serial_cs: oxsemi quirk breaks resume

serial_cs: oxsemi quirk breaks resume

[Pavel Machek]

Quirk is applied on all cards with given manfid (is it that
correct?). Unfortunately, that quirk breaks resume on zaurus with
billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.

Signed-off-by: Pavel Machek <pavel@ucw.cz>

--- a/drivers/serial/serial_cs.c	2009-10-06 13:51:46.000000000 +0200
+++ b/drivers/serial/serial_cs.c	2009-11-09 14:43:25.000000000 +0100
@@ -158,7 +158,8 @@
 {
 	struct serial_info *info = link->priv;
 
-	outb(12, info->c950ctrl + 1);
+	if (info->c950ctrl)
+		outb(12, info->c950ctrl + 1);
 }
 
 /* request_region? oxsemi branch does no request_region too... */


-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: serial_cs: oxsemi quirk breaks resume

[Alan Cox]

On Tue, 10 Nov 2009 08:54:11 +0100
Pavel Machek <pavel@ucw.cz> wrote:

> 
> Quirk is applied on all cards with given manfid (is it that
> correct?). Unfortunately, that quirk breaks resume on zaurus with
> billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.

So how does it end up as zero, do you have a race between insert/remove
somewhere ?

Re: serial_cs: oxsemi quirk breaks resume

[Pavel Machek]

On Tue 2009-11-10 10:32:59, Alan Cox wrote:
> On Tue, 10 Nov 2009 08:54:11 +0100
> Pavel Machek <pavel@ucw.cz> wrote:
> 
> > 
> > Quirk is applied on all cards with given manfid (is it that
> > correct?). Unfortunately, that quirk breaks resume on zaurus with
> > billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.
> 
> So how does it end up as zero, do you have a race between insert/remove
> somewhere ?

It does not look like race. I believe it is simply not a multiport
card. (info->multi == 1). ... ... confirmed by printks.
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: serial_cs: oxsemi quirk breaks resume

[Pavel Machek]

On Tue 2009-11-10 08:54:10, Pavel Machek wrote:
> 
> Quirk is applied on all cards with given manfid (is it that
> correct?). Unfortunately, that quirk breaks resume on zaurus with
> billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.
> 
> Signed-off-by: Pavel Machek <pavel@ucw.cz>

I believe this patch got enough acks and is waiting in the tree
somewhere (greg?) (possibly with slightly different changelog), but I
don't see it in 33-rc4. Could it get pushed? It is quite simple
bugfix, and is important for my zaurus...  Pavel

> --- a/drivers/serial/serial_cs.c	2009-10-06 13:51:46.000000000 +0200
> +++ b/drivers/serial/serial_cs.c	2009-11-09 14:43:25.000000000 +0100
> @@ -158,7 +158,8 @@
>  {
>  	struct serial_info *info = link->priv;
>  
> -	outb(12, info->c950ctrl + 1);
> +	if (info->c950ctrl)
> +		outb(12, info->c950ctrl + 1);
>  }
>  
>  /* request_region? oxsemi branch does no request_region too... */
> 
> 

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: serial_cs: oxsemi quirk breaks resume

[Greg KH]

On Fri, Jan 15, 2010 at 09:01:00PM +0100, Pavel Machek wrote:
> On Tue 2009-11-10 08:54:10, Pavel Machek wrote:
> > 
> > Quirk is applied on all cards with given manfid (is it that
> > correct?). Unfortunately, that quirk breaks resume on zaurus with
> > billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.
> > 
> > Signed-off-by: Pavel Machek <pavel@ucw.cz>
> 
> I believe this patch got enough acks and is waiting in the tree
> somewhere (greg?) (possibly with slightly different changelog), but I
> don't see it in 33-rc4. Could it get pushed? It is quite simple
> bugfix, and is important for my zaurus...  Pavel

I thought others complained about it.

If not, care to resend it to me, with the acks added to it, so that I
can forward it on?

thanks,

greg k-h

Re: serial_cs: oxsemi quirk breaks resume

[Greg KH]

On Fri, Jan 15, 2010 at 03:53:33PM -0800, Greg KH wrote:
> On Fri, Jan 15, 2010 at 09:01:00PM +0100, Pavel Machek wrote:
> > On Tue 2009-11-10 08:54:10, Pavel Machek wrote:
> > > 
> > > Quirk is applied on all cards with given manfid (is it that
> > > correct?). Unfortunately, that quirk breaks resume on zaurus with
> > > billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.
> > > 
> > > Signed-off-by: Pavel Machek <pavel@ucw.cz>
> > 
> > I believe this patch got enough acks and is waiting in the tree
> > somewhere (greg?) (possibly with slightly different changelog), but I
> > don't see it in 33-rc4. Could it get pushed? It is quite simple
> > bugfix, and is important for my zaurus...  Pavel
> 
> I thought others complained about it.
> 
> If not, care to resend it to me, with the acks added to it, so that I
> can forward it on?

Oh nevermind, it's in my tree already, I'll include it in the next round
of patches to go to Linus (looking like Monday now, too late for
today...)

thanks,

greg k-h

Re: serial_cs: oxsemi quirk breaks resume

[Pavel Machek]

On Fri 2010-01-15 15:54:25, Greg KH wrote:
> On Fri, Jan 15, 2010 at 03:53:33PM -0800, Greg KH wrote:
> > On Fri, Jan 15, 2010 at 09:01:00PM +0100, Pavel Machek wrote:
> > > On Tue 2009-11-10 08:54:10, Pavel Machek wrote:
> > > > 
> > > > Quirk is applied on all cards with given manfid (is it that
> > > > correct?). Unfortunately, that quirk breaks resume on zaurus with
> > > > billionton bluetooth card inserted: c950ctrl is 0 and outb() faults.
> > > > 
> > > > Signed-off-by: Pavel Machek <pavel@ucw.cz>
> > > 
> > > I believe this patch got enough acks and is waiting in the tree
> > > somewhere (greg?) (possibly with slightly different changelog), but I
> > > don't see it in 33-rc4. Could it get pushed? It is quite simple
> > > bugfix, and is important for my zaurus...  Pavel
> > 
> > I thought others complained about it.
> > 
> > If not, care to resend it to me, with the acks added to it, so that I
> > can forward it on?
> 
> Oh nevermind, it's in my tree already, I'll include it in the next round
> of patches to go to Linus (looking like Monday now, too late for
> today...)

Thanks!

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html