Date: Wed, 16 Nov 2005 10:06:55 -0800
From: Greg KH <gregkh@suse.de>
To: Adam Belay <abelay@novell.com>
Cc: Linux-pm mailing list <linux-pm@lists.osdl.org>,
       linux-kernel@vger.kernel.org
Subject: Re: [RFC][PATCH 6/6] PCI PM: pci_save/restore_state improvements
Message-ID: <20051116180655.GC6908@suse.de>
References: <1132111902.9809.59.camel@localhost.localdomain> <20051116063125.GE31375@suse.de> <1132125965.3656.15.camel@localhost.localdomain>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1132125965.3656.15.camel@localhost.localdomain>
User-Agent: Mutt/1.5.11
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2114
Lines: 47

On Wed, Nov 16, 2005 at 02:26:04AM -0500, Adam Belay wrote:
> On Tue, 2005-11-15 at 22:31 -0800, Greg KH wrote:
> > On Tue, Nov 15, 2005 at 10:31:42PM -0500, Adam Belay wrote:
> > > This patch makes some improvements to pci_save_state and
> > > pci_restore_state.  Instead of saving and restoring all standard
> > > registers (even read-only ones), it only restores necessary registers.
> > > Also, the command register is handled more carefully.  Let me know if
> > > I'm missing anything important.
> > > 
> > > 
> > > --- a/drivers/pci/pm.c	2005-11-13 20:32:24.000000000 -0500
> > > +++ b/drivers/pci/pm.c	2005-11-13 20:29:32.000000000 -0500
> > > @@ -53,10 +53,13 @@
> > >   */
> > >  int pci_save_state(struct pci_dev *dev)
> > >  {
> > > -	int i;
> > > -	/* XXX: 100% dword access ok here? */
> > > -	for (i = 0; i < 16; i++)
> > > -		pci_read_config_dword(dev, i * 4,&dev->saved_config_space[i]);
> > > +	struct pci_dev_config * conf = &dev->saved_config;
> > > +
> > > +	pci_read_config_word(dev, PCI_COMMAND, &conf->command);
> > > +	pci_read_config_byte(dev, PCI_CACHE_LINE_SIZE, &conf->cacheline_size);
> > > +	pci_read_config_byte(dev, PCI_LATENCY_TIMER, &conf->latency_timer);
> > > +	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &conf->interrupt_line);
> > 
> > Why are we saving and restoring smaller ammounts of config space now?
> 
> After looking at the spec, it seems that most of the registers we were
> restoring were read-only and couldn't possibly need to be restored.
> Also, the PCI PM spec suggests that only a subset of the registers
> should be restored.  Finally, things like BIST should probably never be
> touched.

Ok, but be aware that this _might_ cause problems for some cards/drivers
that were relying on the old way...  As long as you don't mind me
assigning those bugs to you, I don't have a problem with this :)

thanks,

greg k-h
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/