[PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Steffen Sledz]

* .get_sb is called on mounts with automatic fs detection too, so this
  function should not print an error if it cannot read the superblock
  (new behaviour conforms the other fs types)

Signed-off-by: Steffen Sledz <sledz@dresearch.de>
---
 .../linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch |   11 +++++++++++
 recipes/linux/linux_2.6.24.bb                      |    1 +
 2 files changed, 12 insertions(+), 0 deletions(-)
 create mode 100644 recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch

diff --git a/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
new file mode 100644
index 0000000..33f5f9f
--- /dev/null
+++ b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
@@ -0,0 +1,11 @@
+--- linux-2.6.24.orig/fs/ubifs/super.c	2010-09-23 09:25:52.000000000 +0200
++++ linux-2.6.24/fs/ubifs/super.c	2010-09-23 09:36:21.000000000 +0200
+@@ -2020,8 +2020,6 @@
+ 	 */
+ 	ubi = open_ubi(name, UBI_READONLY);
+ 	if (IS_ERR(ubi)) {
+-		ubifs_err("cannot open \"%s\", error %d",
+-			  name, (int)PTR_ERR(ubi));
+ 		return PTR_ERR(ubi);
+ 	}
+ 	ubi_get_volume_info(ubi, &vi);
diff --git a/recipes/linux/linux_2.6.24.bb b/recipes/linux/linux_2.6.24.bb
index 9f3655c..982038b 100644
--- a/recipes/linux/linux_2.6.24.bb
+++ b/recipes/linux/linux_2.6.24.bb
@@ -17,6 +17,7 @@ SRC_URI = "${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/linux-2.6.24.tar.bz2;name=k
            ${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/patch-${PV}.7.bz2;apply=yes;name=stablepatch \
            file://squashfs-lzma-2.6.24.patch \
            file://ubifs-v2.6.24.patch \
+           file://ubifs-v2.6.24-silent-get-sb.patch \
            file://defconfig"
 
 # Moved away temporarely until committed properly (work in progress).
-- 
1.7.1

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Steffen Sledz]

Ping!

Am 23.09.2010 10:33, schrieb Steffen Sledz:
> * .get_sb is called on mounts with automatic fs detection too, so this
>   function should not print an error if it cannot read the superblock
>   (new behaviour conforms the other fs types)
> 
> Signed-off-by: Steffen Sledz <sledz@dresearch.de>
> ---
>  .../linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch |   11 +++++++++++
>  recipes/linux/linux_2.6.24.bb                      |    1 +
>  2 files changed, 12 insertions(+), 0 deletions(-)
>  create mode 100644 recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
> 
> diff --git a/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
> new file mode 100644
> index 0000000..33f5f9f
> --- /dev/null
> +++ b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
> @@ -0,0 +1,11 @@
> +--- linux-2.6.24.orig/fs/ubifs/super.c	2010-09-23 09:25:52.000000000 +0200
> ++++ linux-2.6.24/fs/ubifs/super.c	2010-09-23 09:36:21.000000000 +0200
> +@@ -2020,8 +2020,6 @@
> + 	 */
> + 	ubi = open_ubi(name, UBI_READONLY);
> + 	if (IS_ERR(ubi)) {
> +-		ubifs_err("cannot open \"%s\", error %d",
> +-			  name, (int)PTR_ERR(ubi));
> + 		return PTR_ERR(ubi);
> + 	}
> + 	ubi_get_volume_info(ubi, &vi);
> diff --git a/recipes/linux/linux_2.6.24.bb b/recipes/linux/linux_2.6.24.bb
> index 9f3655c..982038b 100644
> --- a/recipes/linux/linux_2.6.24.bb
> +++ b/recipes/linux/linux_2.6.24.bb
> @@ -17,6 +17,7 @@ SRC_URI = "${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/linux-2.6.24.tar.bz2;name=k
>             ${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/patch-${PV}.7.bz2;apply=yes;name=stablepatch \
>             file://squashfs-lzma-2.6.24.patch \
>             file://ubifs-v2.6.24.patch \
> +           file://ubifs-v2.6.24-silent-get-sb.patch \
>             file://defconfig"
>  
>  # Moved away temporarely until committed properly (work in progress).

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Frans Meulenbroeks]

2010/9/27 Steffen Sledz <sledz@dresearch.de>:
> Ping!
>
> Am 23.09.2010 10:33, schrieb Steffen Sledz:
>> * .get_sb is called on mounts with automatic fs detection too, so this
>>   function should not print an error if it cannot read the superblock
>>   (new behaviour conforms the other fs types)
>>
>> Signed-off-by: Steffen Sledz <sledz@dresearch.de>
>> ---
>>  .../linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch |   11 +++++++++++
>>  recipes/linux/linux_2.6.24.bb                      |    1 +
>>  2 files changed, 12 insertions(+), 0 deletions(-)
>>  create mode 100644 recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
>>
>> diff --git a/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
>> new file mode 100644
>> index 0000000..33f5f9f
>> --- /dev/null
>> +++ b/recipes/linux/linux-2.6.24/ubifs-v2.6.24-silent-get-sb.patch
>> @@ -0,0 +1,11 @@
>> +--- linux-2.6.24.orig/fs/ubifs/super.c       2010-09-23 09:25:52.000000000 +0200
>> ++++ linux-2.6.24/fs/ubifs/super.c    2010-09-23 09:36:21.000000000 +0200
>> +@@ -2020,8 +2020,6 @@
>> +      */
>> +     ubi = open_ubi(name, UBI_READONLY);
>> +     if (IS_ERR(ubi)) {
>> +-            ubifs_err("cannot open \"%s\", error %d",
>> +-                      name, (int)PTR_ERR(ubi));
>> +             return PTR_ERR(ubi);
>> +     }
>> +     ubi_get_volume_info(ubi, &vi);
>> diff --git a/recipes/linux/linux_2.6.24.bb b/recipes/linux/linux_2.6.24.bb
>> index 9f3655c..982038b 100644
>> --- a/recipes/linux/linux_2.6.24.bb
>> +++ b/recipes/linux/linux_2.6.24.bb
>> @@ -17,6 +17,7 @@ SRC_URI = "${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/linux-2.6.24.tar.bz2;name=k
>>             ${KERNELORG_MIRROR}/pub/linux/kernel/v2.6/patch-${PV}.7.bz2;apply=yes;name=stablepatch \
>>             file://squashfs-lzma-2.6.24.patch \
>>             file://ubifs-v2.6.24.patch \
>> +           file://ubifs-v2.6.24-silent-get-sb.patch \
>>             file://defconfig"
>>
>>  # Moved away temporarely until committed properly (work in progress).
>

Hm, I have some concerns:

the first concern is that mainline 2.6.34 still has the code you have deleted.
I'd expect that others would have bumped into this too and that it
would have been fixed this upstream.
Since it isn't, chances are that upstream might have fixed the autofs
detection in a different way.  If they do this differently maybe we
should backport that solution.

the second concern is that you fix it for a single version. If we feel
this patch is needed, I would suggest applying the patch to all linux
recipes.

and the 3rd concern is that I am not too sure whether you do not
remove a message that in other cases would be useful.

Best regards, Frans

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Steffen Sledz]

Am 27.09.2010 09:40, schrieb Frans Meulenbroeks:
>>> * .get_sb is called on mounts with automatic fs detection too, so this
>>>   function should not print an error if it cannot read the superblock
>>>   (new behaviour conforms the other fs types)
>>> ...
>
> Hm, I have some concerns:
> 
> the first concern is that mainline 2.6.34 still has the code you have deleted.
> I'd expect that others would have bumped into this too and that it
> would have been fixed this upstream.
> Since it isn't, chances are that upstream might have fixed the autofs
> detection in a different way.  If they do this differently maybe we
> should backport that solution.

I've suggested the same patch at the LKML and informed the UBIFS authors. But i do believe because of it's "low prio" it will need some time to go into the mainline.

> the second concern is that you fix it for a single version. If we feel
> this patch is needed, I would suggest applying the patch to all linux
> recipes.

Yes, that's true. I'll try to create a new patch.

> and the 3rd concern is that I am not too sure whether you do not
> remove a message that in other cases would be useful.

I did not found any other filesystem printing out such a message at this level.

Steffen

-- 
Steffen Sledz
DResearch Digital Media Systems GmbH
Otto-Schmirgal-Str.3, D-10319 Berlin, Germany
Tel: +49 (30) 515932237 mailto:sledz@DResearch.DE
Fax: +49 (30) 515932299 http://www.DResearch.DE
Geschäftsführer: Dr. Michael Weber, Werner Mögle;
Amtsgericht Berlin Charlottenburg; HRB 54412;
Ust.-IDNr. DE169013825; WEEE Reg.-Nr. DE 85995642

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Tom Rini]

Steffen Sledz wrote:
> Am 27.09.2010 09:40, schrieb Frans Meulenbroeks:
>>>> * .get_sb is called on mounts with automatic fs detection too, so this
>>>>   function should not print an error if it cannot read the superblock
>>>>   (new behaviour conforms the other fs types)
>>>> ...
>> Hm, I have some concerns:
>>
>> the first concern is that mainline 2.6.34 still has the code you have deleted.
>> I'd expect that others would have bumped into this too and that it
>> would have been fixed this upstream.
>> Since it isn't, chances are that upstream might have fixed the autofs
>> detection in a different way.  If they do this differently maybe we
>> should backport that solution.
> 
> I've suggested the same patch at the LKML and informed the UBIFS authors. But i do believe because of it's "low prio" it will need some time to go into the mainline.

So it is queued somewhere then?

-- 
Tom Rini
Mentor Graphics Corporation

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Steffen Sledz]

Am 27.09.2010 22:15, schrieb Tom Rini:
>>> the first concern is that mainline 2.6.34 still has the
>>> code you have deleted. I'd expect that others would have
>>> bumped into this too and that it would have been fixed this
>>> upstream. Since it isn't, chances are that upstream might
>>> have fixed the autofs detection in a different way. If they
>>> do this differently maybe we should backport that solution.
>>
>> I've suggested the same patch at the LKML and informed the
>> UBIFS authors. But i do believe because of it's "low prio"
>> it will need some time to go into the mainline.
> 
> So it is queued somewhere then?

Yes, meanwhile in a modified way.

<http://marc.info/?l=linux-kernel&m=128532926311387&w=2>
<http://marc.info/?l=linux-kernel&m=128559006603975&w=2>

I'll wait the final decision there and then push the commit at least for the 2.6.24 backport.

I don't know, if it's really necessary to patch all linux recipes because there are a lot of them and it's just a cosmetic change.

Steffen

-- 
Steffen Sledz
DResearch Digital Media Systems GmbH
Otto-Schmirgal-Str.3, D-10319 Berlin, Germany
Tel: +49 (30) 515932237 mailto:sledz@DResearch.DE
Fax: +49 (30) 515932299 http://www.DResearch.DE
Geschäftsführer: Dr. Michael Weber, Werner Mögle;
Amtsgericht Berlin Charlottenburg; HRB 54412;
Ust.-IDNr. DE169013825; WEEE Reg.-Nr. DE 85995642

Re: [PATCH] linux-2.6.24: avoid kernel error if ubifs superblock read fails

[Steffen Sledz]

Am 28.09.2010 08:10, schrieb Steffen Sledz:
> Am 27.09.2010 22:15, schrieb Tom Rini:
>>>> the first concern is that mainline 2.6.34 still has the
>>>> code you have deleted. I'd expect that others would have
>>>> bumped into this too and that it would have been fixed this
>>>> upstream. Since it isn't, chances are that upstream might
>>>> have fixed the autofs detection in a different way. If they
>>>> do this differently maybe we should backport that solution.
>>>
>>> I've suggested the same patch at the LKML and informed the
>>> UBIFS authors. But i do believe because of it's "low prio"
>>> it will need some time to go into the mainline.
>>
>> So it is queued somewhere then?
> 
> Yes, meanwhile in a modified way.
> 
> <http://marc.info/?l=linux-kernel&m=128532926311387&w=2>
> <http://marc.info/?l=linux-kernel&m=128559006603975&w=2>
> 
> I'll wait the final decision there and then push the commit at
> least for the 2.6.24 backport.

Pushed after upstream confirmation.

Steffen

-- 
Steffen Sledz
DResearch Digital Media Systems GmbH
Otto-Schmirgal-Str.3, D-10319 Berlin, Germany
Tel: +49 (30) 515932237 mailto:sledz@DResearch.DE
Fax: +49 (30) 515932299 http://www.DResearch.DE
Geschäftsführer: Dr. Michael Weber, Werner Mögle;
Amtsgericht Berlin Charlottenburg; HRB 54412;
Ust.-IDNr. DE169013825; WEEE Reg.-Nr. DE 85995642