Broken serial port

Broken serial port

[Laurent Vivier]

[-- Attachment #1: Type: text/plain, Size: 1161 bytes --]

Hi,

I'm not subscribed to these mailing lists, so CC: me.

I'm playing for some days with a kernel 3.15 on an old mac performa 575
and it seems that the serial port is broken. When I log on using the
console on the serial port it seems there is an screen refresh added
after each character I send.

A little demo :

http://asciinema.org/a/10211 [at 3:20]

I bisect and I found this commit :

commit 717f3bbab3c7628736ef738fdbf3d9a28578c26c
Author: Seth Bollinger <sethb@digi.com>
Date:   Tue Mar 25 12:55:37 2014 -0500

    serial_core: Fix conditional start_tx on ring buffer not empty

    If the serial_core ring buffer empties just as the tty layer receives
    an XOFF, then start_tx will never be called when the tty layer
    receives an XON as the serial_core ring buffer is empty.  This will
    possibly leave a few bytes trapped in the fifo for drivers that
    disable the transmitter when flow controlled.

    Signed-off-by: Seth Bollinger <sethb@digi.com>
    Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>


If I revert it, all works fine :

http://asciinema.org/a/10212

Regards,
Laurent


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 246 bytes --]

Re: Broken serial port

[Peter Hurley]

On 06/17/2014 02:48 PM, Laurent Vivier wrote:
> Hi,
>
> I'm not subscribed to these mailing lists, so CC: me.
>
> I'm playing for some days with a kernel 3.15 on an old mac performa 575
> and it seems that the serial port is broken. When I log on using the
> console on the serial port it seems there is an screen refresh added
> after each character I send.

What uart driver is this happening on?

This is a known bug being worked on for these uart drivers:
sunsab
sunzilog
ip22zilog
pmac_zilog

Regards,
Peter Hurley

> A little demo :
>
> http://asciinema.org/a/10211 [at 3:20]
>
> I bisect and I found this commit :
>
> commit 717f3bbab3c7628736ef738fdbf3d9a28578c26c
> Author: Seth Bollinger <sethb@digi.com>
> Date:   Tue Mar 25 12:55:37 2014 -0500
>
>      serial_core: Fix conditional start_tx on ring buffer not empty
>
>      If the serial_core ring buffer empties just as the tty layer receives
>      an XOFF, then start_tx will never be called when the tty layer
>      receives an XON as the serial_core ring buffer is empty.  This will
>      possibly leave a few bytes trapped in the fifo for drivers that
>      disable the transmitter when flow controlled.
>
>      Signed-off-by: Seth Bollinger <sethb@digi.com>
>      Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>
>
> If I revert it, all works fine :
>
> http://asciinema.org/a/10212
>
> Regards,
> Laurent
>

Re: Broken serial port

[Laurent Vivier]

[-- Attachment #1: Type: text/plain, Size: 1942 bytes --]

Le 17/06/2014 21:28, Peter Hurley a écrit :
> On 06/17/2014 02:48 PM, Laurent Vivier wrote:
>> Hi,
>>
>> I'm not subscribed to these mailing lists, so CC: me.
>>
>> I'm playing for some days with a kernel 3.15 on an old mac performa 575
>> and it seems that the serial port is broken. When I log on using the
>> console on the serial port it seems there is an screen refresh added
>> after each character I send.
> 
> What uart driver is this happening on?
> 
> This is a known bug being worked on for these uart drivers:
> sunsab
> sunzilog
> ip22zilog
> pmac_zilog

Seems to be pmac_zilog, from dmesg :

pmac_zilog: 0.6 (Benjamin Herrenschmidt <benh@kernel.crashing.org>)
scc.0: ttyS0 at MMIO 0x50f0c022 (irq = 4, base_baud = 230400) is a
Z85c30 ESCC -
 Serial port
scc.1: ttyS1 at MMIO 0x50f0c020 (irq = 4, base_baud = 230400) is a
Z85c30 ESCC -
 Serial port

If needed, I can test the fix.

Thank you.

Regards,
Laurent

> Regards,
> Peter Hurley
> 
>> A little demo :
>>
>> http://asciinema.org/a/10211 [at 3:20]
>>
>> I bisect and I found this commit :
>>
>> commit 717f3bbab3c7628736ef738fdbf3d9a28578c26c
>> Author: Seth Bollinger <sethb@digi.com>
>> Date:   Tue Mar 25 12:55:37 2014 -0500
>>
>>      serial_core: Fix conditional start_tx on ring buffer not empty
>>
>>      If the serial_core ring buffer empties just as the tty layer
>> receives
>>      an XOFF, then start_tx will never be called when the tty layer
>>      receives an XON as the serial_core ring buffer is empty.  This will
>>      possibly leave a few bytes trapped in the fifo for drivers that
>>      disable the transmitter when flow controlled.
>>
>>      Signed-off-by: Seth Bollinger <sethb@digi.com>
>>      Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>
>>
>> If I revert it, all works fine :
>>
>> http://asciinema.org/a/10212
>>
>> Regards,
>> Laurent
>>
> 



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 246 bytes --]