Re: -tip: origin tree boot crash

From: Len Brown
Date: Wed Dec 30 2009 - 01:15:30 EST


Thanks for the quick test/debug Paul,

> Fixed this completely with the following two patches on top of 2.6.33-rc2 :
>
> --- wmi.c.orig 2009-12-28 17:27:15.000000000 +0100
> +++ wmi.c 2009-12-28 17:39:01.000000000 +0100
> @@ -488,12 +488,13 @@
> {
> struct wmi_block *block;
> acpi_status status;
> + bool err;
>
> if (!guid || !handler)
> return AE_BAD_PARAMETER;
>
> - find_guid(guid, &block);
> - if (!block)
> + err = find_guid(guid, &block);
> + if (!err)

Unfortunately, find_guid() returns 1 for success
and 0 for failure, making this code look backwards.

I've applied this, but changed it to if (!find_guid())
to be consistent with the other callers.

> return AE_NOT_EXIST;
>
> if (block->handler)
> @@ -517,12 +518,13 @@
> {
> struct wmi_block *block;
> acpi_status status;
> + bool err;
>
> if (!guid)
> return AE_BAD_PARAMETER;
>
> - find_guid(guid, &block);
> - if (!block)
> + err = find_guid(guid, &block);
> + if (!err)
> return AE_NOT_EXIST;
>
> if (!block->handler)
>
> --- dell-wmi.c.orig 2009-12-28 17:49:09.000000000 +0100
> +++ dell-wmi.c 2009-12-28 17:41:36.000000000 +0100
> @@ -343,7 +343,7 @@
> printk(KERN_ERR
> "dell-wmi: Unable to register notify handler - %d\n",
> err);
> - return err;
> + return -err;

This change is not quite correct,
but the correct fix is already in the acpi-test tree in the patch
"dell-wmi: sys_init_module: 'dell_wmi'->init suspiciously returned 21,
it should follow 0/-E convention"

I'll reply with both of the patches in a sec.
Please let me know if they are sufficient to fix your system.

thanks,
-Len Brown, Intel Open Source Technology Center

> }
>
> return 0;
>
>
> Signed-off-by: rol@xxxxxxxxxx <Paul Rolland>
>
> Paul
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/