linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2282 bytes --]

Hi Andrew,

On Thu, 15 Jan 2009 09:35:14 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Thu, 15 Jan 2009 20:05:08 +0900 (JST) KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> 
> > Applied after: linux-next.patch
> > ==
> > Subject: [PATCH] remove duplicate EXPORT_SYMBOL()
> > Impact: fix build error
> > 
> > In past, ia64_ksyms.c had EXPORT_SYMBOL(copy_page) in wrong place.
> > 
> > old code was
> > 
> > 	#include <asm/uaccess.h>
> > 	EXPORT_SYMBOL(__copy_user);
> > 	EXPORT_SYMBOL(__do_clear_user);
> > 	EXPORT_SYMBOL(__strlen_user);
> > 	EXPORT_SYMBOL(__strncpy_from_user);
> > 	EXPORT_SYMBOL(__strnlen_user);
> > 	EXPORT_SYMBOL(copy_page);
> > 
> > (but actually, copy_page is declared in asm/page.h)
> > 
> > And recently, commit bdc2619ab95d45d434c16d5c216bc6243761f6fb insert
> > EXPORT_SYMBOL(copy_page) into correct place.
> > 
> > then, duplicate EXPORT_SYMBOL create following build error.
> > 
> > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__kcrctab_copy_page'
> > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__kcrctab_copy_page' was here
> > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__kstrtab_copy_page'
> > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__kstrtab_copy_page' was here
> > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__ksymtab_copy_page'
> > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__ksymtab_copy_page' was here
> 
> Yes, this is a strange bug added by
> 
> commit 7aed50e09223e7623c7ab826efd53f097bed2f73
> Author:     Prarit Bhargava <prarit@redhat.com>
> AuthorDate: Sat Jan 10 02:25:44 2009 +0000
> Commit:     David Howells <dhowells@redhat.com>
> CommitDate: Sat Jan 10 02:25:44 2009 +0000
> 
>     CacheFiles: Add missing copy_page export for ia64
> 
> I don't see how it could have been tested :(
> 
> I also don't know which linux-next tree added that commit.  There's a
> way of working this out but I forgot it.

In gitk you can search backwards for a merge commit containing "sfr" :-)

I have dropped the nfs tree from linux-next for today.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Andrew Morton]

On Fri, 16 Jan 2009 12:49:49 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:

> Hi Andrew,
> 
> On Thu, 15 Jan 2009 09:35:14 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
> >
> > On Thu, 15 Jan 2009 20:05:08 +0900 (JST) KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> > 
> > > Applied after: linux-next.patch
> > > ==
> > > Subject: [PATCH] remove duplicate EXPORT_SYMBOL()
> > > Impact: fix build error
> > > 
> > > In past, ia64_ksyms.c had EXPORT_SYMBOL(copy_page) in wrong place.
> > > 
> > > old code was
> > > 
> > > 	#include <asm/uaccess.h>
> > > 	EXPORT_SYMBOL(__copy_user);
> > > 	EXPORT_SYMBOL(__do_clear_user);
> > > 	EXPORT_SYMBOL(__strlen_user);
> > > 	EXPORT_SYMBOL(__strncpy_from_user);
> > > 	EXPORT_SYMBOL(__strnlen_user);
> > > 	EXPORT_SYMBOL(copy_page);
> > > 
> > > (but actually, copy_page is declared in asm/page.h)
> > > 
> > > And recently, commit bdc2619ab95d45d434c16d5c216bc6243761f6fb insert
> > > EXPORT_SYMBOL(copy_page) into correct place.
> > > 
> > > then, duplicate EXPORT_SYMBOL create following build error.
> > > 
> > > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__kcrctab_copy_page'
> > > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__kcrctab_copy_page' was here
> > > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__kstrtab_copy_page'
> > > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__kstrtab_copy_page' was here
> > > arch/ia64/kernel/ia64_ksyms.c:65: error: redefinition of '__ksymtab_copy_page'
> > > arch/ia64/kernel/ia64_ksyms.c:43: error: previous definition of '__ksymtab_copy_page' was here
> > 
> > Yes, this is a strange bug added by
> > 
> > commit 7aed50e09223e7623c7ab826efd53f097bed2f73
> > Author:     Prarit Bhargava <prarit@redhat.com>
> > AuthorDate: Sat Jan 10 02:25:44 2009 +0000
> > Commit:     David Howells <dhowells@redhat.com>
> > CommitDate: Sat Jan 10 02:25:44 2009 +0000
> > 
> >     CacheFiles: Add missing copy_page export for ia64
> > 
> > I don't see how it could have been tested :(
> > 
> > I also don't know which linux-next tree added that commit.  There's a
> > way of working this out but I forgot it.
> 
> In gitk you can search backwards for a merge commit containing "sfr" :-)

I do most of my gitting across a DSL connection.  It makes GUIs
painful.

> I have dropped the nfs tree from linux-next for today.

Just for this buglet?  That sounds a bit savage.

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 777 bytes --]

Hi Andrew,

On Thu, 15 Jan 2009 17:59:17 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
>
> I do most of my gitting across a DSL connection.  It makes GUIs
> painful.

Yeah, I will think on it ...

> > I have dropped the nfs tree from linux-next for today.
> 
> Just for this buglet?  That sounds a bit savage.

There was another problem with the fscache code as well.  Also, this was
reported the day before yesterday and got no response.  Also, the other
patch that added EXPORT_SYMBOL(copy_page) "recently" was from
December, *2007* ... Also, this is 2.6.30 material and they have a long
time to fix it.

Also, it is much easier on me :-)
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[KOSAKI Motohiro]

> Hi Andrew,
> 
> On Thu, 15 Jan 2009 17:59:17 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
> >
> > I do most of my gitting across a DSL connection.  It makes GUIs
> > painful.
> 
> Yeah, I will think on it ...
> 
> > > I have dropped the nfs tree from linux-next for today.
> > 
> > Just for this buglet?  That sounds a bit savage.
> 
> There was another problem with the fscache code as well.  Also, this was
> reported the day before yesterday and got no response.  Also, the other
> patch that added EXPORT_SYMBOL(copy_page) "recently" was from
> December, *2007* ... Also, this is 2.6.30 material and they have a long
> time to fix it.
> 
> Also, it is much easier on me :-)

Ghh, sorry for my really poor analysis ;)
thanks stephen.

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

Hi Andrew,

On Fri, 16 Jan 2009 13:49:30 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> On Thu, 15 Jan 2009 17:59:17 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
> >
> > I do most of my gitting across a DSL connection.  It makes GUIs
> > painful.
> 
> Yeah, I will think on it ...

With the help of Paul Mackerras, I have come up with the shell script
below.  You run it with the commit id, the tag of the tree you care
about (e.g. next-20080115)(it defaults to the current top of tree) and
the name of the branch the represents Linus' tree (default to
origin/master).  It will spit out the (or a) tree that contains the
commit you asked about.

-------------------------------------------------------------------------
#!/bin/bash

commit=$1
[ "$commit" ] || {
	echo "$0: <commit id> [<top tag> [<linus branch>]]" 1>&2
	exit 1
}

top=$2
[ "$top" ] || top=HEAD

linus=$3
[ "$linus" ] || linus=origin/master

base=$(git merge-base "$linus" "$top")

git log --first-parent --pretty='format:%H %P' "$base".."$top" |
while read m p1 p2
do
	[ "$p2" ] || continue
	git rev-list $base..$p2 | grep -q "$commit" || continue
	branch=$(git show $m | sed -n "s/[ 	]*Merge .* '\([^']*\)'/\1/p")
	tree=${branch%%/*}
	[ "$tree" = "quilt" ] && tree=${branch##*/}
	echo "$tree"
	break
done

exit 0
----------------------------------------------------------------------------
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Andrew Morton]

On Fri, 16 Jan 2009 15:34:17 +1100
Stephen Rothwell <sfr@canb.auug.org.au> wrote:

> With the help of Paul Mackerras, I have come up with the shell script
> below.  You run it with the commit id, the tag of the tree you care
> about (e.g. next-20080115)(it defaults to the current top of tree) and
> the name of the branch the represents Linus' tree (default to
> origin/master).  It will spit out the (or a) tree that contains the
> commit you asked about.

hm.  Thanks.

y:/usr/src/git26> git-checkout linux-next
Switched to branch "linux-next"
y:/usr/src/git26> who-committed.sh ba9a5918
y:/usr/src/git26> who-committed.sh 2c3fb6b1b1a43ca089032e6523207f6687c962c1
y:/usr/src/git26> 

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 896 bytes --]

Hi Andrew,

On Fri, 16 Jan 2009 14:15:57 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
>
> y:/usr/src/git26> git-checkout linux-next
> Switched to branch "linux-next"
> y:/usr/src/git26> who-committed.sh ba9a5918

Well that commit is from 2.6.23-rc1:

commit ba9a5918c86762fa7fae353c635bb6ed5221dfb8
Author: Douglas Thompson <dougthompson@xmission.com>
Date:   Thu Jul 19 01:50:32 2007 -0700

    drivers/edac: add to maintainers new info

before linux-next even existed.  My script only works for things in
linux-next but not in Linus' tree ...

> y:/usr/src/git26> who-committed.sh 2c3fb6b1b1a43ca089032e6523207f6687c962c1

And that is one I committed myself because Linus' tree is broken.

I'll see if I can do better.  What do you actually want to do?
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 492 bytes --]

Hi Andrew,

On Sat, 17 Jan 2009 09:40:08 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> > y:/usr/src/git26> who-committed.sh 2c3fb6b1b1a43ca089032e6523207f6687c962c1
> 
> And that is one I committed myself because Linus' tree is broken.

It also appears as commit d992c8403a72c4067587a32a7736bcc2af6c408f in the
mfd tree but has not made it into Linus' tree yet.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Andrew Morton]

On Sat, 17 Jan 2009 09:40:08 +1100
Stephen Rothwell <sfr@canb.auug.org.au> wrote:

> Hi Andrew,
> 
> On Fri, 16 Jan 2009 14:15:57 -0800 Andrew Morton <akpm@linux-foundation.org> wrote:
> >
> > y:/usr/src/git26> git-checkout linux-next
> > Switched to branch "linux-next"
> > y:/usr/src/git26> who-committed.sh ba9a5918
> 
> Well that commit is from 2.6.23-rc1:
> 
> commit ba9a5918c86762fa7fae353c635bb6ed5221dfb8
> Author: Douglas Thompson <dougthompson@xmission.com>
> Date:   Thu Jul 19 01:50:32 2007 -0700
> 
>     drivers/edac: add to maintainers new info
> 
> before linux-next even existed.  My script only works for things in
> linux-next but not in Linus' tree ...
> 
> > y:/usr/src/git26> who-committed.sh 2c3fb6b1b1a43ca089032e6523207f6687c962c1
> 
> And that is one I committed myself because Linus' tree is broken.

Oh, OK, sorry, drat.  Two hastily and poorly chosen test cases.

> I'll see if I can do better.  What do you actually want to do?

y:/usr/src/git26> who-committed.sh bcff57c25287216221d0ff82080ac37bb081969e
ide

That ;)

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[David Howells]

Stephen Rothwell <sfr@canb.auug.org.au> wrote:

> > > I have dropped the nfs tree from linux-next for today.
> > 
> > Just for this buglet?  That sounds a bit savage.
> 
> There was another problem with the fscache code as well.  Also, this was
> reported the day before yesterday and got no response.

I've been flying to Tasmania for LCA and spending much time sitting around in
too many different airports.

> Also, the other patch that added EXPORT_SYMBOL(copy_page) "recently" was
> from December, *2007* ... Also, this is 2.6.30 material and they have a long
> time to fix it.

The patch was required Tuesday morning for my patches to build on top of
Linus's vanilla kernel.  This may have changed, but I haven't checked - too
busy enjoying my pre-LCA hols in Tas:-)

David

Re: linux-next: build problem with nfs tree (Was: Re: [mmotm][PATCH] fix ia64 build error)

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 568 bytes --]

Hi David,

On Sat, 17 Jan 2009 04:43:52 +0000 David Howells <dhowells@redhat.com> wrote:
>
> Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> 
> > There was another problem with the fscache code as well.  Also, this was
> > reported the day before yesterday and got no response.
> 
> I've been flying to Tasmania for LCA and spending much time sitting around in
> too many different airports.

Sorry about that ... it should have occurred to me.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]