linux-kernel - Re: [PATCH] fsi: sbefifo: Add sysfs file indicating a timeout error

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <CACPK8XczD=4PXxRQrZ=aGCYtZk47i4-XoFVwep04qszf3Ls6jg@mail.gmail.com>
Date:   Fri, 15 Oct 2021 05:08:47 +0000
From:   Joel Stanley <joel@....id.au>
To:     Eddie James <eajames@...ux.ibm.com>,
        Amitay Isaacs <amitay@...abs.org>
Cc:     linux-fsi@...ts.ozlabs.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Jeremy Kerr <jk@...abs.org>,
        Alistair Popple <alistair@...ple.id.au>,
        OpenBMC Maillist <openbmc@...ts.ozlabs.org>
Subject: Re: [PATCH] fsi: sbefifo: Add sysfs file indicating a timeout error

On Mon, 20 Sept 2021 at 19:00, Eddie James <eajames@...ux.ibm.com> wrote:
>
> The SBEFIFO timeout error requires special handling in userspace
> to do recovery operations. Add a sysfs file to indicate a timeout
> error, and notify pollers when a timeout occurs.

Should this have some documentation too?

What userspace uses this?

Looks good to me otherwise.

Reviewed-by: Joel Stanley <joel@....id.au>

>
> Signed-off-by: Eddie James <eajames@...ux.ibm.com>
> ---
>  drivers/fsi/fsi-sbefifo.c | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
>
> diff --git a/drivers/fsi/fsi-sbefifo.c b/drivers/fsi/fsi-sbefifo.c
> index 84cb965bfed5..b414ab4431ef 100644
> --- a/drivers/fsi/fsi-sbefifo.c
> +++ b/drivers/fsi/fsi-sbefifo.c
> @@ -124,6 +124,7 @@ struct sbefifo {
>         bool                    broken;
>         bool                    dead;
>         bool                    async_ffdc;
> +       bool                    timed_out;
>  };
>
>  struct sbefifo_user {
> @@ -136,6 +137,14 @@ struct sbefifo_user {
>
>  static DEFINE_MUTEX(sbefifo_ffdc_mutex);
>
> +static ssize_t timeout_show(struct device *dev, struct device_attribute *attr,
> +                           char *buf)
> +{
> +       struct sbefifo *sbefifo = container_of(dev, struct sbefifo, dev);
> +
> +       return sysfs_emit(buf, "%d\n", sbefifo->timed_out ? 1 : 0);
> +}
> +static DEVICE_ATTR_RO(timeout);
>
>  static void __sbefifo_dump_ffdc(struct device *dev, const __be32 *ffdc,
>                                 size_t ffdc_sz, bool internal)
> @@ -462,11 +471,14 @@ static int sbefifo_wait(struct sbefifo *sbefifo, bool up,
>                         break;
>         }
>         if (!ready) {
> +               sysfs_notify(&sbefifo->dev.kobj, NULL, dev_attr_timeout.attr.name);
> +               sbefifo->timed_out = true;
>                 dev_err(dev, "%s FIFO Timeout ! status=%08x\n", up ? "UP" : "DOWN", sts);
>                 return -ETIMEDOUT;
>         }
>         dev_vdbg(dev, "End of wait status: %08x\n", sts);
>
> +       sbefifo->timed_out = false;
>         *status = sts;
>
>         return 0;
> @@ -993,6 +1005,8 @@ static int sbefifo_probe(struct device *dev)
>                                  child_name);
>         }
>
> +       device_create_file(&sbefifo->dev, &dev_attr_timeout);
> +
>         return 0;
>   err_free_minor:
>         fsi_free_minor(sbefifo->dev.devt);
> @@ -1018,6 +1032,8 @@ static int sbefifo_remove(struct device *dev)
>
>         dev_dbg(dev, "Removing sbefifo device...\n");
>
> +       device_remove_file(&sbefifo->dev, &dev_attr_timeout);
> +
>         mutex_lock(&sbefifo->lock);
>         sbefifo->dead = true;
>         mutex_unlock(&sbefifo->lock);
> --
> 2.27.0
>