (ret < 0)
-         goto err_reset;
+ if (!wait_for_completion_timeout(&cd->card_boot, TIMEOUT)) {
+         ret = -ETIMEDOUT;
+         goto err_reset;
+ }