linux-next: md tree build failure

linux-next: md tree build failure

[Stephen Rothwell]

Hi Neil,

Today's linux-next build (powerpc ppc64_defconfig) failed like this:

drivers/md/raid0.c: In function 'create_strip_zones':
drivers/md/raid0.c:243: error: 'struct request_queue' has no member named 'hardsect_size'

Caused by commit e1defc4ff0cf57aca6c5e3ff99fa503f5943c1f1 ("block: Do
away with the notion of hardsect_size") from the block tree interacting
with commit 131e4477401ae76a9cbe8539e4b7819e04cfba36 ("md: raid0: chunk
size check in raid0_run") from the md tree.

I applied the following patch.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index b8a0747..0bf275d 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -240,7 +240,7 @@ static int create_strip_zones(mddev_t *mddev)
 	 * now since we have the hard sector sizes, we can make sure
 	 * chunk size is a multiple of that sector size
 	 */
-	if (mddev->chunk_size % mddev->queue->hardsect_size) {
+	if (mddev->chunk_size % queue_logical_block_size(mddev->queue)) {
 		printk(KERN_ERR "%s chunk_size of %d not valid\n",
 		       mdname(mddev),
 		       mddev->chunk_size);

Re: linux-next: md tree build failure

[Jens Axboe]

On Mon, May 25 2009, Stephen Rothwell wrote:
> Hi Neil,
> 
> Today's linux-next build (powerpc ppc64_defconfig) failed like this:
> 
> drivers/md/raid0.c: In function 'create_strip_zones':
> drivers/md/raid0.c:243: error: 'struct request_queue' has no member named 'hardsect_size'
> 
> Caused by commit e1defc4ff0cf57aca6c5e3ff99fa503f5943c1f1 ("block: Do
> away with the notion of hardsect_size") from the block tree interacting
> with commit 131e4477401ae76a9cbe8539e4b7819e04cfba36 ("md: raid0: chunk
> size check in raid0_run") from the md tree.

Martin, this is starting to look like a disaster :-(

> 
> I applied the following patch.
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> 
> diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
> index b8a0747..0bf275d 100644
> --- a/drivers/md/raid0.c
> +++ b/drivers/md/raid0.c
> @@ -240,7 +240,7 @@ static int create_strip_zones(mddev_t *mddev)
>  	 * now since we have the hard sector sizes, we can make sure
>  	 * chunk size is a multiple of that sector size
>  	 */
> -	if (mddev->chunk_size % mddev->queue->hardsect_size) {
> +	if (mddev->chunk_size % queue_logical_block_size(mddev->queue)) {
>  		printk(KERN_ERR "%s chunk_size of %d not valid\n",
>  		       mdname(mddev),
>  		       mddev->chunk_size);

-- 
Jens Axboe

Re: linux-next: md tree build failure

[Martin K. Petersen]

>>>>> "Jens" == Jens Axboe <jens.axboe@oracle.com> writes:

Jens> Martin, this is starting to look like a disaster :-(

Yeah.  I guess you should have pushed the accessors first and let that
patch steep for a while before applying the rest of the kit :|

It's just super bad timing that both Neil and Alasdair had patches to
their respective hardsect size areas right now.  Neither tree has been
touched in that department for ages...

*sigh*

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: linux-next: md tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 913 bytes --]

On Mon, 25 May 2009 08:15:52 +0200 Jens Axboe <jens.axboe@oracle.com> wrote:
>
> On Mon, May 25 2009, Stephen Rothwell wrote:
> > 
> > Today's linux-next build (powerpc ppc64_defconfig) failed like this:
> > 
> > drivers/md/raid0.c: In function 'create_strip_zones':
> > drivers/md/raid0.c:243: error: 'struct request_queue' has no member named 'hardsect_size'
> > 
> > Caused by commit e1defc4ff0cf57aca6c5e3ff99fa503f5943c1f1 ("block: Do
> > away with the notion of hardsect_size") from the block tree interacting
> > with commit 131e4477401ae76a9cbe8539e4b7819e04cfba36 ("md: raid0: chunk
> > size check in raid0_run") from the md tree.
> 
> Martin, this is starting to look like a disaster :-(

So, is anything happening about these?  I still get these (and the dm)
build failures.

-- 
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: md tree build failure

[Martin K. Petersen]

>>>>> "Stephen" == Stephen Rothwell <sfr@canb.auug.org.au> writes:

>> > Caused by commit e1defc4ff0cf57aca6c5e3ff99fa503f5943c1f1 ("block:
>> > Do away with the notion of hardsect_size") from the block tree
>> > interacting with commit 131e4477401ae76a9cbe8539e4b7819e04cfba36
>> > ("md: raid0: chunk size check in raid0_run") from the md tree.

Stephen> So, is anything happening about these?  I still get these (and
Stephen> the dm) build failures.

There's a replacement DM patch kit brewing that will hopefully come your
way shortly.  It replaces the patches currently in Alasdair's tree.
Mike sent the kit out for review last week and I'm currently testing it.


I don't know what Neil's plans are.  Neil: Do you want to rebase your
tree, fix up the conflicting patches or how would you like to handle
this?

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: linux-next: md tree build failure

[NeilBrown]

On Mon, June 1, 2009 3:13 pm, Martin K. Petersen wrote:
>>>>>> "Stephen" == Stephen Rothwell <sfr@canb.auug.org.au> writes:
>
>>> > Caused by commit e1defc4ff0cf57aca6c5e3ff99fa503f5943c1f1 ("block:
>>> > Do away with the notion of hardsect_size") from the block tree
>>> > interacting with commit 131e4477401ae76a9cbe8539e4b7819e04cfba36
>>> > ("md: raid0: chunk size check in raid0_run") from the md tree.
>
> Stephen> So, is anything happening about these?  I still get these (and
> Stephen> the dm) build failures.
>
> There's a replacement DM patch kit brewing that will hopefully come your
> way shortly.  It replaces the patches currently in Alasdair's tree.
> Mike sent the kit out for review last week and I'm currently testing it.
>
>
> I don't know what Neil's plans are.  Neil: Do you want to rebase your
> tree, fix up the conflicting patches or how would you like to handle
> this?

I'm happy to rebase on something if you have a tree that doesn't have too
much more than the changes that are causing conflicts.
Then I'll fix my patches so they work on that tree.

What should I pull??

NeilBrown