Hi all,
After merging the ftrace tree, today's linux-next build (powerpc
ppc64_defconfig) failed like this:
Inconsistent kallsyms data
Try make KALLSYMS_EXTRA_PASS=1 as a workaround
I am not sure what caused this - maybe commit
4f1f18311591 ("scripts: ftrace - move the sort-processing in ftrace_init")
I have used the ftrace tree from next-20220111 for today.
--
Cheers,
Stephen Rothwell
Hi all,
On Wed, 12 Jan 2022 11:31:03 +1100 Stephen Rothwell <[email protected]> wrote:
>
> After merging the ftrace tree, today's linux-next build (powerpc
> ppc64_defconfig) failed like this:
>
> Inconsistent kallsyms data
> Try make KALLSYMS_EXTRA_PASS=1 as a workaround
>
> I am not sure what caused this - maybe commit
>
> 4f1f18311591 ("scripts: ftrace - move the sort-processing in ftrace_init")
>
> I have used the ftrace tree from next-20220111 for today.
I am still getting this.
--
Cheers,
Stephen Rothwell
On Fri, 14 Jan 2022 12:02:40 +1100
Stephen Rothwell <[email protected]> wrote:
> Hi all,
>
> On Wed, 12 Jan 2022 11:31:03 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > After merging the ftrace tree, today's linux-next build (powerpc
> > ppc64_defconfig) failed like this:
> >
> > Inconsistent kallsyms data
> > Try make KALLSYMS_EXTRA_PASS=1 as a workaround
> >
> > I am not sure what caused this - maybe commit
> >
> > 4f1f18311591 ("scripts: ftrace - move the sort-processing in ftrace_init")
> >
> > I have used the ftrace tree from next-20220111 for today.
>
> I am still getting this.
>
OK, I'll take a look.
-- Steve
On Fri, 14 Jan 2022 12:02:40 +1100
Stephen Rothwell <[email protected]> wrote:
> Hi all,
>
> On Wed, 12 Jan 2022 11:31:03 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > After merging the ftrace tree, today's linux-next build (powerpc
> > ppc64_defconfig) failed like this:
> >
> > Inconsistent kallsyms data
> > Try make KALLSYMS_EXTRA_PASS=1 as a workaround
> >
> > I am not sure what caused this - maybe commit
> >
> > 4f1f18311591 ("scripts: ftrace - move the sort-processing in ftrace_init")
> >
> > I have used the ftrace tree from next-20220111 for today.
>
> I am still getting this.
>
What compiler are you using, because I built my branch against all archs
and it didn't have an issue.
I can try to see if it is with linux-next and something my tree has with
other updates.
-- Steve
Hi Steve,
On Thu, 13 Jan 2022 21:16:33 -0500 Steven Rostedt <[email protected]> wrote:
>
> What compiler are you using, because I built my branch against all archs
> and it didn't have an issue.
gcc (Debian 10.3.0-11) 10.3.0
GNU ld (GNU Binutils for Debian) 2.37
I am doing a native build on ppc64le.
> I can try to see if it is with linux-next and something my tree has with
> other updates.
Thanks. There are some changes in the kbuild tree that may interact, I
guess.
--
Cheers,
Stephen Rothwell
Hi Steve,
On Fri, 14 Jan 2022 13:58:08 +1100 Stephen Rothwell <[email protected]> wrote:
>
> On Thu, 13 Jan 2022 21:16:33 -0500 Steven Rostedt <[email protected]> wrote:
> >
> > What compiler are you using, because I built my branch against all archs
> > and it didn't have an issue.
>
> gcc (Debian 10.3.0-11) 10.3.0
> GNU ld (GNU Binutils for Debian) 2.37
>
> I am doing a native build on ppc64le.
>
> > I can try to see if it is with linux-next and something my tree has with
> > other updates.
>
> Thanks. There are some changes in the kbuild tree that may interact, I
> guess.
It also presumably could depend on the actual kernel config being tested.
--
Cheers,
Stephen Rothwell
On Fri, 14 Jan 2022 14:02:58 +1100
Stephen Rothwell <[email protected]> wrote:
> Hi Steve,
>
> On Fri, 14 Jan 2022 13:58:08 +1100 Stephen Rothwell <[email protected]> wrote:
> >
> > On Thu, 13 Jan 2022 21:16:33 -0500 Steven Rostedt <[email protected]> wrote:
> > >
> > > What compiler are you using, because I built my branch against all archs
> > > and it didn't have an issue.
> >
> > gcc (Debian 10.3.0-11) 10.3.0
> > GNU ld (GNU Binutils for Debian) 2.37
> >
> > I am doing a native build on ppc64le.
> >
> > > I can try to see if it is with linux-next and something my tree has with
> > > other updates.
> >
> > Thanks. There are some changes in the kbuild tree that may interact, I
> > guess.
>
> It also presumably could depend on the actual kernel config being tested.
>
I can't reproduce it, and doing searches on this error, seems that it
happens "randomly". I haven't found what can cause it yet.
I don't have a way to build nativly, but if you send me your config, I
can still try to cross compile it to see if it triggers.
Does it fail on anything else?
-- Steve
Hi Steve,
On Thu, 13 Jan 2022 22:25:39 -0500 Steven Rostedt <[email protected]> wrote:
>
> I can't reproduce it, and doing searches on this error, seems that it
> happens "randomly". I haven't found what can cause it yet.
rats :-(
> I don't have a way to build nativly, but if you send me your config, I
> can still try to cross compile it to see if it triggers.
I just use ppc64_defconfig (which may vary depending on what is merged
into my tree at the time). I don't save actual .config files, sorry.
So:
$ make ARCH=powerpc ppc64_defconfig
> Does it fail on anything else?
This is the first build I do after merging each tree, so I don't know
about others. I may be able to do some testing later today (after
linux-next is released).
--
Cheers,
Stephen Rothwell
On Fri, 14 Jan 2022 14:34:18 +1100
Stephen Rothwell <[email protected]> wrote:
> > Does it fail on anything else?
>
> This is the first build I do after merging each tree, so I don't know
> about others. I may be able to do some testing later today (after
> linux-next is released).
Could you see if this fixes it? This is a revert of part of the patch.
Yinan, was there a reason you moved this code?
-- Steve
diff --git a/scripts/link-vmlinux.sh b/scripts/link-vmlinux.sh
index dd9955f45774..5cdd9bc5c385 100755
--- a/scripts/link-vmlinux.sh
+++ b/scripts/link-vmlinux.sh
@@ -400,9 +400,6 @@ if [ -n "${CONFIG_DEBUG_INFO_BTF}" -a -n "${CONFIG_BPF}" ]; then
${RESOLVE_BTFIDS} vmlinux
fi
-info SYSMAP System.map
-mksysmap vmlinux System.map
-
if [ -n "${CONFIG_BUILDTIME_TABLE_SORT}" ]; then
info SORTTAB vmlinux
if ! sorttable vmlinux; then
@@ -411,6 +408,9 @@ if [ -n "${CONFIG_BUILDTIME_TABLE_SORT}" ]; then
fi
fi
+info SYSMAP System.map
+mksysmap vmlinux System.map
+
# step a (see comment above)
if [ -n "${CONFIG_KALLSYMS}" ]; then
mksysmap ${kallsyms_vmlinux} .tmp_System.map
Hi, steven
When sorttable is executed, system.map is queried, so the creation
of system.map is moved forward. Did that make a difference?
在 2022/1/14 上午11:50, Steven Rostedt 写道:
> On Fri, 14 Jan 2022 14:34:18 +1100
> Stephen Rothwell <[email protected]> wrote:
>
>>> Does it fail on anything else?
>>
>> This is the first build I do after merging each tree, so I don't know
>> about others. I may be able to do some testing later today (after
>> linux-next is released).
>
> Could you see if this fixes it? This is a revert of part of the patch.
> Yinan, was there a reason you moved this code?
>
> -- Steve
>
> diff --git a/scripts/link-vmlinux.sh b/scripts/link-vmlinux.sh
> index dd9955f45774..5cdd9bc5c385 100755
> --- a/scripts/link-vmlinux.sh
> +++ b/scripts/link-vmlinux.sh
> @@ -400,9 +400,6 @@ if [ -n "${CONFIG_DEBUG_INFO_BTF}" -a -n "${CONFIG_BPF}" ]; then
> ${RESOLVE_BTFIDS} vmlinux
> fi
>
> -info SYSMAP System.map
> -mksysmap vmlinux System.map
> -
> if [ -n "${CONFIG_BUILDTIME_TABLE_SORT}" ]; then
> info SORTTAB vmlinux
> if ! sorttable vmlinux; then
> @@ -411,6 +408,9 @@ if [ -n "${CONFIG_BUILDTIME_TABLE_SORT}" ]; then
> fi
> fi
>
> +info SYSMAP System.map
> +mksysmap vmlinux System.map
> +
> # step a (see comment above)
> if [ -n "${CONFIG_KALLSYMS}" ]; then
> mksysmap ${kallsyms_vmlinux} .tmp_System.map
Hi Steve
On Thu, 13 Jan 2022 22:50:48 -0500 Steven Rostedt <[email protected]> wrote:
>
> On Fri, 14 Jan 2022 14:34:18 +1100
> Stephen Rothwell <[email protected]> wrote:
>
> > > Does it fail on anything else?
> >
> > This is the first build I do after merging each tree, so I don't know
> > about others. I may be able to do some testing later today (after
> > linux-next is released).
>
> Could you see if this fixes it? This is a revert of part of the patch.
> Yinan, was there a reason you moved this code?
It is my fault, sorry. I got the merge resolution wrong and managed to
remove both copies of:
info SYSMAP System.map
mksysmap vmlinux System.map
putting it back before
if [ -n "${CONFIG_BUILDTIME_TABLE_SORT}" ]; then
makes it work!
I am so sorry. I will fix up my merge resolution for Monday.
--
Cheers,
Stephen Rothwell