Kernel feature support - filesystems and networking


Ben Hutchings <ben.hutchings@...>
 

I apologise for the long delay in responding to the kernel config files
I was given. Here's a first list of features I would like to *not*
support. I expect to be able to send additional lists soon. Let me
know if you really need some of the features listed below, and if you
have any ideas about how to mitigate the problems I see.

The following features are enabled in some of the configurations I was
sent but seem to be unsupportable over CIP's time scale:

File systems:
- btrfs (CONFIG_BTRFS_FS)
- Too rapidly changing to backport most fixes
- ceph (CONFIG_CEPH_LIB, CONFIG_CEPH_FS)
- Complex network protocol which is likely to need many fixes
- cifs (CONFIG_CIFS_FS)
- Complex network protocol which is likely to need many fixes
- nfs (CONFIG_NFS_FS, CONFIG_NFSD)
- Complex network protocol which is likely to need many fixes
- ufs (CONFIG_UFS_FS)
- No work by current maintainer for 7 years
- xfs (CONFIG_XFS_FS)
- Maintainers unsupportive of backporting fixes

Networking:
- batman-adv (CONFIG_BATMAN_ADV)
- Rapidly changing
- caif (CONFIG_CAIF)
- No work by current maintainer for 5 years
- dcb (CONFIG_DCB)
- No current maintainer
- hsr (CONFIG_HSR)
- No work by current maintainer for 2 years
- llc2 (CONFIG_LLC2)
- No work by current maintainer for 7 years
- phonet (CONFIG_PHONET)
- Dead protocol
- sctp (CONFIG_IP_SCTP)
- Poor security record
- wimax (CONFIG_WIMAX)
- Dead protocol
- x25 (CONFIG_X25)
- Poor security record

I have some doubts about whether these can be supported:

- ntfs (CONFIG_NTFS_FS)
- Fixes are never marked for stable
- overlayfs (CONFIG_OVERLAY_FS)
- I found many apparently important bug fixes in later kernel
versions that have not been backported

Ben.

--
Ben Hutchings
Software Developer, Codethink Ltd.


Takuo Koguchi
 

Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.

Best regards,

Takuo Koguchi,
Hitachi, Ltd.


-----Original Message-----
From: cip-dev-bounces@...
[mailto:cip-dev-bounces@...] On Behalf Of Ben Hutchings
Sent: Monday, May 29, 2017 11:13 AM
To: cip-dev@...
Subject: [!][cip-dev] Kernel feature support - filesystems and networking

I apologise for the long delay in responding to the kernel config files I was given. Here's a
first list of features I would like to *not* support. I expect to be able to send additional lists
soon. Let me know if you really need some of the features listed below, and if you have
any ideas about how to mitigate the problems I see.

The following features are enabled in some of the configurations I was sent but seem to
be unsupportable over CIP's time scale:

File systems:
- btrfs (CONFIG_BTRFS_FS)
- Too rapidly changing to backport most fixes
- ceph (CONFIG_CEPH_LIB, CONFIG_CEPH_FS)
- Complex network protocol which is likely to need many fixes
- cifs (CONFIG_CIFS_FS)
- Complex network protocol which is likely to need many fixes
- nfs (CONFIG_NFS_FS, CONFIG_NFSD)
- Complex network protocol which is likely to need many fixes
- ufs (CONFIG_UFS_FS)
- No work by current maintainer for 7 years
- xfs (CONFIG_XFS_FS)
- Maintainers unsupportive of backporting fixes

Networking:
- batman-adv (CONFIG_BATMAN_ADV)
- Rapidly changing
- caif (CONFIG_CAIF)
- No work by current maintainer for 5 years
- dcb (CONFIG_DCB)
- No current maintainer
- hsr (CONFIG_HSR)
- No work by current maintainer for 2 years
- llc2 (CONFIG_LLC2)
- No work by current maintainer for 7 years
- phonet (CONFIG_PHONET)
- Dead protocol
- sctp (CONFIG_IP_SCTP)
- Poor security record
- wimax (CONFIG_WIMAX)
- Dead protocol
- x25 (CONFIG_X25)
- Poor security record

I have some doubts about whether these can be supported:

- ntfs (CONFIG_NTFS_FS)
- Fixes are never marked for stable
- overlayfs (CONFIG_OVERLAY_FS)
- I found many apparently important bug fixes in later kernel
versions that have not been backported

Ben.

--
Ben Hutchings
Software Developer, Codethink Ltd.


_______________________________________________
cip-dev mailing list
cip-dev@...
https://lists.cip-project.org/mailman/listinfo/cip-dev


Agustin Benito Bethencourt <agustin.benito@...>
 

Hi Ben,

On 29/05/17 03:40, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
- I found many apparently important bug fixes in later kernel
versions that have not been backported

Is there a technical reason for these patches not being backported or is it simply that never was requested/needed before?

OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Based on this request, is it possible for you to estimate the effort to do this?

@Takuo is there a kernel developer from Hitachi that would be willing to work under Ben's supervision/review in backporting these patches?

Best Regards

--
Agustin Benito Bethencourt
Principal Consultant - FOSS at Codethink
agustin.benito@...


Jan Kiszka
 

On 2017-05-29 04:40, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
We plan to use this feature with CIP kernels as well.

Regards,
Jan

Best regards,

Takuo Koguchi,
Hitachi, Ltd.


-----Original Message-----
From: cip-dev-bounces@...
[mailto:cip-dev-bounces@...] On Behalf Of Ben Hutchings
Sent: Monday, May 29, 2017 11:13 AM
To: cip-dev@...
Subject: [!][cip-dev] Kernel feature support - filesystems and networking

I apologise for the long delay in responding to the kernel config files I was given. Here's a
first list of features I would like to *not* support. I expect to be able to send additional lists
soon. Let me know if you really need some of the features listed below, and if you have
any ideas about how to mitigate the problems I see.

The following features are enabled in some of the configurations I was sent but seem to
be unsupportable over CIP's time scale:

File systems:
- btrfs (CONFIG_BTRFS_FS)
- Too rapidly changing to backport most fixes
- ceph (CONFIG_CEPH_LIB, CONFIG_CEPH_FS)
- Complex network protocol which is likely to need many fixes
- cifs (CONFIG_CIFS_FS)
- Complex network protocol which is likely to need many fixes
- nfs (CONFIG_NFS_FS, CONFIG_NFSD)
- Complex network protocol which is likely to need many fixes
- ufs (CONFIG_UFS_FS)
- No work by current maintainer for 7 years
- xfs (CONFIG_XFS_FS)
- Maintainers unsupportive of backporting fixes

Networking:
- batman-adv (CONFIG_BATMAN_ADV)
- Rapidly changing
- caif (CONFIG_CAIF)
- No work by current maintainer for 5 years
- dcb (CONFIG_DCB)
- No current maintainer
- hsr (CONFIG_HSR)
- No work by current maintainer for 2 years
- llc2 (CONFIG_LLC2)
- No work by current maintainer for 7 years
- phonet (CONFIG_PHONET)
- Dead protocol
- sctp (CONFIG_IP_SCTP)
- Poor security record
- wimax (CONFIG_WIMAX)
- Dead protocol
- x25 (CONFIG_X25)
- Poor security record

I have some doubts about whether these can be supported:

- ntfs (CONFIG_NTFS_FS)
- Fixes are never marked for stable
- overlayfs (CONFIG_OVERLAY_FS)
- I found many apparently important bug fixes in later kernel
versions that have not been backported

Ben.

--
Ben Hutchings
Software Developer, Codethink Ltd.


_______________________________________________
cip-dev mailing list
cip-dev@...
https://lists.cip-project.org/mailman/listinfo/cip-dev
_______________________________________________
cip-dev mailing list
cip-dev@...
https://lists.cip-project.org/mailman/listinfo/cip-dev


Ben Hutchings <ben.hutchings@...>
 

On Mon, 2017-05-29 at 02:40 +0000, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Yes, I thought this was popular. I think we will have to find some way
to support it.

Ben.

--
Ben Hutchings
Software Developer, Codethink Ltd.


Ben Hutchings <ben.hutchings@...>
 

On Mon, 2017-05-29 at 04:58 +0100, Agustin Benito Bethencourt wrote:
Hi Ben,

On 29/05/17 03:40, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
>> - I found many apparently important bug fixes in later kernel
versions that have not been backported

Is there a technical reason for these patches not being backported or is
it simply that never was requested/needed before?
I don't know yet. I will ask upstream.

OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Based on this request, is it possible for you to estimate the effort to
do this?
Not yet, but I will try to get an idea of how large a change is needed.

Ben.

@Takuo is there a kernel developer from Hitachi that would be willing to
work under Ben's supervision/review in backporting these patches?
--
Ben Hutchings
Software Developer, Codethink Ltd.


--
Ben Hutchings
Software Developer, Codethink Ltd.


Daniel Sangorrin <daniel.sangorrin@...>
 

-----Original Message-----
From: cip-dev-bounces@... [mailto:cip-dev-bounces@...] On Behalf Of Ben Hutchings
Sent: Wednesday, May 31, 2017 7:16 PM
To: 小口琢夫 / KOGUCHI,TAKUO
Cc: cip-dev@...
Subject: Re: [cip-dev] Kernel feature support - filesystems and networking

On Mon, 2017-05-29 at 02:40 +0000, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Yes, I thought this was popular. I think we will have to find some way
to support it.
During the CIP TSC meeting I proposed using Ubuntu kernel backports.
http://kernel.ubuntu.com/git/ubuntu/ubuntu-xenial.git/log/fs/overlayfs

What do you think?

Thanks,
Daniel


Jan Kiszka
 

On 2017-06-01 15:30, Daniel Sangorrin wrote:
-----Original Message-----
From: cip-dev-bounces@... [mailto:cip-dev-bounces@...] On Behalf Of Ben Hutchings
Sent: Wednesday, May 31, 2017 7:16 PM
To: 小口琢夫 / KOGUCHI,TAKUO
Cc: cip-dev@...
Subject: Re: [cip-dev] Kernel feature support - filesystems and networking

On Mon, 2017-05-29 at 02:40 +0000, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Yes, I thought this was popular. I think we will have to find some way
to support it.
During the CIP TSC meeting I proposed using Ubuntu kernel backports.
http://kernel.ubuntu.com/git/ubuntu/ubuntu-xenial.git/log/fs/overlayfs

What do you think?
Did overlayfs folks looked at/commented on this already? Ideally,
maintainers should have at least a rough look on our backports whenever
possible.

Jan


Daniel Sangorrin <daniel.sangorrin@...>
 

-----Original Message-----
From: Jan Kiszka [mailto:jan.kiszka@...]
Sent: Thursday, June 01, 2017 10:50 PM
To: Daniel Sangorrin; 'Ben Hutchings'; '小口琢夫 / KOGUCHI,TAKUO'
Cc: cip-dev@...
Subject: Re: [cip-dev] Kernel feature support - filesystems and networking

On 2017-06-01 15:30, Daniel Sangorrin wrote:
-----Original Message-----
From: cip-dev-bounces@... [mailto:cip-dev-bounces@...] On Behalf Of Ben Hutchings
Sent: Wednesday, May 31, 2017 7:16 PM
To: 小口琢夫 / KOGUCHI,TAKUO
Cc: cip-dev@...
Subject: Re: [cip-dev] Kernel feature support - filesystems and networking

On Mon, 2017-05-29 at 02:40 +0000, 小口琢夫 / KOGUCHI,TAKUO wrote:
Hi Ben,

Thanks for your efforts.

- overlayfs (CONFIG_OVERLAY_FS)
OVERLAY_FS is one of the features Hitachi wants to use.
I hope you look into this further and I would like to know if other members also want it.
Yes, I thought this was popular. I think we will have to find some way
to support it.
During the CIP TSC meeting I proposed using Ubuntu kernel backports.
http://kernel.ubuntu.com/git/ubuntu/ubuntu-xenial.git/log/fs/overlayfs

What do you think?
Did overlayfs folks looked at/commented on this already? Ideally,
maintainers should have at least a rough look on our backports whenever
possible.
The Ubuntu LTS kernels contain patches back ported by them from upstream and
their own patches which are usually prefixed by UBUNTU. UBUNTU patches are often
based on BUG reports on launchpad.

As for overlayfs, it seems that UBUNTU has cherry-picked quite a few patches by
the main maintainer Miklos Szeredi. UBUNTU patches were mostly added by Seth Forshee
but I couldn't find any review online.

I think it would be nice to have a rough overlook by the maintainers. And we should also
test them ourselves thoroughly. For example, running LXC with overlayfs.

Thanks,
Daniel