Re: [PATCH 4.19 00/98] 4.19.321-rc1 review

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Mon, Sep 02, 2024 at 02:44:44PM +0530, Harshit Mogalapalli wrote:
> Hi Greg,
> 
> On 01/09/24 21:45, Greg Kroah-Hartman wrote:
> > This is the start of the stable review cycle for the 4.19.321 release.
> > There are 98 patches in this series, all will be posted as a response
> > to this one.  If anyone has any issues with these being applied, please
> > let me know.
> > 
> > Responses should be made by Tue, 03 Sep 2024 16:07:34 +0000.
> > Anything received after that time might be too late.
> 
> Build fails on our infrastructure.
> 
> 
> BUILDSTDERR: In file included from /builddir/build/BUILD/kernel-4.19.321/linux-4.19.321-master.20240901.el7.dev/tools/include/linux/bitmap.h:6,
> BUILDSTDERR:                  from /builddir/build/BUILD/kernel-4.19.321/linux-4.19.321-master.20240901.el7.dev/tools/perf/util/include/../../util/pmu.h:5,
> BUILDSTDERR:                  from arch/x86/util/pmu.c:9:
> BUILDSTDERR: /builddir/build/BUILD/kernel-4.19.321/linux-4.19.321-master.20240901.el7.dev/tools/include/linux/align.h:6:10:
> fatal error: uapi/linux/const.h: No such file or directory
> BUILDSTDERR:  #include <uapi/linux/const.h>
> BUILDSTDERR:           ^~~~~~~~~~~~~~~~~~~~
> BUILDSTDERR: compilation terminated.
> 
> 
> Looked at the commits:
> 
> This commit 993a20bf6225c: ("tools: move alignment-related macros to new
> <linux/align.h>") is causing that perf build to fail.
> 
> Solution is not to drop this patch as this is probably pulled in to support
> bitmap_size() macros in these commits(which are also part of this release):
> 
> 6fbe5a3920f48 fix bitmap corruption on close_range() with
> CLOSE_RANGE_UNSHARE
> ef9ebc42c10f8 bitmap: introduce generic optimized bitmap_size()
> 
> 
> 
> Applying the below diff, helps the perf build to pass: I think we should
> fold this into: commit 993a20bf6225c: ("tools: move alignment-related macros
> to new <linux/align.h>")
> 
> diff --git a/tools/include/linux/align.h b/tools/include/linux/align.h
> index 14e34ace80dda..a27bc1edf6e5c 100644
> --- a/tools/include/linux/align.h
> +++ b/tools/include/linux/align.h
> @@ -3,7 +3,7 @@
>  #ifndef _TOOLS_LINUX_ALIGN_H
>  #define _TOOLS_LINUX_ALIGN_H
> 
> -#include <uapi/linux/const.h>
> +#include <linux/const.h>
> 
>  #define ALIGN(x, a)            __ALIGN_KERNEL((x), (a))
>  #define ALIGN_DOWN(x, a)       __ALIGN_KERNEL((x) - ((a) - 1), (a))

Change now made, thanks.

> !! But this breaks the build for arm here.
> !! Not sure what is the best way to solve this problem.

Are people building perf for arm on 4.19.y?  If so, wow, this thing is
about to go end-of-life any week now, and I would be amazed if it built
at all anymore as I can't get perf to build on _ANY_ lts kernel these
days.

Anyway, I'll make this change and if any arm build fixes want to show
up, I'll be glad to take them.

thanks,

greg k-h