[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20221122094422.000032f0@Huawei.com>
Date: Tue, 22 Nov 2022 09:44:22 +0000
From: Jonathan Cameron <Jonathan.Cameron@...wei.com>
To: Davidlohr Bueso <dave@...olabs.net>
CC: <linux-cxl@...r.kernel.org>, <dan.j.williams@...el.com>,
<bwidawsk@...nel.org>, <ira.weiny@...el.com>,
<alison.schofield@...el.com>, <vishal.l.verma@...el.com>,
<a.manzanares@...sung.com>, <mcgrof@...nel.org>,
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] cxl/mbox: Add background operation handling
machinery
On Mon, 21 Nov 2022 13:48:04 -0800
Davidlohr Bueso <dave@...olabs.net> wrote:
> On Thu, 25 Aug 2022, Jonathan Cameron wrote:
>
> >> +/*
> >> + * Ensure that ->mbox_send() can run safely when a background
> >> + * command is running. If so, returns zero, otherwise -EBUSY.
> >
> >As above, I'm not sure this is necessary given you should get
> >a sensible error code in response to any such message.
> >E.g. media disabled.
>
> So this is necessary to prevent mischief in the polling case
> where you can have windows where the driver is out of sync
> with the hardware - hw finishes async command but driver
> (sleeping poller) has not acknowledged this. In the irq case,
> yeah this would be redundant and not needed.
>
Ok. A worked example might be useful for where this matters
at a higher level. For example, if there is no conflict at
the hardware level what goes wrong in the software as a result
of dropping the protection on the one that we haven't yet detected
as finished...
Jonathan
> Thanks,
> Davidlohr
Powered by blists - more mailing lists