User-mode Linux kernel port

On Tue, Jun 19, 2012 at 10:55 AM, David Fernández <david@...> wrote:
> Any idea of the cause of the problem? A simple 'diff' between 3.3 and
> 3.4 versions shows the option parsing code has changed significantly.

Can you do a git bisect?

-- 
Thanks,
//richard

El 19/06/2012, a las 11:03, richard -rw- weinberger escribió:

> On Tue, Jun 19, 2012 at 10:55 AM, David Fernández <david@...> wrote:
>> Any idea of the cause of the problem? A simple 'diff' between 3.3 and
>> 3.4 versions shows the option parsing code has changed significantly.
> 
> Can you do a git bisect?

I've some problems with git bisect, but I manage to make the tests manually and the result is:

v3.3.8 -> OK
v3.4-rc1 -> Could not compile (implicit declaration of switch_to function)
v3.4-rc2 -> Could not compile (implicit declaration of switch_to function)
v3.4-rc3 -> Fails

Let me know if I can help with any other test.

Best regards,
David

> 
> -- 
> Thanks,
> //richard

On Wed, Jun 20, 2012 at 12:33 AM, David Fernández <david@...> wrote:
>
> El 19/06/2012, a las 11:03, richard -rw- weinberger escribió:
>
>> On Tue, Jun 19, 2012 at 10:55 AM, David Fernández <david@...> wrote:
>>> Any idea of the cause of the problem? A simple 'diff' between 3.3 and
>>> 3.4 versions shows the option parsing code has changed significantly.
>>
>> Can you do a git bisect?
>
> I've some problems with git bisect, but I manage to make the tests manually and the result is:

You have to bisect between v3.3 (Linus' tree) and v3.4-rc1.
To make 3.4-rc1 compile you have to always to apply a fix.
AFAIK the commit a3a85a763 (um: Disintegrate asm/system.h) fixes this.

-- 
Thanks,
//richard

On 20/06/12 09:17, richard -rw- weinberger wrote:
> On Wed, Jun 20, 2012 at 12:33 AM, David Fernández<david@...>  wrote:
>> El 19/06/2012, a las 11:03, richard -rw- weinberger escribió:
>>
>>> On Tue, Jun 19, 2012 at 10:55 AM, David Fernández<david@...>  wrote:
>>>> Any idea of the cause of the problem? A simple 'diff' between 3.3 and
>>>> 3.4 versions shows the option parsing code has changed significantly.
>>> Can you do a git bisect?
>> I've some problems with git bisect, but I manage to make the tests manually and the result is:
> You have to bisect between v3.3 (Linus' tree) and v3.4-rc1.
> To make 3.4-rc1 compile you have to always to apply a fix.
> AFAIK the commit a3a85a763 (um: Disintegrate asm/system.h) fixes this.

Finally I managed to do the git bisect with your indications. The result:

# git bisect bad
026cee0086fe1df4cf74691cf273062cc769617d is the first bad commit
commit 026cee0086fe1df4cf74691cf273062cc769617d
Author: Pawel Moll <pawel.moll@...>
Date:   Mon Mar 26 12:50:51 2012 +1030

     params: <level>_initcall-like kernel parameters

     This patch adds a set of macros that can be used to declare
     kernel parameters to be parsed _before_ initcalls at a chosen
     level are executed.  We rename the now-unused "flags" field of
     struct kernel_param as the level.  It's signed, for when we
     use this for early params as well, in future.

     Linker macro collating init calls had to be modified in order
     to add additional symbols between levels that are later used
     by the init code to split the calls into blocks.

     Signed-off-by: Pawel Moll <pawel.moll@...>
     Signed-off-by: Rusty Russell <rusty@...>

:040000 040000 f5622d1cb7f6933aa18fae3e30069d759deafc71 
84f4f9747e3094c061308589b001c29c5d3234ab M    arch
:040000 040000 4809a8f0ac006bd718c16a33c987e593c4673654 
4e6b5d41ce4ad1ea695463a95dfce5595181435d M    include
:040000 040000 36f9650000abe4d15d8859985f93bda1b8ef6cdf 
644e4fe5f6501505144dc0de6c8ff9b79d3b5d65 M    init
:040000 040000 edd20969e15db976a1923f3bc63315c5ff58eecf 
005f7039b2853d3f2571b50d695a4aaffbe4f1d0 M    kernel

An  the log:

# git bisect log
git bisect start
# good: [c16fa4f2ad19908a47c63d8fa436a1178438c7e7] Linux 3.3
git bisect good c16fa4f2ad19908a47c63d8fa436a1178438c7e7
# bad: [dd775ae2549217d3ae09363e3edb305d0fa19928] Linux 3.4-rc1
git bisect bad dd775ae2549217d3ae09363e3edb305d0fa19928
# good: [b2094ef840697bc8ca5d17a83b7e30fad5f1e9fa] Merge tag 'sound-3.4' 
of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
git bisect good b2094ef840697bc8ca5d17a83b7e30fad5f1e9fa
# good: [1bfecd935849a45b6b47d9f011e1c278ff880512] Merge tag 'boards' of 
git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
git bisect good 1bfecd935849a45b6b47d9f011e1c278ff880512
# bad: [47b816ff7d520509176154748713e7d66b3ad6ac] Merge branch 'next' of 
git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
git bisect bad 47b816ff7d520509176154748713e7d66b3ad6ac
# good: [69e1aaddd63104f37021d0b0f6abfd9623c9134c] Merge tag 
'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
git bisect good 69e1aaddd63104f37021d0b0f6abfd9623c9134c
# good: [30304e5a79d424eb2c8707b3ff0e9b8bf6ab3e8f] Merge tag 'mfd_3.4-1' 
of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-2.6
git bisect good 30304e5a79d424eb2c8707b3ff0e9b8bf6ab3e8f
# bad: [d25413efa9536e2f425ea45c7720598035c597bc] Merge 
git://github.com/rustyrussell/linux
git bisect bad d25413efa9536e2f425ea45c7720598035c597bc
# good: [ade28abdcb474531bb7045c032a286812c7f6d2a] asus-wmi: don't 
update power and brightness when using scalar
git bisect good ade28abdcb474531bb7045c032a286812c7f6d2a
# good: [5430528b8bfa2c4d10df3bd59b73ab5faf6fa6cb] Merge branch 
'for_3.4/fixes/gpio-2' of 
git://git.kernel.org/pub/scm/linux/kernel/git/khilman/linux-omap-pm into 
gpio/next
git bisect good 5430528b8bfa2c4d10df3bd59b73ab5faf6fa6cb
# good: [61e5191c9d96268746bd57ed55d035678a1a2cf9] Merge branch 
'for_linus' of git://cavan.codon.org.uk/platform-drivers-x86
git bisect good 61e5191c9d96268746bd57ed55d035678a1a2cf9
# good: [0142877aa4e54dd9943fb727e9b386c36c8e3ab7] i5400_edac: Avoid 
calling pci_put_device() twice
git bisect good 0142877aa4e54dd9943fb727e9b386c36c8e3ab7
# bad: [f946eeb9313ff1470758e171a60fe7438a2ded3f] module: Remove module 
size limit
git bisect bad f946eeb9313ff1470758e171a60fe7438a2ded3f
# good: [8b8252813dee8e8cd453bb219731c36b268c69a7] module_param: remove 
support for bool parameters which are really int.
git bisect good 8b8252813dee8e8cd453bb219731c36b268c69a7
# bad: [d53799be6758841e1ffb1fd3780f73d0ffe44432] module: move 
__module_get and try_module_get() out of line.
git bisect bad d53799be6758841e1ffb1fd3780f73d0ffe44432
# bad: [026cee0086fe1df4cf74691cf273062cc769617d] params: 
<level>_initcall-like kernel parameters
git bisect bad 026cee0086fe1df4cf74691cf273062cc769617d

Let me know if you need any other test.

Best regards,
David

On Wed, Jun 20, 2012 at 7:02 PM, David Fernández <david@...> wrote:
> Finally I managed to do the git bisect with your indications. The result:
>
> # git bisect bad
> 026cee0086fe1df4cf74691cf273062cc769617d is the first bad commit
> commit 026cee0086fe1df4cf74691cf273062cc769617d
> Author: Pawel Moll <pawel.moll@...>
> Date:   Mon Mar 26 12:50:51 2012 +1030
>
>    params: <level>_initcall-like kernel parameters
>
>    This patch adds a set of macros that can be used to declare
>    kernel parameters to be parsed _before_ initcalls at a chosen
>    level are executed.  We rename the now-unused "flags" field of
>    struct kernel_param as the level.  It's signed, for when we
>    use this for early params as well, in future.
>
>    Linker macro collating init calls had to be modified in order
>    to add additional symbols between levels that are later used
>    by the init code to split the calls into blocks.
>
>    Signed-off-by: Pawel Moll <pawel.moll@...>
>    Signed-off-by: Rusty Russell <rusty@...>

Interesting.
Reverting this commit makes the problem go away?

-- 
Thanks,
//richard

On 20/06/12 19:13, richard -rw- weinberger wrote:
> On Wed, Jun 20, 2012 at 7:02 PM, David Fernández<david@...>  wrote:
>> Finally I managed to do the git bisect with your indications. The result:
>>
>> # git bisect bad
>> 026cee0086fe1df4cf74691cf273062cc769617d is the first bad commit
>> commit 026cee0086fe1df4cf74691cf273062cc769617d
>> Author: Pawel Moll<pawel.moll@...>
>> Date:   Mon Mar 26 12:50:51 2012 +1030
>>
>>     params:<level>_initcall-like kernel parameters
>>
>>     This patch adds a set of macros that can be used to declare
>>     kernel parameters to be parsed _before_ initcalls at a chosen
>>     level are executed.  We rename the now-unused "flags" field of
>>     struct kernel_param as the level.  It's signed, for when we
>>     use this for early params as well, in future.
>>
>>     Linker macro collating init calls had to be modified in order
>>     to add additional symbols between levels that are later used
>>     by the init code to split the calls into blocks.
>>
>>     Signed-off-by: Pawel Moll<pawel.moll@...>
>>     Signed-off-by: Rusty Russell<rusty@...>
> Interesting.
> Reverting this commit makes the problem go away?

What command should I use to revert that specific commit? I'm not an 
expert in git, in fact I just learnt what a 'git bisect' is after your 
message ...:-)

On Thu, Jun 21, 2012 at 4:29 PM, David Fernández <david@...> wrote:
>
> What command should I use to revert that specific commit? I'm not an expert
> in git, in fact I just learnt what a 'git bisect' is after your message
> ...:-)

E.g:
$ git show 026cee00 > sucker.diff
$ patch -p1 -R < sucker.diff

-- 
Thanks,
//richard

On Thu, Jun 21, 2012 at 4:42 PM, richard -rw- weinberger
<richard.weinberger@...> wrote:
> On Thu, Jun 21, 2012 at 4:29 PM, David Fernández <david@...> wrote:
>>
>> What command should I use to revert that specific commit? I'm not an expert
>> in git, in fact I just learnt what a 'git bisect' is after your message
>> ...:-)
>
> E.g:
> $ git show 026cee00 > sucker.diff
> $ patch -p1 -R < sucker.diff

    git revert 026cee00

or

    git revert -n 026cee00

if you don't want to commit the revert.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Hi, 

I have just repeated the git bisect and confirmed that:
- the problem was introduced in 026cee00  
- reverting 026cee00 the problem goes away

That commit includes some modifications in kernel/params.c file. Probably the bug has been introduced by that modifications.

Best regards,
David


El 21/06/2012, a las 16:49, Geert Uytterhoeven escribió:

> On Thu, Jun 21, 2012 at 4:42 PM, richard -rw- weinberger
> <richard.weinberger@...> wrote:
>> On Thu, Jun 21, 2012 at 4:29 PM, David Fernández <david@...> wrote:
>>> 
>>> What command should I use to revert that specific commit? I'm not an expert
>>> in git, in fact I just learnt what a 'git bisect' is after your message
>>> ...:-)
>> 
>> E.g:
>> $ git show 026cee00 > sucker.diff
>> $ patch -p1 -R < sucker.diff
> 
>    git revert 026cee00
> 
> or
> 
>    git revert -n 026cee00
> 
> if you don't want to commit the revert.
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds