sunzilog possible bug

sunzilog possible bug

[Benjamin Herrenschmidt]

Hi !

I found recently a bug in pmac_zilog, which originates from sunzilog who so might
have the same problem. The "timeout" member of the port structure was never
initialized, causing very long pauses to happen on port close, that is pretty much
all the time when hitting userland.

Quick untested patch (I don't have a sparc)

=== drivers/serial/sunzilog.c 1.46 vs edited ==--- 1.46/drivers/serial/sunzilog.c	2004-09-21 06:46:03 +10:00
+++ edited/drivers/serial/sunzilog.c	2004-11-07 10:05:37 +11:00
@@ -974,6 +974,8 @@
 
 	sunzilog_maybe_update_regs(up, ZILOG_CHANNEL_FROM_PORT(port));
 
+	uart_update_timeout(port, termios->c_cflag, baud);
+
 	spin_unlock_irqrestore(&up->port.lock, flags);
 }
 

Re: sunzilog possible bug

[William Lee Irwin III]

On Sun, Nov 07, 2004 at 10:34:49AM +1100, Benjamin Herrenschmidt wrote:
> I found recently a bug in pmac_zilog, which originates from sunzilog
> who so might have the same problem. The "timeout" member of the port
> structure was never initialized, causing very long pauses to happen
> on port close, that is pretty much all the time when hitting userland.
> Quick untested patch (I don't have a sparc)

Thanks ben.

If those having trouble with serial ports could try this patch and report
success (or failure if that's the case) I would be much obliged.


-- wli

Re: sunzilog possible bug

[David S. Miller]

On Sun, 07 Nov 2004 10:34:49 +1100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> I found recently a bug in pmac_zilog, which originates from sunzilog who so might
> have the same problem. The "timeout" member of the port structure was never
> initialized, causing very long pauses to happen on port close, that is pretty much
> all the time when hitting userland.
> 
> Quick untested patch (I don't have a sparc)

Looks good to me, I'll apply this to my tree.

Thanks Benjamic.