Foo Tze Wan | 1 Aug 02:28 2006
Picon

Re: AT91 - Achieving UART baud rate of 921600 bps

Chip and all,

Just to update the group after about 2 weeks...
Thanks for all the assistance!

History:
My group had wanted to test the AT91RM9200 UART transmission and reception 
at 921600 bps. (Send from one UART and receive on another.) I am using the 
2.6.17 kernel + the PDC patch from http://maxim.org.za.

Status:
I have managed to compile, but was still not able to run well without errors 
at 921600 bps. My group found that the TTL to RS232 transceiver on the DK 
board was spec at maximum 230400 bps. This could be causing some errors. We 
will be trying to get a transceiver with higher max baud to try.

Thanks again,
Tze Wan

On Fri, 14 Jul 2006, Chip Coldwell wrote:

> If you need a cross-compiler, you can get the one I built and used
> from
>
> http://frank.harvard.edu/~coldwell/toolchain/
>
> The easiest thing to do is download the rpm packages; they were built
> on Red Hat Enterprise Linux 4, but should work fine on Fedora Core 5
> or SuSE -- if they don't, let me know.
>
(Continue reading)

ken | 1 Aug 03:32 2006
Picon

UDMA cannot burst initiation(Linux-2.6.8) on EP9315

Hello all:
 The development environment I used is: 
  a. Linux kernel version: Linux-2.6.8.1(ep9315 patch: linux-2.6.8.1-crus-1.07);
  b. ARM: EP9315
  c. IDE disk: MINI IDE Flash Disk
I configure Linux kernel BLK_DEV_IDE_DMA_EP93XX=y, and compiling is ok. But when the kernel is booting,
the IDE's UDMA mode can not initual correctly. Following is the kernel debug message:

ide0 at 0x800-0x807,0x406 on irq 40
hda: max request size: 128KiB
hda: 6094368 sectors (3120 MB), CHS=6046/16/63ep93xx_ide_dma_verbose

 /dev/ide/host0/bus0/target0/lun0:ep93xx_ide_dma_read
DMA Config 
        1 dmas pending
        add buf: handle=0xa0000001, base=0x800a0024, dest=0x05f44000, ep93xx_dma_add_buffer
dma_get_channel_from_handle
command=0xc8
ep93xx_set_pio
ep93xx_ide_dma_begin
ep93xx_rwproc
ep93xx_set_pio
rwproc in udma mode (udma starts at 64): 66
rwproc in udma 2, action is 0
ep93xx_dma_start
dma_get_channel_from_handle
DMA Start 
dma_start_m2m
DMA - It's been started!!CONTROL - 0x3c0540c 
STATUS - 0x1013 
(Continue reading)

stanliao | 1 Aug 10:34 2006

any document describe the macroes in the Makefile


Dear All,

Is there any document describing the macroes in the Linux kernel's
Makefiles?
e.g. O_TARGET, export-objs, mod-subdirs, obj-y, suddir-y, subdir-m.
I'm trying to learn from existing examples but I have no confidence whether
or not I use it correctly. Would anybody help on this? Thanks.

Best Regards,
Stan Liao

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php

Ram | 1 Aug 11:29 2006
Picon

u-boot and CCS (off topic)

Hi,
   apologies for this is offtopic - the only thing that is related is my
target processor is ARM.

   Im using CCS to compile a hello world program (CCS is TI Code composer
studio).

   The output CCS produces is a hello.out file (COFF format).

    I then used a coff2elf converter to convert into an elf.

   Now, im trying to run it using u-boot. - I first download the
hello.elfto RAM and then
    go the entry point of my program (some address).

   However, as soon as i do this - i get "undefined instruction"

  Is this possible?. compilation using CCS and loading the program using
u-boot 1.1.4?

   What else do you think needs to be done?

Regards,
sriram

   undefined instruction
pc : [<10c08060>]    lr : [<110861dc>]
sp : 1103fc44  ip : 0080c010  fp : 00000001
r10: 10000000  r9 : 00000000  r8 : 1103ffdc
r7 : 00000400  r6 : 01000001  r5 : 0ff3bc7a  r4 : 10000040
(Continue reading)

Russell King - ARM Linux | 1 Aug 11:39 2006
Picon

Re: any document describe the macroes in the Makefile

On Tue, Aug 01, 2006 at 04:34:27PM +0800, stanliao <at> globalunichip.com wrote:
> Is there any document describing the macroes in the Linux kernel's
> Makefiles?
> e.g. O_TARGET, export-objs, mod-subdirs, obj-y, suddir-y, subdir-m.
> I'm trying to learn from existing examples but I have no confidence whether
> or not I use it correctly. Would anybody help on this? Thanks.

Those are for the 2.4 kernel.  I've forgotten all about such kernels
long ago.

	http://www.arm.linux.org.uk/mailinglists/faq.php#f11

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php

Marco Cavallini | 1 Aug 12:25 2006

Re: [PATCH] AT91RM9200 slow clock suspend

Savin Zlobec ha scritto:
 > I've been working on low power suspend mode which switches
 > the CPU to 32Khz clock and turns off  the main oscillater and PLLA.
 > I'am attaching my patches which are highly experimental, but work very
 > well for me - i.e. no SDRAM corruption and no wakeup failures :-)
 >
 > Attached is also a patch for arch/arm/mach-at91rm9200/time.c
 > which removes the annoying, but harmless (?) 'BUG: soft lockup
 > detected on CPU#0!' message when exiting suspended states.
 >
 > And a patch for drivers/usb/host/ohci-at91.c which fixes the increase
 > in user count of usb clocks when mixing 'standby' and 'mem' suspend
 > states resulting in suspend failures. Inside is also a small fix for
 > disabling
 > the second usb port on non-BGA parts - I've got one.
 >

Hi Savin,
thank you for the patch.

I am testing your patch on my custom AT91 board,
I can confirm that the "soft lockup" message message no longer appears.

About low power tests I am facing to problems with
  echo mem > /sys/power/state

normal 	189 mA	
standby 169 mA	wake up with chars from debug port
mem 	144 mA doesn't wake up :-(

(Continue reading)

Holger Schurig | 1 Aug 13:56 2006
Picon

Re: u-boot and CCS (off topic)

>    Im using CCS to compile a hello world program (CCS is TI Code
> composer studio).

Ask TI for support.

Ask in the right mailing list (U-Boot has it's own mailing list).

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php

David Brownell | 1 Aug 18:05 2006
Picon

Re: Using ic2 chip from another driver

On Monday 31 July 2006 1:06 am, Avinash Bhatia wrote:
> > 
> >	 But how can i use this existing driver from 
> > another driver or do i have to implement the control of the
> > pca9539 again in my driver ?
> 
> I am running into a similar design issue - apparently there's
> no 'established' or clean way to do it. 

I've had to do that with a few I2C drivers, and the solution
adopted there was to just define an API and export it through
a driver-specific #include <...> file.  See how the tps65010
and isp1301_omap drivers do that.

I don't know what you mean by "clean" there; it's not like
there's another solution than exposing an API!!

The tricky bits involve init sequencing (I2C drivers are way
late in the static init sequence), API design (the usual stuff),
chip addressing where that make sense (e.g. multiple i2c i/o
expanders could be identified by board specific #defines for
the chip addresses, but only one OTG transceiver may exist on
a system), and similar integration issues.

- Dave

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php
(Continue reading)

David Brownell | 1 Aug 17:54 2006
Picon

Re: [PATCH] AT91RM9200 slow clock suspend


> normal  189 mA	
> standby 169 mA    wake up with chars from debug port
> mem     144 mA    doesn't wake up :-(

"mem" doesn't wake for me either, or at least not so far as getting
to Linux again.

The same wakeup events should work as would kick in with

	echo mem > /sys/power/state

as would kick in _before_ applying the patch that adds slow clock mode
to "mem" ... which wouldn't include DBGU (unless it's reclocked to run
with MCK=32kHz), but can include things like GPIO irqs (for MMC or CF
insert/remove events) or IRQ0 (maybe that's conveniently hooked up to
a pushbutton).

Re "standby" not saving so much power, you should be able to save a
few more mA per device you've told not to be a wakeup event source,
e.g. by "echo disabled > /sys/devices/.../power/wakeup".

- Dave

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php

(Continue reading)

stanliao | 1 Aug 10:42 2006

not see ".depend" file generated after "make dep"


Dear All,

I put my driver ("Config.in" and "Makefile" also) into a new directory (say
.../drivers/mydrv) and this driver can be compiled automatically. However,
the ".depend" file isn't seen after "make dep". Excuse me, how to get this
".depend" file? Thanks.

Best Regards,
Stan Liao

-------------------------------------------------------------------
List admin: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ:        http://www.arm.linux.org.uk/mailinglists/faq.php
Etiquette:  http://www.arm.linux.org.uk/mailinglists/etiquette.php


Gmane