Hi Wolfgang,

This is a customized board, which is very similar with Freescale's P1020RDB (which is p1_p2_rdb in u-boot), only we replace the process with a compatible one-core version. And the board support files are mostly copied from the p1_p2_rdb.

And the following is my updated patch. My git does not have "git send-mail" supported, so I have to paste the formatted commit below. Sorry for that.

From 9758bc34e8459c8b906029105e41fe8adf9e79ca Mon Sep 17 00:00:00 2001

From: Ke Pan pppeterpppan@gmail.com Date: Thu, 13 Oct 2011 10:34:23 +0800 Subject: [PATCH] common/cmd_elf : Disable interrupts before boot vxWorks

The vxWorks needs all interrupts to be disabled before its boot. --- common/cmd_elf.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/common/cmd_elf.c b/common/cmd_elf.c index bf32612..a5ef9b4 100644 --- a/common/cmd_elf.c +++ b/common/cmd_elf.c @@ -228,6 +228,8 @@ int do_bootvx (cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[])

printf ("## Using bootline (@ 0x%lx): %s\n", bootaddr, (char *) bootaddr); + printf("## Disableing interrupts ...\n"); + disable_interrupts(); printf ("## Starting vxWorks at 0x%08lx ...\n", addr);

((void (*)(void)) addr) ();

Hi Wolfgang,

I've installed the git-email packet now, but I can not make it work. The "Unable to initialize SMTP properly" always pop up. T_T This is the first time I send any patch to the community, sorry for the incontinence.

This is tested on our board, and I think at least for any Freescale E500v2 core based board, this patch is needed. The TCR[DIE] is set in function interrupt_init_cpu of mpc85xx/interrupts.c, which enables decrementer interrupts.

The following is my patch. I've changed the printf to a comment.

Subject: [PATCH] common/cmd_elf : Disable interrupts before bootup vxWorks

vxWorks needs all interrupts to be disabled before bootup. Signed-off-by: Ke Pan pppeterpppan@gmail.com --- common/cmd_elf.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/common/cmd_elf.c b/common/cmd_elf.c index bf32612..7a7a6bb 100644 --- a/common/cmd_elf.c +++ b/common/cmd_elf.c @@ -230,6 +230,8 @@ int do_bootvx (cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) (char *) bootaddr); printf ("## Starting vxWorks at 0x%08lx ...\n", addr);

+ /* Disable all interrupts before boot up vxWorks */ + disable_interrupts(); ((void (*)(void)) addr) ();

puts ("## vxWorks terminated\n");

Hallo,

i am out of office at the moment, so i can not test, but i have tracked all my vxWorks start problems down to cache issues, even if they look like irq issues at a first glance, but disabling the irq's may not harm.

Adding the boot parameter and a cache flush makes much more sense for me.

I will come with a formal patch later next week.

Kind regards

Reinhard

-----Original-Nachricht-----

Von: Wolfgang Denk wd@denx.de

An: Peter Pan pppeterpppan@gmail.com

Cc: u-boot@lists.denx.de

Betreff: Re: [U-Boot] Interrupt issue about bootvx command

Datum: Fri, 14 Oct 2011 22:46:28 +0200

Dear Peter Pan,

In message CAGQXtZEXL+psHAL2RgcimeQo_CQdTw13agSRDT5fQ4tmkmzAjQ@mail.gmail.com you wrote:

>

> I've installed the git-email packet now, but I can not make it work.

> The "Unable to initialize SMTP properly" always pop up. T_T

> This is the first time I send any patch to the community, sorry for

> the incontinence.

>

> This is tested on our board, and I think at least for any Freescale

> E500v2 core based board, this patch is needed. The TCR[DIE] is set in

> function interrupt_init_cpu of mpc85xx/interrupts.c, which enables

> decrementer interrupts.

>

> The following is my patch. I've changed the printf to a comment.

We've been through a number of iterations now. Your technical

problems aside, I don't understand why you cannot follow simple

instructions.

I pointed you at http://www.denx.de/wiki/U-Boot/Patches ; there you

can read:

* In the message body, include a description of your changes.

- For bug fixes: a description of the bug and how your

patch fixes this bug. Please try to include a way of

demonstrating that the patch actually fixes something.

- For new features: a description of the feature and your

implementation.

- Additional comments which you don't want included in

U-Boot's history can be included below the first "---" in

the message body.

Do you think above text should become part of the commit message?

Probably not. So why don't you treat it as "Additional comments which

you don't want included in U-Boot's history"?

>

> Subject: [PATCH] common/cmd_elf : Disable interrupts before bootup vxWorks

>

> vxWorks needs all interrupts to be disabled before bootup.

> Signed-off-by: Ke Pan pppeterpppan@gmail.com

> ---

Here is the place where such comments should go...

And when reposting a patch, you are supposed to follow the steps

described here:

http://www.denx.de/wiki/view/U-Boot/Patches#Sending_updated_patch_versions

> common/cmd_elf.c | 2 ++

> 1 files changed, 2 insertions(+), 0 deletions(-)

>

> diff --git a/common/cmd_elf.c b/common/cmd_elf.c

> index bf32612..7a7a6bb 100644

> --- a/common/cmd_elf.c

> +++ b/common/cmd_elf.c

> @@ -230,6 +230,8 @@ int do_bootvx (cmd_tbl_t *cmdtp, int flag, int

> argc, char * const argv[])

> (char *) bootaddr);

> printf ("## Starting vxWorks at 0x%08lx ...n", addr);

>

> + /* Disable all interrupts before boot up vxWorks */

> + disable_interrupts();

> ((void (*)(void)) addr) ();

>

> puts ("## vxWorks terminatedn");

Finally, as I wrote before, I would like to see some additional

feedback for other VxWorks users.

Any testers around?

Best regards,

Wolfgang Denk

--

DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel

HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de

Computers are not intelligent. They only think they are.

_______________________________________________

U-Boot mailing list

U-Boot@lists.denx.de

http://lists.denx.de/mailman/listinfo/u-boot