RTAS error logging of EEH errors

Nathan Fontenot nfont at austin.ibm.com
Tue Feb 3 03:01:44 EST 2004

I ended up leaving this on the stack because we call panic() right
after using the buffer.  The patch below changes this so that the
buffer is allocated.  If this is preferred let me know and I will
push the change to Ameslab.

I don't think this should be a statically defined buffer, no sense
in keeping around a buffer this big that will only be used once
right before we die.

-Nathan Fontenot

On Sun, 2004-02-01 at 01:36, Anton Blanchard wrote:
> Hi,
> I was reviewing patches to send onto Andrew Morton and noticed:
>                 if (ret == 0 && rets[1] == 1 && rets[0] >= 2) {
> +                       unsigned char slot_err_buf[RTAS_ERROR_LOG_MAX];
> +                       unsigned long   slot_err_ret;
> Thats 2kB on the stack, it really wants to be kmalloced (GFP_ATOMIC
> since it can be called in interrupt context) or statically defined.
> Anton
-------------- next part --------------
A non-text attachment was scrubbed...
Name: eeh_update.patch
Type: text/x-patch
Size: 1470 bytes
Desc: not available
Url : http://ozlabs.org/pipermail/linuxppc64-dev/attachments/20040202/e964cd4a/attachment.bin 

More information about the Linuxppc64-dev mailing list