Hi Ilya,

Thanks for your reply.

Can you please help me understand what params you are referring to ?

I had changed my u-boot configs file to use: #define CONFIG_BOOTARGS "console=ttyS1,115200 root=/dev/mtdblock4 rw, mem=64M@0x00000000"

Note the change from default ttyS0 to ttyS1.

And linux bootargs I see xconf as : console=ttyS1,115200 root=/dev/mtdblock4 mem=64M@0x00000000

So, they look more or less ok ??

With same u-boot I can get 2.4 working. So, what could I be missing here.

Wolfgang: Apologise for html mail, my mistake. But I do see this issue as falling on the line between u-boot and linux. Not sure if linux guys will also push me to u-boot back again. So, please see if you can offer some advice.

Regards, Alfred

On 6/27/06, Ilya Yanok ilya.yanok@gmail.com wrote:

alfred hitch wrote:

...

I am trying to get 2.6 kernel boot on a plattfor very similar to ixdp425. We already have u-boot 1.1.1 and 2.4 kernel working for us. But, while trying to get the 2.6 working.

I get the Booting image at XXXX XXXX Starting Kernel ...

And after that nothing.

At this stage I am expecting the "Uncompressing Linux ......"

Looks like kernel doesn't use serial console, have you provided correct console= boot argument?

...

What could be going wrong ? Same board, if I put in 2.4 on a 1.1.4 version u-boot, it boots fine.

Is it some tty redirection issue ?? Only change from IXDP425 board is sdram configuration and ttyS1 for console instead of ttyS0. Suspecting later but then 2.4 shouldnt' have also booted up ?

I'm pretty sure that serial device of your board was renamed from ttyS to something more specific in 2.6 kernel. Check this.

Second possible reason is that you don't have boot arguments support in u-boot (your 2.4 kernel may have precompiled boot arguments, check kernel configuration).

    Ilya.

Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&da... _______________________________________________ U-Boot-Users mailing list U-Boot-Users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/u-boot-users

Dear Alfred!

alfred hitch schrieb:

Note the change from default ttyS0 to ttyS1.

Have you enabled PXA serial ports in your Linux configuration? And also the "Console on PXA serial port" option?

I have just taken a look into 2.6.15 and have found out that the PXA serial ports are also named ttyS, so it should be okay.

Regards Andreas Schweigstill

Hi Michael, Everyone,

Thanks for your replies. Michael: Looks like we do have something in common. This is a very serious observation. So, correct me if I a wrong. Did the same kernel give you traces with redboot but not u-boot. (We are together)

Did same kernel crash later on with u-boot and not redboot ? As in I am curious to know that did change in boot loader cause some other issue apart from crash also ?

Hey guys, looks there is more to it than just saying linux issue.

Is someone else out there running u-boot 1.1.4 successfully with 2.6 kernel ??

What could be going wrong ??

Cheers, Alfred

On 6/27/06, Michael Schwingen rincewind@discworld.dascon.de wrote:

alfred hitch wrote:

...

Hi All,

I am trying to get 2.6 kernel boot on a plattfor very similar to ixdp425. We already have u-boot 1.1.1 and 2.4 kernel working for us. But, while trying to get the 2.6 working.

I get the Booting image at XXXX XXXX Starting Kernel ...

And after that nothing.

At this stage I am expecting the "Uncompressing Linux ......"

What could be going wrong ? Same board, if I put in 2.4 on a 1.1.4 version u-boot, it boots fine.

Is it some tty redirection issue ?? Only change from IXDP425 board is sdram configuration and ttyS1 for console instead of ttyS0. Suspecting later but then 2.4 shouldnt' have also booted up ?

Maybe.

I just ran into similar problems. Since I had trouble getting a kernel to boot on my custom IXP425 board, I borrowed a IXDP425 board last weekend and tried that.

I have:

• IXDP425 board with redboot in flash
• a working kernel that can be booted by redboot via tftp.

Replacing redboot by u-boot (pulled from git last week), the kernel is started, but the early output (like "uncompressing kernel") does not appear on any of the UART ports.

When the kernel finally has the console set up, the messages appear - but if it panics before that point, there is no output at all.

Tracing into the kernel startup using gdb + bdi2000, I found no obvious reason as to why this happens, but I did not have enough time to really dig into the problem - and since I found the reason for my hang, this has now moved down in priority a bit. I can boorow the IXDP425 board again if absolutely necessary.

My guess would be that u-boot leaves the hardware in some different state than redboot, so that the early UART output in the linux kernel fails, but this is only a guess. However, since this happens only with u-boot and not with redboot, it does not seem to be a pure linux problem.

As a sidenote, the SDRAM config for the IXDP425 seems wrong to me: the board has 256MB SDRAM, but config.h shows only 16MB?

cu Michael

Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&da... _______________________________________________ U-Boot-Users mailing list U-Boot-Users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/u-boot-users

Hey guys, looks there is more to it than just saying linux issue. Is someone else out there running u-boot 1.1.4 successfully with 2.6 kernel ??

Running all 2.4 and 2.6 kernels since the release of 1.1.4 without problems.

Hi All,

I am still at logger heads with this issue.

Michael: What was ur typo ?? And just wondering if kernel default command line wasn't configured for u ?

Someone else also posted today on mailing list on kernel arg's not being passed ? Observations looks similar for now.

Cheers, Alfred

On 6/28/06, Michael Schwingen rincewind@discworld.dascon.de wrote:

alfred hitch wrote:

...

Hi Michael, Everyone,

Thanks for your replies. Michael: Looks like we do have something in common. This is a very serious observation. So, correct me if I a wrong. Did the same kernel give you traces with redboot but not u-boot. (We are together)

Correct. I could change between redboot and u-boot and observe this behaviour.

...

Did same kernel crash later on with u-boot and not redboot ? As in I am curious to know that did change in boot loader cause some other issue apart from crash also ?

No. Once I corrected the startup issues (which turned out to be a typo in the commandline arguments), the kernel started up fine (I did not test further than mounting the root fs).

cu Michael

Hi Alfred,

On Friday 30 June 2006 08:04, alfred hitch wrote:

I am sending dumps of my logs and what I did.. Anyone can give any pointers, I will really appreciate it.

Boot up log:

U-Boot 1.1.1.2 (Nov 17 2004 - 12:54:00)

Pretty old. A major update of the IXP4xx port was done a few weeks ago. Please try to port your board support to the current version.

sB_CONFIG: 32MB_2CHIP

U-Boot code: 00F80000 -> 00F90E64 BSS: -> 00F92374 RAM Configuration: Bank #0: 00000000 16 MB Flash: 32 MB *** Warning - bad CRC, using default environment

In: serial Out: serial Err: serial Hit any key to stop autoboot: 0 => setenv bootargs 'console=ttySAC01,115200 root=/dev/mtdblock4 rw,

Try "ttyS0" or "ttyS1" as console parameter.

<snip>

I created uImage using this command:

->> u-boot-mkimage -A arm -O linux -T kernel -C none -a 0x01600000 -e 0x01600000 -n "Linux" -d ./zImage uImage

Why didn't you just type "make uImage"? This result in a proper uImage to boot from U-Boot.

BTW: We are using current U-Boot and current Linux 2.6 kernel on different IXP4xx boards without any problems (e.g. ixdpg425, pdnb3).

Best regards, Stefan

Michael Schwingen wrote:

...

Did same kernel crash later on with u-boot and not redboot ? As in I am curious to know that did change in boot loader cause some other issue apart from crash also ?

No. Once I corrected the startup issues (which turned out to be a typo in the commandline arguments), the kernel started up fine (I did not test further than mounting the root fs).

cu Michael

Just for the record: my problems have disappeared now. It turns out the problem was caused by the LL_DEBUG code in Linux, which expects the UART to be enabled and initialized. Now, when u-boot and kernel disagree about which UART is to be used for console output, the kernel messages simply disappear.

This was complicated by the fact that the kernel LL_DEBUG code for IXP4xx always uses UART1, while my board only has UART2 connected, and I configured the IXDP425 board for a similar setup.

I modified my u-boot to always initialize both UARTS - that way, you can at least see the messages on boards that have both UARTS connected. As long as noone writes to the unused UART, this should not cause any harm.

The reason why I got no "normal" console output was a typo in the "mem=..." commandline parameter, which lead to errors early in memory management initialization in the kernel.

cu Michael