2013-03-16 19:36:06

by Martin Walch

[permalink] [raw]
Subject: RAMster selects HAVE_ALIGNED_STRUCT_PAGE on 64 bit systems when it should not

At drivers/staging/zcache/Kconfig:13 it says

config RAMSTER
bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem"
depends on CONFIGFS_FS=y && SYSFS=y && !HIGHMEM && ZCACHE=y
depends on NET
# must ensure struct page is 8-byte aligned
select HAVE_ALIGNED_STRUCT_PAGE if !64_BIT
default n

However, there is no config section for the symbol 64_BIT, so !64_BIT will
always be evaluated to true, and therefore RAMSTER will always select
HAVE_ALIGNED_STRUCT_PAGE, disregarding any breakage.

64_BIT should probably say 64BIT which has corresponding config sections in
arch/{x86,ia64,mips,s390,tile,alpha,arm64,sparc,parisc,powerpc}/Kconfig and
arch/x86/um/Kconfig.

I do not know if there are any bad consequences about this, but I guess this
should be reviewed and fixed, probably by just removing the underscore from
64_BIT in drivers/staging/zcache/Kconfig.

Regards
Martin Walch
--


Attachments:
signature.asc (198.00 B)
This is a digitally signed message part.

2013-03-16 20:47:27

by Paul Bolle

[permalink] [raw]
Subject: Re: RAMster selects HAVE_ALIGNED_STRUCT_PAGE on 64 bit systems when it should not

On Sat, 2013-03-16 at 20:35 +0100, Martin Walch wrote:
> 64_BIT should probably say 64BIT which has corresponding config sections in
> arch/{x86,ia64,mips,s390,tile,alpha,arm64,sparc,parisc,powerpc}/Kconfig and
> arch/x86/um/Kconfig.
>
> I do not know if there are any bad consequences about this, but I guess this
> should be reviewed and fixed, probably by just removing the underscore from
> 64_BIT in drivers/staging/zcache/Kconfig.

It's an obvious typo. See the patch I sent last week:
https://lkml.org/lkml/2013/3/9/106 . I have not yet received any
feedback.


Paul Bolle

2013-03-16 22:15:10

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: RAMster selects HAVE_ALIGNED_STRUCT_PAGE on 64 bit systems when it should not

On Sat, Mar 16, 2013 at 09:47:24PM +0100, Paul Bolle wrote:
> On Sat, 2013-03-16 at 20:35 +0100, Martin Walch wrote:
> > 64_BIT should probably say 64BIT which has corresponding config sections in
> > arch/{x86,ia64,mips,s390,tile,alpha,arm64,sparc,parisc,powerpc}/Kconfig and
> > arch/x86/um/Kconfig.
> >
> > I do not know if there are any bad consequences about this, but I guess this
> > should be reviewed and fixed, probably by just removing the underscore from
> > 64_BIT in drivers/staging/zcache/Kconfig.
>
> It's an obvious typo. See the patch I sent last week:
> https://lkml.org/lkml/2013/3/9/106 . I have not yet received any
> feedback.

Your patch is in my "to-apply" queue that I should get to soon. It's
only 137 patches back at the moment...

thanks,

greg "my queue is bigger than yours" k-h