Re: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1, 0)

Thanks for your reply Aras. Yes. I have selected I have compiled kernel with
fileystem support.

Can you suugest any other reson for such kernel behaviour.?

TIA
satpal

On 3/1/07, Aras Vaichas <arasv <at> magtech.com.au> wrote:
>
> satpal parmar wrote:
> ...
> > No filesystem could mount root, tried:  ext2
>
> Have you compiled in support for the filing system used to create the
> initrd?
> e.g. cramfs, squashfs?
>
> Also, I use a 9MB initrd image without any problems.
>
> regards,
>
> Aras Vaichas
>
> ______________________________________________________________________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email
> ______________________________________________________________________
>
-------------------------------------------------------------------

Re: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1, 0)

Hi all

         I am able to solve the problem.It was related to boot parameters.
as we can see from log oneline realted to
I was passing wrong parameters so its was not loooking for initram.Once I
passed right params we got the shell.

Warning: bad configuration page, trying to continue: this give a hint that
some thing is wrong with params or ATAG list.

>> this is not present in my log
 checking if image is initramfs...it isn't (bad gzip magic numbers); looks
like an initrd

>From these two observations we got the solution.

Anyway thanks a lot.
satpal

On 3/1/07, satpal parmar <parmarsatpal <at> gmail.com> wrote:
>
> Thanks for your reply Aras. Yes. I have selected I have compiled kernel
> with fileystem support.
>
> Can you suugest any other reson for such kernel behaviour.?
>
> TIA
> satpal
>
>

Reg maping and unmaping RAM buffer

Hi,
  I'm in the process of mapping and unmapping a RAM
buffer from the driver to user space. My kernel
version is 2.4.10. I've the following problem for
which I'm unable to find a proper solution

>From our driver, if we allocate memory using
get_dma_pages call and map it to the user space. How
to free this memory when the application unmaps it ? 

I have tried the following

Registered my own handler functions in the
vm_operations structure that is passed to the driver
when mmap is called. I'm seeing that the vm_close
funciton is called when the application munmaps the
mapped buffer. I tried the following code in the
vm_close implementation without any success.

void my_vm_close (struct vm_area_struct *vma)
{ 
	struct mmap_info    *mmapInfo;
	unsigned int        len;
	int                 order;
	struct dev_data     *devdata;
	char                *kbuf;
	unsigned int        count = 0;
	unsigned int        limit; 
	struct page         *page;

[PATCH] fix uncorrect defines of virtual addresses of NPE[ABC] registers

This patch fixes wrong defines of NPE[ABC] registers. Currently
they are defined as physical addresses, but they must be virtual...

Russel, is this patch ok? Should I post it via your patch system?

Signed-off-by: Milan Svoboda <msvoboda <at> ra.rockwell.com>
---

Index: linux/include/asm-arm/arch-ixp4xx/ixp4xx-regs.h
===================================================================
--- linux.orig/include/asm-arm/arch-ixp4xx/ixp4xx-regs.h
+++ linux/include/asm-arm/arch-ixp4xx/ixp4xx-regs.h
 <at>  <at>  -146,9 +146,9  <at>  <at> 
 #define IXP4XX_INTC_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0x3000)
 #define IXP4XX_GPIO_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0x4000)
 #define IXP4XX_TIMER_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0x5000)
-#define IXP4XX_NPEA_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_PHYS + 0x6000)
-#define IXP4XX_NPEB_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_PHYS + 0x7000)
-#define IXP4XX_NPEC_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_PHYS + 0x8000)
+#define IXP4XX_NPEA_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_VIRT + 0x6000)
+#define IXP4XX_NPEB_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_VIRT + 0x7000)
+#define IXP4XX_NPEC_BASE_VIRT   	(IXP4XX_PERIPHERAL_BASE_VIRT + 0x8000)
 #define IXP4XX_EthB_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0x9000)
 #define IXP4XX_EthC_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0xA000)
 #define IXP4XX_USB_BASE_VIRT		(IXP4XX_PERIPHERAL_BASE_VIRT + 0xB000)

-------------------------------------------------------------------

cardctl ident command does not identify the wireless cf card

Hello,
I am working with the board i.MXL Litekit based on i.MXL processor (Cogent 
CSB536) from Freescale and I need to connect to him a Netgear MA701 WIFI 
compact flash card. I am using the kernel 2.6.11.4. 

When I insert the WiFi card.. I get: 
 # cardctl status
Socket 0: 3.3V 16-bit PC Card
  function 0: [ready], [bat dead]
# cardctl config
Socket 0:
  Vcc 3.3V  Vpp1 3.3V  Vpp2 3.3V
# cardctl ident
Socket 0:
  no product info available
/ # cardctl info
PRODID_1=""
PRODID_2=""
PRODID_3=""
PRODID_4=""
MANFID=0000,0000
FUNCID=255

I load properly pcmcia_core, pcmcia, imxl_csb536 (custom driver), orinoco 
and orinoco_cs, with my wifi card in. 
# cat /etc/sysconfig/pcmcia
PCMCIA=yes
PCIC=imxl_csb536
PCIC_OPTS=
CORE_OPTS=

Re: Lost serial console after resume

>From http://lists.openwall.net/linux-kernel/2006/08/29/342
This MACRO is used to help debugging suspend/resume. But it makes serial sonsole
out of work on PXA after resume. How doesn it help resume debug on
PXA? Does this
MACRO depends on some other one to enable the serial console after resume?

Regards
Yin, Fengwei

On 3/1/07, Rodolfo Giometti <giometti <at> enneenne.com> wrote:
> On Wed, Feb 28, 2007 at 06:09:22PM +0000, Russell King - ARM Linux wrote:
>
> > It's fairly simple.  If you disable console suspend, the console is taken
> > out of the suspend/resume path and _will_ _not_ be suspended and resumed.
> >
> > So, if you want the console to be suspend/resumed, don't enable the option.
>
> I see. But the problem is that I'm working on suspend/resume on UDC
> and keyboard devices so I _need_ that support. :)
>
> Are there no ways to not disable the serial console but saving serial
> status anyway?
>
> Thanks,
>
> Rodolfo
>
> --
>
> GNU/Linux Solutions                  e-mail:    giometti <at> enneenne.com

Re: [RFC, PATCH 2.6.20] MXC platform and i.MX31ADS core support

Damien,

> In its current state the iMX31 linux port is able to address 128 MB.
> This is definitely not enough !
>
> You'll have :
> - to make VMALLOC_END a little higher to get some space
> for ioremap() and vmalloc() even if the board has 512 MB of RAM.
> (may I suggest
> #define VMALLOC_END       (PAGE_OFFSET + 0x30000000)
> ?)
> - to change your static virtual addresses away from the newly
> defined vmalloc space in include/asm-arm/arch-mxc/mx31.h
> - make sure compiling with DISCONTIGMEM works
>
> I've this changes in my local tree (as we have 128M <at> 0x80000000
> and  128M <at> 0x90000000) and  the MMU translation is  correctly
> set up,  so says my JTAG debugger)

Can you provide a patch of what you've done so far?  Or, at least
what you've done to meet the requirements that you've outlined
above?

> Careful,  this changes are breaking  your binary only OpenGL driver
> as is uses mem_map (through mm subsystem defines & functions)
>
> I do have those sources under NDA, but a simple compilation is not
> enough to make it work again, I've already solved a bug, but still the
> OpenGL driver does not free its memory properly with DISCONTIGMEM.
> This is my current ongoing work, pretty hairy for a starter in the mm

Any possible to split the pxa-regs.h?

Hi,
We are trying to clean the code for PXA3XX. And we find that there are two
many #ifdef #endif in pxa-regs.h after add the PXA3XX registers definition.
The code is unclear and error-prone. I just wonder whether it's acceptable to
split the pxa-regs.h to pxa2xx-regs.h and pxa3xx-regs.h?

Regards
Yin, Fengwei

-------------------------------------------------------------------
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

Re: Reg maping and unmaping RAM buffer

On Thu, Mar 01, 2007 at 06:49:04AM -0800, shankoo77 <at> yahoo.com wrote:
>   I'm in the process of mapping and unmapping a RAM
> buffer from the driver to user space. My kernel
> version is 2.4.10. I've the following problem for
> which I'm unable to find a proper solution

Please read http://www.arm.linux.org.uk/mailinglists/faq.php#f11 .
2.4 knowledge is getting very stale...

Erik

--

-- 
----  Erik Mouw  ----  www.bitwizard.nl  ----  +31 15 2600 998  ----

-------------------------------------------------------------------
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

Reg unmmap-ing RAM buffer to user space

Hi,
  I'm in the process of mapping and unmapping a RAM
buffer from the driver to user space. My kernel
version is 2.4.10. I've the following problem for
which I'm unable to find a proper solution

>From our driver, if we allocate memory using
get_dma_pages call and map it to the user space. How
to free this memory when the application unmaps it ? 

I have tried the following

Registered my own handler functions in the
vm_operations structure that is passed to the driver
when mmap is called. I'm seeing that the vm_close
funciton is called when the application munmaps the
mapped buffer. I tried the following code in the
vm_close implementation without any success.

void my_vm_close (struct vm_area_struct *vma)
{ 
	struct mmap_info    *mmapInfo;
	unsigned int        len;
	int                 order;
	struct dev_data     *devdata;
	char                *kbuf;
	unsigned int        count = 0;
	unsigned int        limit; 
	struct page         *page;