All of lore.kernel.org
 help / color / mirror / Atom feed
From: Corey Minyard <minyard@acm.org>
To: Ryan O'Leary <ryanoleary@google.com>
Cc: openipmi-developer@lists.sourceforge.net, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/1] ipmi/watchdog: Add WDIOC_GETTIMELEFT ioctl
Date: Wed, 5 May 2021 16:36:19 -0500	[thread overview]
Message-ID: <20210505213619.GS18645@minyard.net> (raw)
In-Reply-To: <CAKVu-U=Om6CQF7DVkzp487=R_a77Qbi_grKQ7J-HPVZ_QR7DfA@mail.gmail.com>

On Tue, May 04, 2021 at 01:59:49PM -0700, Ryan O'Leary wrote:
> Hi, please let me know if there's anything missing for review.

Only my negligence.

Unfortunately I can't add anything to my next tree at the moment, I have
to wait until the merge window closes.

Anyway, on the patch, if you look at the definition of msg_wait, it
says:

/*
 * We use a mutex to make sure that only one thing can send a set a
 * message at one time.  The mutex is claimed when a message is sent
 * and freed when both the send and receive messages are free.
 */

So if you use msg_wait, you need to grab that mutex, or bad things can
happen.

However, that means that a non-critical operation, grabbing the time
left, can result in blocking a more critical operations, pinging the
watchdog.  So it would be better if you could use another set of
messages for doing this.

-corey

> 
> On Tue, Mar 16, 2021 at 5:11 PM Ryan O'Leary <ryanoleary@google.com> wrote:
> 
> > This is the same ioctl the rest of the watchdogs support. GETTIMELEFT
> > returns the number of seconds in the countdown -- useful for testing
> > whether the watchdog is functioning.
> >
> > Signed-off-by: Ryan O'Leary <ryanoleary@google.com>
> > ---
> >  drivers/char/ipmi/ipmi_watchdog.c | 75 +++++++++++++++++++++++++++++++
> >  1 file changed, 75 insertions(+)
> >
> > diff --git a/drivers/char/ipmi/ipmi_watchdog.c
> > b/drivers/char/ipmi/ipmi_watchdog.c
> > index 32c334e34d55..f253c8667395 100644
> > --- a/drivers/char/ipmi/ipmi_watchdog.c
> > +++ b/drivers/char/ipmi/ipmi_watchdog.c
> > @@ -456,6 +456,71 @@ static int ipmi_set_timeout(int do_heartbeat)
> >         return rv;
> >  }
> >
> > +static unsigned int __ipmi_get_timeout(struct ipmi_smi_msg  *smi_msg,
> > +                                      struct ipmi_recv_msg *recv_msg,
> > +                                      int                  *countdown)
> > +{
> > +       struct kernel_ipmi_msg            msg;
> > +       int                               rv = 0;
> > +       struct ipmi_system_interface_addr addr;
> > +
> > +
> > +       addr.addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE;
> > +       addr.channel = IPMI_BMC_CHANNEL;
> > +       addr.lun = 0;
> > +
> > +       msg.netfn = 0x06;
> > +       msg.cmd = IPMI_WDOG_GET_TIMER;
> > +       msg.data = NULL;
> > +       msg.data_len = 0;
> > +       rv = ipmi_request_supply_msgs(watchdog_user,
> > +                                     (struct ipmi_addr *) &addr,
> > +                                     0,
> > +                                     &msg,
> > +                                     NULL,
> > +                                     smi_msg,
> > +                                     recv_msg,
> > +                                     1);
> > +       if (rv) {
> > +               pr_warn("get timeout error: %d\n", rv);
> > +               return rv;
> > +       }
> > +
> > +       wait_for_completion(&msg_wait);
> > +
> > +       if (recv_msg->msg.data_len < 9) {
> > +               pr_warn("get timeout response size: %d (expected 9)\n",
> > +                       recv_msg->msg.data_len);
> > +               return -EIO;
> > +       }
> > +
> > +       if (recv_msg->msg.data[0] != 0)  {
> > +               pr_warn("get timeout completion code error: %d\n",
> > +                       recv_msg->msg.data[0]);
> > +               return -EIO;
> > +       }
> > +
> > +       *countdown = WDOG_GET_TIMEOUT(recv_msg->msg.data[7],
> > recv_msg->msg.data[8]);
> > +
> > +       return rv;
> > +}
> > +
> > +static int _ipmi_get_timeout(int *countdown)
> > +{
> > +       int rv;
> > +
> > +       if (!watchdog_user)
> > +               return -ENODEV;
> > +
> > +       atomic_set(&msg_tofree, 2);
> > +
> > +       rv = __ipmi_get_timeout(&smi_msg,
> > +                               &recv_msg,
> > +                               countdown);
> > +
> > +       return rv;
> > +}
> > +
> >  static atomic_t panic_done_count = ATOMIC_INIT(0);
> >
> >  static void panic_smi_free(struct ipmi_smi_msg *msg)
> > @@ -729,6 +794,16 @@ static int ipmi_ioctl(struct file *file,
> >                         return -EFAULT;
> >                 return 0;
> >
> > +       case WDIOC_GETTIMELEFT:
> > +               val = 0;
> > +               i = _ipmi_get_timeout(&val);
> > +               if (i)
> > +                       return i;
> > +               i = copy_to_user(argp, &val, sizeof(val));
> > +               if (i)
> > +                       return -EFAULT;
> > +               return 0;
> > +
> >         default:
> >                 return -ENOIOCTLCMD;
> >         }
> > --
> > 2.31.0.rc2.261.g7f71774620-goog
> >
> >

      parent reply	other threads:[~2021-05-05 21:36 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-17  0:10 [PATCH 1/1] ipmi/watchdog: Add WDIOC_GETTIMELEFT ioctl Ryan O'Leary
     [not found] ` <CAKVu-U=Om6CQF7DVkzp487=R_a77Qbi_grKQ7J-HPVZ_QR7DfA@mail.gmail.com>
2021-05-05 21:36   ` Corey Minyard [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=20210505213619.GS18645@minyard.net \
    --to=minyard@acm.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=openipmi-developer@lists.sourceforge.net \
    --cc=ryanoleary@google.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.