[oe] Serial Console name for i.mx31

Mon Aug 1 01:44:51 UTC 2011

Charles...

And as I said earlier, I have been looking in drivers/serial/mx.c, that
is where I saw that the function serial_imx_probe was not being called
if you remember  ...

My problem is what bit of code calls that function .. It is NOT in that
module .. I am trying to get to the code that calls the ...probe to see
Why it is not being called ... That is in some other module .. But I
can't trace it due to all the vectored execution .. 

All the CONFIG_xxx defines are defined and yes, it is compiled in,
because if you remember earlier, I said I put debug statements in the ..
Init()
Function and that was being called, I put debug statements in the
...probe() and that is NOT being called .... I see these printk outputs
in the kernel ring buffer ..... All the obvious things are covered
Charles ..

>the driver layers (and hence serial drivers) only get set up later in 
>the piece by serial_imx_probe() once the OS is running.

It's the "later" part I am trying to find ... I repeat, the bit of code
that calls serial_imx_probe(), I may be thick, but I can't trace where
it is ..
All I know is that the driver is "registered" correctly
(platform_register_driver() outputs no error messages), but the ..probe
is not called.

I am trying to find the bit of code that finds that there is an builtin
imx31 serial peripheral and trys to load the driver for it .... The only
code
I could find is the function platform_driver_register() in
../drivers/base which loads all the "platform" drivers .. I am currently
trying to put some debug stuff in that file to see why the driver is not
being loaded/probed ..

I hope I am being clear this time ..

-----Original Message-----
From: Charles Manning [mailto:manningc2 at actrix.gen.nz] 
Sent: Monday, 1 August 2011 1:24 p.m.
To: openembedded-devel at lists.openembedded.org
Cc: Bernard Mentink
Subject: Re: [oe] Serial Console name for i.mx31

Bernie...

As I said, look in the serial driver code. Where that is will depend on
what Linux tree you are using.

On a tree I looked at now it is in drivers/serial/imx.c . It could also
be under drivers/tty/...

As I said, there are setups for two different sets of code in there: one
for 
the console and one for drivers. The console is used during early boot 
because the driver layers (and hence serial drivers) only get set up
later in 
the piece by serial_imx_probe() once the OS is running.

imx_console_xxx functions relate to the console support.
imx_console_setup() sets up the actual console.

There are a bunch of defines CONFIG_SERIAL_IMX_CONSOLE, IMX_CONSOLE etc.
Make 
sure those are configured and all the console code is actually being
compiled 
in and initialised properly.

On Monday 01 August 2011 13:03:16 Bernard Mentink wrote:
> Funny Man!  ..... What code .... The problem is finding the relevent
> bits ... As everything seems to be executed as tables of function
> pointers, it is very hard to fine what links with what ..
>
> (remember Charles, this is my 1st time getting into Embedded linux
code
> .... Give me a break!)
>
> -----Original Message-----
> From: Charles Manning [mailto:manningc2 at actrix.gen.nz]
> Sent: Monday, 1 August 2011 11:34 a.m.
> To: openembedded-devel at lists.openembedded.org
> Cc: Bernard Mentink
> Subject: Re: [oe] Serial Console name for i.mx31
>
> Look at the code.
>
> That will tell you exactly what is going on.
>
> On Monday 01 August 2011 11:24:12 Bernard Mentink wrote:
> > Hi Charles,
> >
> > Thanks, but I thought that it was obvious that I am hooked upto the
> > 1st
> > imx31 serial port as I am using ttymxc0 .. ;-) I have tried seeing
if
> > I am getting output on any other serial port (we have 4 of the 5
> > brought out to connectors ..)
> >
> > The example .config file for the imx shows a kernel argument of
> > "console=ttymxc0,115200" with no parity/bits setting so I don't
think
> > That will make any difference ... But I will try it.
> >
> > Cheers,
> > B.
> >
> >
> >
> >
> > -----Original Message-----
> > From: Charles Manning [mailto:manningc2 at actrix.gen.nz]
> > Sent: Monday, 1 August 2011 11:17 a.m.
> > To: openembedded-devel at lists.openembedded.org
> > Cc: Bernard Mentink
> > Subject: Re: [oe] Serial Console name for i.mx31
> >
> > Bernie
> >
> > You have to hook up to the UART that is wired to what you're
watching.
> > Nobody on this list knows what serial port you are connected to.
> >
> > NB That the early console stuff is set up as a console through
console
> >
> > initialisation. THis is not the same as setting up the serial ports
> > for driver usage.
> >
> > As for the actual setting, it does not look good compared to what I
am
> >
> > used to because it does not set all the parameters.
> >
> > console=ttymxc0,115200n8
> >
> > But what actually is used will depend on the console setup (search
for
> >
> > console in the imx31 serial driver code.
> >
> > On Monday 01 August 2011 10:44:07 Bernard Mentink wrote:
> > > Hi,
> > >
> > > Can anyone verify what the kernel command line option should be to
> > > use the first imx31 uart? As far as I can tell it is
> > > "console=ttymxc0,115200"
> > >
> > > The reason I ask, is that I am seeing the kernel messages in the
ram
> > >
> > > kernel ring buffer, but not seing them come out the serial
console.
> > >
> > > I am porting linux to our own hardware which is similar to the
> > > LogicPD LiteKit platform (I have compiled the kernel for this
> > > platform ..)
> > >
> > > Debugging into the serial driver in ../drivers/serial/mx31  .. I
see
> >
> > the
> >
> > > function init_serial_init() being called ( .. and it's printk
> > > message
> >
> > is
> >
> > > in the buffer) but I am not seeing the serial_imx_probe() being
> > > called ..
> > >
> > > I do not understand the the registering of platform drivers enough
> > > to work out how the kernel connects the diver with the hardware
....
> > >
> > > Can someone enlighten me on how to debug this further ..... Or
even
> > > better just tell me to use the correct command line option and it
> > > all bursts into life  :-)
> > >
> > > Thanks,
> > > Bernie
> > >
> > > In addition:
> > >
> > > I added the earlyprintk command to the kernel command line and I
do
> >
> > see
> >
> > > these messages. i.e:
> > >
> > > Starting kernel ...
> > >
> > > Uncompressing Linux... done, booting the kernel.
> > > Linux version 2.6.36.1 (bmentink at 2570-ArchBox) (gcc version 4.5.3
> > > 20110311 (prerelease) (GCC) ) #7 P
> > > REEMPT Mon Aug 1 09:51:34 NZST 2011
> > > CPU: ARMv6-compatible processor [4107b364] revision 4 (ARMv6TEJ),
> > > cr=00c5387f
> > > CPU: VIPT nonaliasing data cache, VIPT nonaliasing instruction
cache
> > > Machine: LogicPD i.MX31 SOM
> > > bootconsole [earlycon0] enabled
> > > Memory policy: ECC disabled, Data cache writeback Built 1
zonelists
> > > in Zone order, mobility grouping on.  Total pages:
> > > 32512
> > > Kernel command line: console=ttymxc,115200
> > > earlyprintk=serial,ttymxc0,115200,keep
> > > PID hash table entries: 512 (order: -1, 2048 bytes) Dentry cache
> > > hash table entries: 16384 (order: 4, 65536 bytes) Inode-cache hash
> > > table entries: 8192 (order: 3, 32768 bytes)
> > > Memory: 128MB = 128MB total
> > > Memory: 126744k/126744k available, 4328k reserved, 0K highmem
> > > Virtual kernel memory layout:
> > >     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
> > >     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
> > >     DMA     : 0xffc00000 - 0xffe00000   (   2 MB)
> > >     vmalloc : 0xc8800000 - 0xf4000000   ( 696 MB)
> > >     lowmem  : 0xc0000000 - 0xc8000000   ( 128 MB)
> > >     modules : 0xbf000000 - 0xc0000000   (  16 MB)
> > >       .init : 0xc0008000 - 0xc001e000   (  88 kB)
> > >       .text : 0xc001e000 - 0xc02e1000   (2828 kB)
> > >       .data : 0xc02f6000 - 0xc030c6e0   (  90 kB)
> > > Hierarchical RCU implementation.
> > >         RCU-based detection of stalled CPUs is disabled.
> > >         Verbose stalled-CPUs detection is disabled.
> > > NR_IRQS:180
> > > MXC IRQ initialized
> > > MXC GPIO hardware
> > >
> > > So it seems there is nothing wrong with the "ttymxc0" port
> > > assignment ... It's just the main serial console is just not
working
>
> ..
>
> > > Cheers,
> > > B.
> > >
> > >
> > >
> > >
> > > _______________________________________________
> > > Openembedded-devel mailing list
> > > Openembedded-devel at lists.openembedded.org
> > >
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-dev
> > > el
> > >
> > > _______________________________________________
> > > Openembedded-devel mailing list
> > > Openembedded-devel at lists.openembedded.org
> > >
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-dev
> > > el
> >
> > _______________________________________________
> > Openembedded-devel mailing list
> > Openembedded-devel at lists.openembedded.org
> >
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel