[PATCH] Documentation: devices.txt: reconcile serial/ucc_uart minor numers

[Christophe Leroy]

Le 01/08/2023 à 07:30, Randy Dunlap a écrit :
> Hi Christophe,
> 
> On 7/31/23 22:21, Christophe Leroy wrote:
>>
>>
>> Le 24/07/2023 à 08:33, Randy Dunlap a écrit :
>>> Reconcile devices.txt with serial/ucc_uart.c regarding device number
>>> assignments. ucc_uart.c supports 4 ports and uses minor devnums
>>> 46-49, so update devices.txt with that info.
>>> Then update ucc_uart.c's reference to the location of the devices.txt
>>> list in the kernel source tree.
>>
>> Devices ttyCPM* belong to cpm_uart driver. As explained in the comment
>> you have modified in UCC uart driver, UCC uart borrows those devices and
>> shall not be considered as the reference. But the lines you modify in
>> device.txt doesn't mention QE UCC, it mentions CPM SCC and CPM SMC.
>>
>> CPM uart driver supports up to 6 ports (4 SCC and 2 SMC).
>>
> 
> Thank you for replying.
> 
> Does this mean that the patch should be reverted?

Not sure it should be reverted, the best would be to fix it up because 
anyway what we had was wrong.

See below.

> 
>> On one of my boards which has a powerpc mpc866 CPU, I have:
>>
>> [    2.393872] ff000a80.serial: ttyCPM0 at MMIO 0xfff00a80 (irq = 19,
>> base_baud = 8250000) is a CPM UART
>> [    2.411899] ff000a90.serial: ttyCPM1 at MMIO 0xfff00a90 (irq = 20,
>> base_baud = 8250000) is a CPM UART
>> [    2.430352] ff000a00.serial: ttyCPM2 at MMIO 0xfff00a00 (irq = 30,
>> base_baud = 8250000) is a CPM UART
>> [    2.448944] ff000a20.serial: ttyCPM3 at MMIO 0xfff00a20 (irq = 29,
>> base_baud = 8250000) is a CPM UART
>> [    2.467435] ff000a40.serial: ttyCPM4 at MMIO 0xfff00a40 (irq = 28,
>> base_baud = 8250000) is a CPM UART
>> [    2.485924] ff000a60.serial: ttyCPM5 at MMIO 0xfff00a60 (irq = 27,
>> base_baud = 8250000) is a CPM UART
>>
>> # ll /dev/ttyCPM*
>> crw-------    1 root     root      204,  46 Jan  1 01:01 /dev/ttyCPM0
>> crw-------    1 root     root      204,  47 Jan  1 01:00 /dev/ttyCPM1
>> crw-------    1 root     root      204,  48 Jan  1 01:00 /dev/ttyCPM2
>> crw-------    1 root     root      204,  49 Jan  1 01:00 /dev/ttyCPM3
>> crw-------    1 root     root      204,  50 Jan  1 01:00 /dev/ttyCPM4
>> crw-------    1 root     root      204,  51 Jan  1 01:00 /dev/ttyCPM5
> 
> I don't see minors 50-51 allocated in devices.txt for use by this device/driver.
> Am I overlooking that allocation somewhere?

I looked into history.

ttyCPM do not appear in devices.txt in v2.6.9.

It was introduced in v2.6.10 as :

		 46 = /dev/ttyCPM0		PPC CPM (SCC or SMC) - port 0
		    ...
		 49 = /dev/ttyCPM5		PPC CPM (SCC or SMC) - port 5

Then in v2.6.11 we find:

		 46 = /dev/ttyCPM0		PPC CPM (SCC or SMC) - port 0
		    ...
		 47 = /dev/ttyCPM5		PPC CPM (SCC or SMC) - port 5
		 50 = /dev/ttyIOC40		Altix serial card
		    ...
		 81 = /dev/ttyIOC431		Altix serial card

So it looks like from the begining the intention was to have 6 ports 
(from 0 to 5) and it should have been up to 51. For some reason it was 
messed up.

If you look into v2.6.10 drivers/serial/cpm_uart/cpm_uart.h you see 
there are 6 ports. Even in v2.6.9 that driver exists and has 6 ports.


Anyway, the ttyIOC* were removed by commit a017ef17cfd8 ("tty/serial: 
remove the ioc4_serial driver"), so it should be no problem to fix 
devices.txt and set ttyCPM from 46 to 51.

Christophe


> 
> Thanks for your help.
> 
>>
>> # cat /proc/tty/drivers
>> /dev/tty             /dev/tty        5       0 system:/dev/tty
>> /dev/console         /dev/console    5       1 system:console
>> /dev/ptmx            /dev/ptmx       5       2 system
>> ttyCPM               /dev/ttyCPM   204 46-51 serial
>> pty_slave            /dev/pts      136 0-1048575 pty:slave
>> pty_master           /dev/ptm      128 0-1048575 pty:master
>> pty_slave            /dev/ttyp       3 0-4 pty:slave
>> pty_master           /dev/pty        2 0-4 pty:master
>>
>> Christophe
>>
>>
>>>
>>> Fixes: d7584ed2b994 ("[POWERPC] qe-uart: add support for Freescale QUICCEngine UART")
>>> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
>>> Signed-off-by: Randy Dunlap <rdunlap at infradead.org>
>>> Cc: Timur Tabi <timur at kernel.org>
>>> Cc: Kumar Gala <galak at kernel.crashing.org>
>>> Cc: linuxppc-dev at lists.ozlabs.org
>>> Cc: Greg Kroah-Hartman <gregkh at linuxfoundation.org>
>>> Cc: Jiri Slaby <jirislaby at kernel.org>
>>> Cc: linux-serial at vger.kernel.org
>>> Cc: Jonathan Corbet <corbet at lwn.net>
>>> Cc: linux-doc at vger.kernel.org
>>> ---
>>>    Documentation/admin-guide/devices.txt |    2 +-
>>>    drivers/tty/serial/ucc_uart.c         |    2 +-
>>>    2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff -- a/drivers/tty/serial/ucc_uart.c b/drivers/tty/serial/ucc_uart.c
>>> --- a/drivers/tty/serial/ucc_uart.c
>>> +++ b/drivers/tty/serial/ucc_uart.c
>>> @@ -59,7 +59,7 @@ static int firmware_loaded;
>>>    /* #define LOOPBACK */
>>>    
>>>    /* The major and minor device numbers are defined in
>>> - * http://www.lanana.org/docs/device-list/devices-2.6+.txt.  For the QE
>>> + * Documentation/admin-guide/devices.txt.  For the QE
>>>     * UART, we have major number 204 and minor numbers 46 - 49, which are the
>>>     * same as for the CPM2.  This decision was made because no Freescale part
>>>     * has both a CPM and a QE.
>>> diff -- a/Documentation/admin-guide/devices.txt b/Documentation/admin-guide/devices.txt
>>> --- a/Documentation/admin-guide/devices.txt
>>> +++ b/Documentation/admin-guide/devices.txt
>>> @@ -2691,7 +2691,7 @@
>>>    		 45 = /dev/ttyMM1		Marvell MPSC - port 1 (obsolete unused)
>>>    		 46 = /dev/ttyCPM0		PPC CPM (SCC or SMC) - port 0
>>>    		    ...
>>> -		 47 = /dev/ttyCPM5		PPC CPM (SCC or SMC) - port 5
>>> +		 49 = /dev/ttyCPM5		PPC CPM (SCC or SMC) - port 3
>>>    		 50 = /dev/ttyIOC0		Altix serial card
>>>    		    ...
>>>    		 81 = /dev/ttyIOC31		Altix serial card
>