Re: linux-next: Tree for April 21 (i82975x_edac)

[Arvind R <arvino55@gmail.com>]

On Thu, Apr 21, 2011 at 9:21 PM, Randy Dunlap <randy.dunlap@oracle.com> wrote:
> On Thu, 21 Apr 2011 13:20:45 +1000 Stephen Rothwell wrote:
>
>> Changes since 20110420:
>
>
> on i386:
>
> drivers/edac/i82975x_edac.c:298: warning: integer constant is too large for 'unsigned long' type
>
> which is:
>
>        if (info->xeap & 1)
>                page |= 0x100000000ul; <<<<<<<<<<<<<<<
>
> but page is unsigned long (32 bits), while 0x100000000ul is more than 32 bits,
> so that line isn't setting a bit in <page>.  Is it a bug, or this driver isn't
> applicable to X86_32, or what?
>
would defining page variable to unsigned long long solve it for x86_32?

That would still not fix the wrong error-info reported by the driver - this is
fixed by the patch-set titled '[RFC PATCH 2.6.37 0/4] i3000_edac : merge
i82975x support' dated 2011-02-14 in the linux-edac archives.

Please advise on getting the patch into main-stream. The I3000 is a re-labeled
i82975x (or vice-versa). Hence the patch-set to merge the drivers. If separate
i82975x driver is needed - I can provide the patch. The reported bug is not
present in the merge-patched driver.

Thanks,
Arvind.