Hi Mike,
After merging the clk tree, today's linux-next build (x86_64 allmodconfig)
failed like this:
drivers/clk/clk.c: In function 'clk_disable_unused_subtree':
drivers/clk/clk.c:514:3: error: label 'out' used but not defined
goto out;
^
Caused by commit a2146f032294 ("clk: Use lockdep asserts to find
missing hold of prepare_lock"). Commit c440525cb967 ("clk: Remove
unneeded NULL checks") removed that label along with the NULL check
that a2146f032294 reintroduces (was this a bad rebase?). Please do
simple build tests.
I have used the clk tree from next-20150225 for today.
--
Cheers,
Stephen Rothwell [email protected]
On czw, 2015-02-26 at 13:14 +1100, Stephen Rothwell wrote:
> Hi Mike,
>
> After merging the clk tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
>
> drivers/clk/clk.c: In function 'clk_disable_unused_subtree':
> drivers/clk/clk.c:514:3: error: label 'out' used but not defined
> goto out;
> ^
>
> Caused by commit a2146f032294 ("clk: Use lockdep asserts to find
> missing hold of prepare_lock"). Commit c440525cb967 ("clk: Remove
> unneeded NULL checks") removed that label along with the NULL check
> that a2146f032294 reintroduces (was this a bad rebase?). Please do
> simple build tests.
>
> I have used the clk tree from next-20150225 for today.
Mike,
It seems that my patch did not applied cleanly and the merge introduced
such artifacts. My patch adds only lockdep_asserts and does not
influence the program flow. I can rebase and resend the patch if you
wish.
Best regards,
Krzysztof
On Thu, Feb 26, 2015 at 1:09 AM, Krzysztof Kozlowski
<[email protected]> wrote:
> On czw, 2015-02-26 at 13:14 +1100, Stephen Rothwell wrote:
>> Hi Mike,
>>
>> After merging the clk tree, today's linux-next build (x86_64 allmodconfig)
>> failed like this:
>>
>> drivers/clk/clk.c: In function 'clk_disable_unused_subtree':
>> drivers/clk/clk.c:514:3: error: label 'out' used but not defined
>> goto out;
>> ^
>>
>> Caused by commit a2146f032294 ("clk: Use lockdep asserts to find
>> missing hold of prepare_lock"). Commit c440525cb967 ("clk: Remove
>> unneeded NULL checks") removed that label along with the NULL check
>> that a2146f032294 reintroduces (was this a bad rebase?). Please do
>> simple build tests.
>>
>> I have used the clk tree from next-20150225 for today.
>
> Mike,
>
> It seems that my patch did not applied cleanly and the merge introduced
> such artifacts. My patch adds only lockdep_asserts and does not
> influence the program flow. I can rebase and resend the patch if you
> wish.
It's OK. I fixed it up locally the same day that I applied it but
didn't push the change out to my mirror in time to catch -next. It
will be fixed the next time Stephen pulls the clk tree.
Regards,
Mike
>
> Best regards,
> Krzysztof
>
Hi Mike,
On Thu, 26 Feb 2015 16:01:24 -0800 Mike Turquette <[email protected]> wrote:
>
> It's OK. I fixed it up locally the same day that I applied it but
> didn't push the change out to my mirror in time to catch -next. It
> will be fixed the next time Stephen pulls the clk tree.
Which I have just done. Thanks.
--
Cheers,
Stephen Rothwell [email protected]