Archive for January, 2011

Brad Figg

Kernel Team Meeting Minutes

Meeting Minutes

IRC Log of the meeting.

Meeting minutes.

Agenda

20110125 Meeting Agenda

ARM Status

Nothing new

Release Metrics

Release Meeting Bugs (11 bugs, 11 Blueprints)

Alpha 2 Milestoned Bugs (22 across all packages (down 7))

  • 1 linux kernel bugs (no change)
  • 0 linux-ti-omap bugs (no change)
  • 0 linux-meta-ti-omap bug (no change)

Release Targeted Bugs (211 across all packages (up 67))

  • 20 linux kernel bugs (up 8)
  • 0 linux-ti-omap bugs (no change)
  • 0 linux-meta-ti-omap bug (no change)

Milestoned Features

  • 6 blueprints (Including HWE Blueprints)

Maverick Updates Bugs

  • 56 Linux Bugs (no change)

Lucid Updates Bugs

  • 95 Linux Bugs (down 2)

Bugs with Patches Attached:104 (down 41)

Blueprints: Natty Bug Handling

There are a number of items in various stages of flight. I will be following up with those parties I am holding for information on, but I will begin aggressively closing these items.

Blueprints: Kernel Configuration Review

PV on HVM support (XEN_PCI_PLATFORMDEV) testing is mow complete. The only other remaining action here currently is to report on the final config at beta time, this is waiting on the final kernel version.

Blueprints: Enhancements to the firmware test suite

Changes to fwts (natty development branch):

  • add uefidump to dump out annotated uefi vars from /sys/firmware/efi/vars
  • fix cmosdump hex dump (it was dumping too much data)
  • klog test: add test to check for “Denied AML access to ports”
  • add -s4-min-delay, –s4-max-delay, –s4-delay-delta options

Blueprints: Review of the Stable Maintenance Process (sconklin / bjf)

Nothing new

Status: Cert. Team

So, I need to cover a couple of things today. The first one is the SRU testing for Maverick and Lucid.
Lucid is looking well, we have already covered 62 of the scheduled 75 systems, and we haven’t found any regressions on those. Maverick, we have covered 67 systems of the 75 that we have scheduled. Same thing, so far, we haven’t found any regressions. I would like to point out that we are now using the SRU test suite that Jeff put together.
As usual, you can check the progress of our testing at:
We want to finish testing Lucid today, to be able to have CDs from -updates as soon as possible. We will follow up the tracker bugs once we are done.
The second thing I wanted to cover is those bugs that are blocking certification for some of the systems that should be certified for 11.04. The list of those bugs can be retrieved at:
Please, check the list to avoid those bugs be forgotten.

Why is bug #686333 on that list? Isn’t that a wishlist item?

Blueprints: Ubuntu Kernel Delta Review

All 19 of the 19 personal patch reviews are now done, and all of the outcomes applied. Some work remains on compcache which is now superceeded by zram in staging, but some userspace work is required to switch to it.

Blueprints: Kernel Version and Flavours

Nothing new.

Status: Ecryptfs

– moved underlying fs checks to mount time
– fixed uninitialized value bug
– fixed bug where using interposed dentry value before dentry was interposed
– fixed unencoded shortname being containing invalid values (/00) this would cause problem with older versions of ecryptfs if they tried to list the files in the directory
– experimenting with Graceful fallback, where shortname can be used regardless of longname xattr – this allows
– graceful fallback if longname xattr is missing
– older versions of ecryptfs to mount dirs with longname xattrs, though the longname xattrs may become wrong then
– still need to deal with:
– collision of shortnames with
– other shortnames
– plaintext pass through files
– refcounting of shared xattr name, or making xattr name uniq to dir + shortname

Question: Are upstream taking your patches?

Answer: I am going to post out the latest to ecryptfs mailing list today, and am hoping to push out to LKML later this week.

Status: Natty

The main distro kernel remains at 2.6.37-12.26 (v2.6.37 final based). We are about to upload a v2.6.38-rc2 + git based kernel, this seems to be the first mostly stable version. Much testing will be needed. Good progress the last couple of weeks on work-items, and we are close to caught up for alpha-2 and well below the line overall.

Security & bugfix kernels – Maverick/Lucid/Karmic/Hardy/Dapper (sconklin / bjf)

We’ve entered the testing phase of the current kernel cycle. The next cycle timing may
be affected by the Lucid .2 release, but the stable kernel team will proceed as normal
and produce kernels for -proposed.

We encountered a regression in the Maverick -proposed kernel that affected user with
Radeon hardware. The offending patch was identified and reverted, and the new kernel
in -proposed has been verified to be fixed. This is the bug for that issue:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/703553

Package    Upd/Sec    Proposed    TiP    Verified   
              
dapper linux-source-2.6.15    2.6.15-55.90    2.6.15-55.91    0    0   
              
hardy linux    2.6.24-28.81    2.6.24-28.84    0    0   
              
karmic linux-ec2    2.6.31-307.23    2.6.31-307.24    0    0   
— linux    2.6.31-22.70    2.6.31-22.71    0    0   
              
lucid linux-ec2    2.6.32-311.23    2.6.32-312.24    5    5   
— linux-meta-lts-backport-maverick    2.6.35.22.34    2.6.35.23.35    0    0   
— linux-lts-backport-maverick    2.6.35-22.34~lucid1    2.6.35-23.41~lucid1    0    0   
— linux    2.6.32-27.49    2.6.32-28.55    5    5   
— linux-meta    2.6.32.27.29    2.6.32.28.31    0    0   
— linux-meta-ec2    2.6.32.311.12    2.6.32.312.13    0    0   
              
maverick linux-backports-modules-2.6.35    2.6.35-24.15    2.6.35-25.16    0    0   
— linux-meta-ti-omap4       2.6.35.903.6    0    0   
— linux-meta    2.6.35.24.28    2.6.35.25.32    0    0   
— linux    2.6.35-24.42    2.6.35-25.44    21    21   
              

Incoming Bugs: Regressions

Incoming Bugs
93 Natty Bugs (up 36)
1116 Maverick Bugs (down 7)
994 Lucid Bugs (down 51)
Current regression stats (broken down by release):

regression-update

  • 30 maverick bugs (up 5)
  • 74 lucid bugs (down 1)
  • 6 karmic bugs (no change)
  • 0 hardy bugs (no change)

regression-release

  • 218 maverick bugs (up 30)
  • 203 lucid bugs (up 8)
  • 38 karmic bugs (down 1)
  • 2 hardy bugs (no change)

regression-proposed

  • 17 maverick bugs (up 4)
  • 3 lucid bugs (no change)
  • 1 karmic bug (no change)

Incoming Bugs: Bug day report

The next bug day will be on Tuesday next week. I have sent out the e-mail for it and I will also be blogging to our team voices page.

Triage Status

We have gotten a lot of help from various people in the #ubuntu-bugs channel. I’d like to thank charlie-tca specifically and several others in there whose nicks escape me at the moment for directing questions about kernel bugs to me. There have also been a ton of requests through irc to look at specific bugs. In those cases, and many like them, I am directing folks to the bug triage pages of our wiki. I hope that some of these folks will become regular triagers, but that remains to be seen. :)

Open Discussion or Questions: Raise your hand to be recognized

  • David Henningsson asked about the status of the “document derivative flavour” work item. Andy reported that
    it is still on his list but not done and not a high priority. They took the discussion to the #ubuntu-kernel
    irc channel to hash it out some more.
Brad Figg

Kernel Team Meeting Minutes

Meeting Minutes

IRC Log of the meeting.

Meeting minutes.

Agenda

20110118 Meeting Agenda

ARM Status

Nothing new

Release Metrics

Nothing new

Blueprints: Kernel Configuration Review

Nothing to report here.

Blueprints: Enhancements to the firmware test suite

  • Changes to fwts (natty development branch):

    • 0.21.0 now in Natty universe.
    • wmi test: handle table not found error more gracefully.
    • report off by one error on month in date.
    • add –disassemble-aml.
    • fix null pointer segfault.
    • syntaxcheck test: add 40 advice lines on specific error output.

Blueprints: Handling of Deviations from Standard Kernels

Done.

Blueprints: Review of the Stable Maintenance Process

  • The kernel team will take responsibility for setting the verification-needed tags on bugs at the beginning of the verification phase

Status: Cert. Team

Nothing new

Blueprints: Ubuntu Kernel Delta Review

16 of the 19 personal patch reviews are now done. Some work remains on compcache which is now superceeded by zram in staging, but some userspace work is required to switch to it. Nothing else of interest occuring.

Blueprints: Kernel Version and Flavours

Master kernels are now at v2.6.37 and remain stable. OMAP3 kernels now switched over to the distro master branch and being used in ARM images. ti-omap4 is a little behind the curve and being investigated.

Status: Ecryptfs

Currently working with light testing, if you don’t mind losing your data you can play with it (it doesn’t handle unexpected stuff yet)
current issues, that need to be resolved
– code cleanup
– name collisions
– actual fs name
– multiple uses of same name from different directories
– dentry tag on lookup
– could use inode tag (force verifying for all names on tagged file)
– Graceful fallback
– should fallback to shortname if xattr is missing, currently fill_dir fails
– should fallback to something that earlier versions of ecryptfs can support (ie. a valid shortname).
– currently Fill dir, detects unencrypted names with leading // as shortname tag
– need to expand current testing into a full regression test suite

Status: Natty

(apw) The main distro kernel is now at 2.6.37-12.26 (v2.6.37 final based). We are still in an extended merge window for v2.6.38-rc1, but testing of Ubuntu kernels based on this is looking ok so far.

Security & bugfix kernels – Maverick/Lucid/Karmic/Hardy/Dapper

Package    Upd/Sec    Proposed    TiP    Verified   
              
dapper linux-source-2.6.15    2.6.15-55.90    2.6.15-55.91    0    0   
              
hardy linux    2.6.24-28.81    2.6.24-28.84    0    0   
              
karmic linux-ec2    2.6.31-307.23    2.6.31-307.24    0    0   
— linux    2.6.31-22.70    2.6.31-22.71    0    0   
              
lucid linux-ec2    2.6.32-311.23    2.6.32-312.24    5    5   
— linux-meta-lts-backport-maverick    2.6.35.22.34    2.6.35.23.35    0    0   
— linux-lts-backport-maverick    2.6.35-22.34~lucid1    2.6.35-23.41~lucid1    0    0   
— linux-firmware    1.34.1    1.34.3    2    2   
— linux    2.6.32-27.49    2.6.32-28.55    5    5   
— linux-meta    2.6.32.27.29    2.6.32.28.31    0    0   
              
maverick linux-meta    2.6.35.24.28    2.6.35.25.31    0    0   
— linux-firmware       1.38.3    3    3   
— linux    2.6.35-24.42    2.6.35-25.43    20    18   
              
  • New kernels were uploaded and published to -proposed for all supported releases. We are now in the verification phase for these kernels.
  • There is a regression in Maverick proposed having to do with Radeo graphics ( bug 703553 ) that we’ll be looking into.
  • The kernel currently in -proposed for Lucid is the one which we anticipate being released for the .2 release.

Open Discussion or Questions:

Nothing new.

Jeremy Foshee

Why has my Kernel bug been marked Invalid/Won’t Fix/Fix Released?

As many of you have experienced this week, I have begun a (somewhat) aggressive review of old bugs for the Kernel package. In some cases bugs are being marked Invalid or Won’t Fix, and I am sure some of you will take issue with that. I’d like to explain some of the rationale behind these decisions and, hopefully, allay the majority of your fears concerning these issues.

First, I am not attempting to lessen the difficulty of any of these reported issues. I am marking the majority of the bugs Invalid due to the overwhelming responses of reporters having ‘similar’ issues. This has resulted in a watering down of the issue as reported by the Original Reporter. I do not wish to convey the notion that those of you encountering those issues are not important, quite the opposite. I am simply wishing to close an issue that is not really able to be solved and asking those of you experiencing issues of a similar nature in unique bugs of your own. This empowers me as a triager to drill down to individual causes of your issue and get that information in front of the team as opposed to trying to understand a variety of, possibly unrelated, issues from a multitude of reporters.

In some cases, you will see the Invalid status being applied to bugs that are very old. In these cases my preference is for affected users to, again, submit me a new bug with all of the relevant information. This gives me all of the benefits I have outlined above in addition to giving us brand new evidence of an issue in the current development release and will give the team a chance to take a fresh look at a longstanding issue.

There are also those bugs that have been Fix Released at some point in the past due to updates to a kernel. Well meaning reporters who experience a seemingly similar issue have, in the past, reopened these under an incorrect status and they have been lost in the shuffle of thousands of other issues. Let me say right now that it is not our policy to reopen Fixed bugs and I consider it a major no-no. :-) To those of you I fuss at for doing this, I apologize in advance.

Some of you will have seen bugs that are set to Won’t Fix. These are issues that are (once again, in most cases) being defined as hardware faults, BIOS issues or unrelated software configuration problems and cannot be fixed by changes in the kernel. In these cases, they are being marked Won’t Fix not our of a desire to keep from actually fixing an issue but in an effort to allow those invested in a product to take the opportunity to solve the problem in the correct location.

I speak from an experience around the team when I say that we would love to solve all of the issues presented to us. The reality is that we are able to take time on very few reported issues directly. The vast majority of fixes come through upstream commits or via new drivers provided by third parties.

To those of you who get annoyed when it seems that all I ever say on a bug is, “Is this still a problem in ” all I can do is shrug. Given the avenues open to us whereby a possible fix could come, I can only ask in that way. Please understand, I’d prefer there were some easier way. :-) Thanks for your patience as the team and I work to make this one of the easiest to use distributions while supporting the most hardware possible.

Finally, I’d like to say that, with such a broad install base and with such a large number of legacy devices available in the wild, there will be some cases where our answer to whether we will be able to support your specific device may end up being “No”. Please forgive me in those instances when I must be the bearer of bad news.

Thank you to all of you who help to triage bugs, update wiki pages and attend learning events put out by the Community team. With all of your help we can continue to improve our chosen operating system and make it more useful to those who would be free.

~JFo

Brad Figg

Kernel Team Meeting Minutes

Meeting Minutes

IRC Log of the meeting.

Meeting minutes.

Agenda

20110104 Meeting Agenda

ARM Status

It looks like we will start supplying an omap3 natty kernel.

Release Metrics

Release Meeting Bugs (6 bugs, 14 Blueprints)

Alpha 2 Milestoned Bugs (29 across all packages (up 1))

  • 1 linux kernel bugs (no change)
  • 0 linux-ti-omap bugs (no change)
  • 0 linux-meta-ti-omap bug (no change)

Release Targeted Bugs (144 across all packages (up 11))

  • 12 linux kernel bugs (up 1)
  • 0 linux-ti-omap bugs (no change)
  • 0 linux-meta-ti-omap bug (no change)

Milestoned Features

  • 5 blueprints (Including HWE Blueprints)

Maverick Updates Bugs

  • 56 Linux Bugs (up 2)

Lucid Updates Bugs

  • 97 Linux Bugs (down 14)

Bugs with Patches Attached:145 (up 3)

Blueprints: Natty Bug Handling

Nothing to report.

Blueprints: Kernel Configuration Review

PV on HVM support (XEN_PCI_PLATFORMDEV) testing is still pending for natty-alpha-2. The only other remaining action here currently is to report on the final config at beta time, this is waiting on the final kernel version.

Blueprints: Enhancements to the firmware test suite

Nothing to report

Blueprints: Handling of Deviations from Standard Kernels

I got a little farther with the tool by now. Detection seems to work. The output needs a bit of polish still.

Blueprints: Review of the Stable Maintenance Process (sconklin / bjf)

The new cadence is working pretty well, but we continue to address issues as they arise. These mostly have to do with
insuring adequate communication between various teams. I’m concerned that thw two-week cycle may not leave us enough
time for exceptions and tools development, and we will be discussing this at the platform rally next week.
We’re planning to begin a cycle by uploading new kernels this Friday, which will put us in verification phase during
the rally.

Status: Cert. Team

We found this bug before the holidays:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/693042

and it wasn’t being worked upon since then, but I have seen that now apw is on it.

That one is looking like it may be an issue with upstart, see comment #7.

I will try this week with the latest ISOs, although it seems it is still there.

Blueprints: Ubuntu Kernel Delta Review

15 of the 19 personal patch reviews are now done. All of the ubuntu drivers are now updated and reviewed. Some work remains on compcache which is now superceeded by zram in staging, but some userspace work is required to switch to it.

Blueprints: Kernel Version and Flavours

We are still working to determine how to provide omap3 kernels, testing is ongoing of a master based kernel.

Status: Ecryptfs

Prototype is in progress, but isn’t up yet. It should be working in a day or two. I started messing with putting the long names in the header but switched to xattrs due to the factors detailed below.

  • Storing long names in meta data is problematic for hardlinks – this can be handled somewhat by either allowing a single longname or by allowing multiple longnames as space allows fail over to storing long names in header
  • long directory names – problematic. There is no header for directories
  • long symlink names – problematic.
  • require update of header (reencryption) on rename fail over to storing long names in xattrs
  • problematic for filesystems that don’t support xattrs
  • leaks fact that long name is present unless all files are given longname xattr
  • requires update of xattr (reencryption) on rename
    Combine current FNEK names and longnames requires encoding information to that a given name was long and stored differently, this leaks some information about the filename. This leak provides more information than the xattr on the file, in that it provides information on which dentry is long when a file has multiple names.

Status: Natty

The main distro kernel is now at 2.6.37-11.25 (v2.6.37-rc7 based). v2.6.37-rc8 has already released and we are expecting v2.6.37-final shortly, we are planning one more upload before the Rally.

Security & bugfix kernels – Maverick/Lucid/Karmic/Hardy/Dapper (sconklin / bjf)

   Package    Upd/Sec    Proposed    TiP    Verified   
                 
   dapper linux    2.6.15.56    2.6.15.56    0    0   
                 
   karmic linux-ec2    2.6.31-307.22    2.6.31-307.23    0    0   
   — linux    2.6.31-22.69    2.6.31-22.70    0    0   
                 
   lucid linux    2.6.32.27.49    2.6.32.27.49    0    0   
   — linux-meta-lts-backport-maverick    2.6.35.22.34    2.6.35.23.35    0    0   
   — linux-lts-backport-maverick    2.6.35-22.34~lucid1    2.6.35-23.41~lucid1    0    0   
   — linux-firmware    1.34.1    1.34.2    1    0   
                 
   maverick linux    2.6.35-24.42    2.6.35-903.20    0    0   
   — linux-ti-omap4    2.6.35-903.19    2.6.35-903.20    1    0   
   — linux-meta    2.6.35.24.28    2.6.35.24.29    3    1   
   — linux-firmware       1.38.2    2    0   
                 

Incoming Bugs: Regressions

Incoming Bugs
57 Natty Bugs (up 19)
1123 Maverick Bugs (down 4)
1045 Lucid Bugs (down 51)
Current regression stats (broken down by release):

regression-update

  • 25 maverick bugs (down 1)
  • 75 lucid bugs (down 1)
  • 6 karmic bugs (no change)
  • 0 hardy bugs (no change)

regression-release

  • 188 maverick bugs (up 11)
  • 195 lucid bugs (down 3)
  • 39 karmic bugs (down 1)
  • 2 hardy bugs (no change)

regression-proposed

  • 13 maverick bugs (no change)
  • 3 lucid bugs (no change)
  • 1 karmic bug (no change)

Incoming Bugs: Bug day report

The next bug day will be next week. As we are just returning from the holiday’s and we are winding back up,
I’d like to focus on bugs in the new state and getting them moved in the appropriate direction. I’ll send
the information out later today or tomorrow.

Triage Status

I am continuing testing on the script that replaces one tag with another. I should have it finished this week.

Open Discussion or Questions:

Nothing new.