Yong-iL Joh | 1 Sep 2003 08:32

Re: Problems with pxa-ac97 sleep

Ashwin Patwekar wrote:
>
> We had a simple application which did the same. We observed that the 
> control reached the application and as soon as application called sleep 
> every thing freezed.

check DMA handler in pxa-audio.c

--

-- 
Joh, Yong-iL
E-mail: tolkien <at> nownuri.net tolkien <at> mizi.com

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

Andre | 1 Sep 2003 10:45
Picon
Favicon

Re: boot from CF, lilo style?

Tom Walsh <tom <at> openhardware.net> wrote:
> 
> I've done work with CF before and have booted from it on an
> MC68EZ328 system I designed, but that was using the FAT
> filesystem.  I would guess that ext3 filesystem structure is
> going to be a bit more complex?
> 

You could always partition the CF card to create a dedicated
partition for the kernel. If you only need to store one kernel, you
could store it directly in the partition - ie no file system driver
required in the bootloader.

Andre
--

__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

jiunnder | 1 Sep 2003 11:52
Picon

Fw: what the kernel core dump means while I was executing "/sbin/init"

I bought a Altera EPXA10DB development board and want to make some demo on
this board.
I download the kernel form
ftp://ftp.kernel.org/pub/linux/kernel/v2.4/linux-2.4.19.tar.gz
ftp://ftp.arm.linux.org.uk/pub/armlinux/kernel/v2.4/patch-2.4.19-rmk6.gz
And used the patches for EPXA10DB:
1385/1
1386/1
1387/1
1388/1
1413/1
Building the kernel with arm-linux tool chain, and mount the Monta Vista
ARM920T root file distribution
After that, I find the libc of the root file system  are incompatible with
the arm-linux kernel libc.
So I change arm-linux tool chain to the Monta Vista tool chain to re-build
it
again.
This time it doesn't implys the libc is incompatible.
But the kernel panics while executing the "/sbin/init" program.

    execve("/sbin/init",argv_init,envp_init);

The kernel dump following messages, but I don't know what's happened.
How should I do? I need some suggestions to resolve this problem?

Sorry for my rude requirement and thank you in advance.

best regards,
Jiunnder
(Continue reading)

Picon

what's the problem with DSR module on IXPD425 board?

Hi,all

I have booted my own ixp425 board(designed according to IXDP425 DEMO board)
with redboot + linux2.4.19 + patch-linux-2.4.19-rmk7-ds2. 
Till i insmod csr.o, ixp425_eth.o and  dsr1_1_module.o, everything seems OK. 
while i run a very simple module which only has a "xDspInit()" fuction to init DSP,
kernel will has a pointer error such as ----

ASSERTION 'pMem != NULL' failed.

Since there is no open codes of DSR, i dont know what's the exact problem. 
I have followed all instructions of Intel's docs, 
and have tried on all versions of CSR from 1.1, 1.2.1, till 1.2.2., but the problem is the same.

has someone encountered the similar question before? Please help me.
Thanks a lot.

max 

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

stefan.eletzhofer | 1 Sep 2003 18:18
Picon

Re: boot from CF, lilo style?

On Sat, Aug 30, 2003 at 10:52:57AM -0400, Tom Walsh wrote:
> Hello,
> 
> I've looked at a number of ARM bootloaders and none seem to fit what I 
> would like a bootloader to do.  I would like to know if there is some 
> source that I can start with to build from?  I've been considering 
> ripping blob apart and using peices of it to build my own..
> 
> My system (currently) is the WebPal, armv3, arm710, 1 Meg Flash.  There 
> is an IDE controller on the board (Cirrus FDC37C665) and I've added a 
> custom board (mine) to the system where I have a Compact Flash attached 
> to the IDE of the Webpal.  The Webpal is merely a handy ARM system to do 
> this on, the bootloader would be incorporated into an arm720 system 
> later (probably a Sharp LH79520 based controller)
> 
> I can currently load a cramfs (kernel + busybox), then pivot_root to the 
> Compact Flash.  The problem is that I need to write, and add, more code 
> into my kernel sources and I don't have the room in the current cramfs 
> to hold the larger kernel.
> 
> Ideally, I would like the bootloader to come up, access the ext3 
> filesystem of the Compact Flash, pull in the kernel image and boot it. 
> I would rather that the kernel reside on the Compact Flash filesystem so 
> that I can easily update it.  The concept is to have field personell 
> download a new kernel into the Compact Flash to upgrade it, or simply 
> replace the CF itself (field people can be so dumb sometimes).

CVS blob can update a system using a CF card. There's IDE and CF support
in blob now. You can even use TFTP, if you have a NIC on your board.

(Continue reading)

stefan.eletzhofer | 2 Sep 2003 02:20
Picon

[PATCH] 2.4.21-rmk1 sa1111 usb ohci

Hi,
this is a patch to get 2.4.21-rmk1 SA1111 ohci usb working
again.

drivers/usb/hub.c refers to usb_dev dev->bus->bus_name, which
doesn't get initialized in case we have a sa1111, because we only
have "fake" PCI.

The new arch/arm/common/pcipool.c has a slot_name() function, but I
dont know how to get this to work with the fake struct pci_dev, thus
this hack.

diff -Nru a/arch/arm/mach-sa1100/Makefile b/arch/arm/mach-sa1100/Makefile
--- a/arch/arm/mach-sa1100/Makefile	Tue Sep  2 02:12:17 2003
+++ b/arch/arm/mach-sa1100/Makefile	Tue Sep  2 02:12:17 2003
 <at>  <at>  -18,7 +18,7  <at>  <at> 
 export-objs :=	assabet.o consus.o badge4.o dma-sa1100.o dma-sa1111.o \
 		flexanet.o freebird.o frodo.o generic.o h3600.o \
 		huw_webpanel.o irq.o sa1111.o sa1111-pcibuf.o \
-		system3.o yopy.o usb_ctl.o usb_recv.o usb_send.o simputer.o ssp.o
+		pcipool.o system3.o yopy.o usb_ctl.o usb_recv.o usb_send.o simputer.o ssp.o

 # These aren't present yet, and prevents a plain -ac kernel building.
 # hwtimer.o
 <at>  <at>  -40,7 +40,7  <at>  <at> 

 # Next, the SA1111 stuff.
 obj-$(CONFIG_SA1111) += sa1111.o dma-sa1111.o
-obj-$(CONFIG_USB_OHCI_SA1111) += sa1111-pcibuf.o
+obj-$(CONFIG_USB_OHCI_SA1111) += sa1111-pcibuf.o pcipool.o
(Continue reading)

Syed Mohammed Khasim | 2 Sep 2003 07:21

why MOD_INC_USE_COUNT in two places for MMC

Hi
 
I am using the mmc driver for omap_1510.
I am not understanding why the MOD_INC_USE_COUNT is done in two places
mmc_media_open and mmc_media_load.
 
When the driver enters mmc_media_load.
 
regards
Khasim.
-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

Syed Mohammed Khasim | 2 Sep 2003 07:18

MMC problem in mmc_media_open

Hi,
 
I am using the MMC driver for omap1510.
 
When I give mount -t vfat /dev/mmc /mnt/mmc
It enters the mmc_media_open function but it enters this if and returns and
unknown device /dev/mmc is displayed.
   
             dev = &g_media_dev[num]; 
              if ( !dev->slot )
 
The dev -> slot is updated in mmc_configure_card but the control never reaches till there.
 
Can any body suggest what may be the problem
 
with regards
Khasim
 
 
 
-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

Russell King - ARM Linux | 2 Sep 2003 09:56
Picon

Re: why MOD_INC_USE_COUNT in two places for MMC

On Tue, Sep 02, 2003 at 10:51:00AM +0530, Syed Mohammed Khasim wrote:
> I am using the mmc driver for omap_1510.
> I am not understanding why the MOD_INC_USE_COUNT is done in two places
> mmc_media_open and mmc_media_load.

Generally, if a module is using MOD_INC_USE_COUNT, it's broken.  You
shouldn't increase a module use count from zero when you're executing
code from that module.  The same is true of MOD_DEC_USE_COUNT
decrementing a modules use count to zero while executing module code.

IOW, try_module_get() (2.6) or try_inc_mod_count() (2.4) must be used
from the module caller.

Anyone not following this, well, good luck to them.  Getting this wrong
_and_ having preempt turned on is a recipee for an oops just waiting to
happen.

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

Syed Mohammed Khasim | 2 Sep 2003 10:58

Similar behaviour with MMC alternate driver by tony.

Hi,
 
I used the MMC alternate driver given by Tony.
The behaviour of this is same as that of our mmc driver.
The open call fails and the mount says unknown device /dev/mmc.
 
The driver is entering mmc open
 
    if (mmc_blk_devs[DEVICE_NR(inode->i_rdev)] == 0){
        return -ENODEV;
    }   
 
and returning due to this if.
 
Can any body suggest what is the problem.
 
regards
Khasim
-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php


Gmane