Re: Please include pagemap header cleanup in -next

Re: Please include pagemap header cleanup in -next

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1154 bytes --]

On Mon, Aug 24, 2026 at 06:00:47PM +0100, Matthew Wilcox wrote:
> Hi Mark,
> 
> As discussed on IRC, please include
> 
> git://git.infradead.org/users/willy/pagecache.git headers
> 
> The net effect is to reduce the number of files depending on pagemap.h
> from ~9100 to ~1400 on an x86 allmodconfig build.  The Intel test robot
> has been steadily building random configs on random architectures and
> finding new missing includes, so there may be some build breakages that
> it hasn't tested.
> 
> My plan is to send this to Linus on Thursday or Friday.  I don't see the
> point in letting it soak longer; any problems it causes are build
> problems.

As I noted in the regular build error mail it hit a build fail on an arm
multi_v7_defconfig so I didn't merge this today.  It's pretty surprising
that the build bot didn't hit that one TBH since the file that broke
hasn't changed since v7.2 AFAICT, I would guess that there's some other
implicit include cleanup work that's interacting badly with this effort
but I haven't investigated at all.

BTW I added the -next list, please CC requests for -next there for
visibility.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: Please include pagemap header cleanup in -next

[Matthew Wilcox]

On Tue, Aug 25, 2026 at 02:44:54PM +0100, Mark Brown wrote:
> On Mon, Aug 24, 2026 at 06:00:47PM +0100, Matthew Wilcox wrote:
> > Hi Mark,
> > 
> > As discussed on IRC, please include
> > 
> > git://git.infradead.org/users/willy/pagecache.git headers
> > 
> > The net effect is to reduce the number of files depending on pagemap.h
> > from ~9100 to ~1400 on an x86 allmodconfig build.  The Intel test robot
> > has been steadily building random configs on random architectures and
> > finding new missing includes, so there may be some build breakages that
> > it hasn't tested.
> > 
> > My plan is to send this to Linus on Thursday or Friday.  I don't see the
> > point in letting it soak longer; any problems it causes are build
> > problems.
> 
> As I noted in the regular build error mail it hit a build fail on an arm
> multi_v7_defconfig so I didn't merge this today.  It's pretty surprising
> that the build bot didn't hit that one TBH since the file that broke
> hasn't changed since v7.2 AFAICT, I would guess that there's some other
> implicit include cleanup work that's interacting badly with this effort
> but I haven't investigated at all.
> 
> BTW I added the -next list, please CC requests for -next there for
> visibility.

Thanks; I looked for it in the Documentation and didn't find it.

I fixed up the one you mentioned   Then I installed an arm32
crosscompiler and built arm allmodconfig.  And then Intel's test robot
found another.  So in total, three more fixes:

drivers/dma/ste_dma40.c
drivers/ufs/host/ufs-qcom.c
drivers/gpu/drm/amd/amdkfd/kfd_migrate.c

Pushed out a new version (e02cb91d4644)

Re: Please include pagemap header cleanup in -next

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 814 bytes --]

On Tue, Aug 25, 2026 at 08:57:48PM +0100, Matthew Wilcox wrote:
> On Tue, Aug 25, 2026 at 02:44:54PM +0100, Mark Brown wrote:

> > As I noted in the regular build error mail it hit a build fail on an arm
> > multi_v7_defconfig so I didn't merge this today.  It's pretty surprising
> > that the build bot didn't hit that one TBH since the file that broke
> > hasn't changed since v7.2 AFAICT, I would guess that there's some other
> > implicit include cleanup work that's interacting badly with this effort
> > but I haven't investigated at all.

> I fixed up the one you mentioned   Then I installed an arm32
> crosscompiler and built arm allmodconfig.  And then Intel's test robot
> found another.  So in total, three more fixes:

Looks like that did the trick for the build tests I do as part of -next
at least.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]