[meta-oe] fio: upgrade 3.17 -> 3.18

Submitted by Wang Mingyu on Feb. 14, 2020, 11:43 p.m. | Patch ID: 170042

Details

Message ID 1581723823-21171-2-git-send-email-wangmy@cn.fujitsu.com
State Superseded
Commit 07030ea0e87a64c84b74faebf2620eefcf492d19
Headers show

Commit Message

Wang Mingyu Feb. 14, 2020, 11:43 p.m.
Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com>
---
 meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
 rename meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} (96%)

Patch hide | download patch | download mbox

diff --git a/meta-oe/recipes-benchmark/fio/fio_3.17.bb b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
similarity index 96%
rename from meta-oe/recipes-benchmark/fio/fio_3.17.bb
rename to meta-oe/recipes-benchmark/fio/fio_3.18.bb
index 482928c31..b6484290f 100644
--- a/meta-oe/recipes-benchmark/fio/fio_3.17.bb
+++ b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
@@ -22,7 +22,7 @@  PACKAGECONFIG_NUMA_armeb = ""
 PACKAGECONFIG ??= "${PACKAGECONFIG_NUMA}"
 PACKAGECONFIG[numa] = ",--disable-numa,numactl"
 
-SRCREV = "08ce9dc20b8a4e55db7af6d869ddfa49b4a02d03"
+SRCREV = "ac694f66968fe7b18c820468abd8333f3df333fb"
 SRC_URI = "git://git.kernel.dk/fio.git \
           file://0001-update-the-interpreter-paths.patch \
 "

Comments

Khem Raj Feb. 15, 2020, 8:40 a.m.
https://errors.yoctoproject.org/Errors/Details/391129/

On Fri, Feb 14, 2020 at 8:11 AM Wang Mingyu <wangmy@cn.fujitsu.com> wrote:
>
> Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com>
> ---
>  meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>  rename meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} (96%)
>
> diff --git a/meta-oe/recipes-benchmark/fio/fio_3.17.bb b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
> similarity index 96%
> rename from meta-oe/recipes-benchmark/fio/fio_3.17.bb
> rename to meta-oe/recipes-benchmark/fio/fio_3.18.bb
> index 482928c31..b6484290f 100644
> --- a/meta-oe/recipes-benchmark/fio/fio_3.17.bb
> +++ b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
> @@ -22,7 +22,7 @@ PACKAGECONFIG_NUMA_armeb = ""
>  PACKAGECONFIG ??= "${PACKAGECONFIG_NUMA}"
>  PACKAGECONFIG[numa] = ",--disable-numa,numactl"
>
> -SRCREV = "08ce9dc20b8a4e55db7af6d869ddfa49b4a02d03"
> +SRCREV = "ac694f66968fe7b18c820468abd8333f3df333fb"
>  SRC_URI = "git://git.kernel.dk/fio.git \
>            file://0001-update-the-interpreter-paths.patch \
>  "
> --
> 2.17.1
>
>
>
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel
I was just about to submit a patch to remove the disabling of compiler
optimizations to fio. On our cortex A72 this kills the performance of fio
by 25%. That would solve this build issue also. I haven't seen any crashes
with fio. If there are some architecture specific issues can the removal
of optimizations be arch specific only and not the default?

On Sat, Feb 15, 2020 at 12:40 AM Khem Raj <raj.khem@gmail.com> wrote:
>
> https://errors.yoctoproject.org/Errors/Details/391129/
>
> On Fri, Feb 14, 2020 at 8:11 AM Wang Mingyu <wangmy@cn.fujitsu.com> wrote:
> >
> > Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com>
> > ---
> >  meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >  rename meta-oe/recipes-benchmark/fio/{fio_3.17.bb => fio_3.18.bb} (96%)
> >
> > diff --git a/meta-oe/recipes-benchmark/fio/fio_3.17.bb b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
> > similarity index 96%
> > rename from meta-oe/recipes-benchmark/fio/fio_3.17.bb
> > rename to meta-oe/recipes-benchmark/fio/fio_3.18.bb
> > index 482928c31..b6484290f 100644
> > --- a/meta-oe/recipes-benchmark/fio/fio_3.17.bb
> > +++ b/meta-oe/recipes-benchmark/fio/fio_3.18.bb
> > @@ -22,7 +22,7 @@ PACKAGECONFIG_NUMA_armeb = ""
> >  PACKAGECONFIG ??= "${PACKAGECONFIG_NUMA}"
> >  PACKAGECONFIG[numa] = ",--disable-numa,numactl"
> >
> > -SRCREV = "08ce9dc20b8a4e55db7af6d869ddfa49b4a02d03"
> > +SRCREV = "ac694f66968fe7b18c820468abd8333f3df333fb"
> >  SRC_URI = "git://git.kernel.dk/fio.git \
> >            file://0001-update-the-interpreter-paths.patch \
> >  "
> > --
> > 2.17.1
> >
> >
> >
> > --
> > _______________________________________________
> > Openembedded-devel mailing list
> > Openembedded-devel@lists.openembedded.org
> > http://lists.openembedded.org/mailman/listinfo/openembedded-devel
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel
Adrian Bunk Feb. 18, 2020, 10:09 p.m.
On Tue, Feb 18, 2020 at 11:37:33AM -0800, Jonathan Richardson via Openembedded-devel wrote:
> I was just about to submit a patch to remove the disabling of compiler
> optimizations to fio. On our cortex A72 this kills the performance of fio
> by 25%. That would solve this build issue also.
>...
> > https://errors.yoctoproject.org/Errors/Details/391129/
>...

t/io_uring.c:103:17: error: '__NR_sys_io_uring_register' undeclared (first use in this function); did you mean 'SYS_io_uring_register'?

This error does not look related to optimization.

This looks like https://github.com/axboe/fio/issues/923

cu
Adrian
On Tue, Feb 18, 2020 at 2:09 PM Adrian Bunk <bunk@stusta.de> wrote:
>
> On Tue, Feb 18, 2020 at 11:37:33AM -0800, Jonathan Richardson via Openembedded-devel wrote:
> > I was just about to submit a patch to remove the disabling of compiler
> > optimizations to fio. On our cortex A72 this kills the performance of fio
> > by 25%. That would solve this build issue also.
> >...
> > > https://errors.yoctoproject.org/Errors/Details/391129/
> >...
>
> t/io_uring.c:103:17: error: '__NR_sys_io_uring_register' undeclared (first use in this function); did you mean 'SYS_io_uring_register'?
>
> This error does not look related to optimization.
>
> This looks like https://github.com/axboe/fio/issues/923

Yes you're right. The error is another issue. The message I was referring
to is actually only a warning. It should fix the warning only:

  397 | #  warning _FORTIFY_SOURCE requires compiling with optimization (-O)
Khem Raj Feb. 19, 2020, 2:36 a.m.
On Tue, Feb 18, 2020 at 2:09 PM Adrian Bunk <bunk@stusta.de> wrote:
>
> On Tue, Feb 18, 2020 at 11:37:33AM -0800, Jonathan Richardson via Openembedded-devel wrote:
> > I was just about to submit a patch to remove the disabling of compiler
> > optimizations to fio. On our cortex A72 this kills the performance of fio
> > by 25%. That would solve this build issue also.
> >...
> > > https://errors.yoctoproject.org/Errors/Details/391129/
> >...
>
> t/io_uring.c:103:17: error: '__NR_sys_io_uring_register' undeclared (first use in this function); did you mean 'SYS_io_uring_register'?
>
> This error does not look related to optimization.
>
> This looks like https://github.com/axboe/fio/issues/923
>

Thanks, I think this will fix the issue we have at hand, Wang would
you include this in a v2 ?

> cu
> Adrian