kernel weekly news – 20.11.2010

Posted: November 16, 2010 in kernel

Hello! Let’s start this week’s news piece with …

-…Mauro Carvalho Chehab, who updated the V4L tree with mostly bug fixes (git pull
for -rc2), then it’s Joel Becker with an important fix for ocfs2 – “a signed-vs-unsigned
error that I should have caught”, also we have Frederic Weisbecker with a two-piece
git pull request regarding the random-tracing tree and Paul Mundt with a few updates/
fixes for a few trees, all targetting -rc2, and those trees are sh, rmobile and fbdev.

-Steven Whitehouse presented a series of updates for gfs2, here is his description :
“GFS2: Fix inode deallocation race

This area of the code has always been a bit delicate due to the
subtleties of lock ordering. The problem is that for “normal”
alloc/dealloc, we always grab the inode locks first and the rgrp lock

In order to ensure no races in looking up the unlinked, but still
allocated inodes, we need to hold the rgrp lock when we do the lookup,
which means that we can’t take the inode glock.

The solution is to borrow the technique already used by NFS to solve
what is essentially the same problem (given an inode number, look up
the inode carefully, checking that it really is in the expected

We cannot do that directly from the allocation code (lock ordering
again) so we give the job to the pre-existing delete workqueue and
carry on with the allocation as normal.

If we find there is no space, we do a journal flush (required anyway
if space from a deallocation is to be released) which should block
against the pending deallocations, so we should always get the space

-Len Brown issued a pull request for the intel_idle subtree, targetted at
.35-stable; the bulk of the commits are his and consist of these modifications :
intel_idle: delete substates DEBUG modparam
intel_idle: delete power_policy modparam, and choose substate functions
intel_idle: add support for Westmere-EX
intel_idle: enable Atom C6
intel_idle: simplify test for leave_mm()
intel_idle: delete bogus data from cpuidle_state.power_usage
intel_idle: add initial Sandy Bridge support
intel_idle: do not use the LAPIC timer for ATOM C2

-Len also has updates for the same area, only this time for .36-stable.
See below:
H. Peter Anvin (1):
x86, mwait: Move mwait constants to a common header file

Len Brown (4):
intel_idle: simplify test for leave_mm()
intel_idle: delete bogus data from cpuidle_state.power_usage
intel_idle: add initial Sandy Bridge support
intel_idle: do not use the LAPIC timer for ATOM C2

-Does anyone hear the bells and whistles? Well, I do, since I’m reading
Linus’ announcement about 2.6.37-rc2. Here it is:
“It’s been two weeks since -rc1, largely because one of those weeks was
the kernel summit and plumbers conference, and as a result very quiet

And in fact, despite being two weeks, the amount of changes between
-rc1 and -rc2 is pretty normal, both by number of commits (it tends to
be around 500+ commits, although we’ve had outliers: 2.6.34 was 1700,
2.6.36-rc2 was just 237, and 2.6.32-rc2 was the same as -rc1 because I
screwed up the tagging ;) and by actual size of the diff. In fact,
doing the stats, the diff looks a bit smaller than normal. Which is
all good.

And it all looks the way I like to see my -rc2’s: nothing really
interesting there. We had a SCSI calling convention change patch that
got punted due to some discussion about details (and which I’ll
probably accept for -rc3 with the fixed model).

In fact, the biggest change in the diffstat is just due to some
staging driver reverts. So hopefully this whole 2.6.37 release cycle
will be as painless as it looks so far.

Or maybe people are still just not very active after the KS. So go out and test,


-James Bottomley put up a pull request titled “SCSI queuecommand API change for 2.6.37-rc1”
Here is the announcement : “This patch set contains a single patch modifying
the SCSI queuecommand host template API to go from being called with the
host lock held to being called locklessly. The transformation is a directly equivalent
one (i.e. the locking is simply pushed into each HBA) but will form the
basis for optimising locking in the driver patch for the next merge

-Jesse Barnes issued also a pull request, this time for the PCI tree :
“This set includes a fix for a long standing mmap issue on /proc/bus/pci
(present since we added the map range checking during that e1000 EEPROM
fire drill) and a fix from Bjorn for a recent resource handling
regression. It also adds a little code to better handle device power
states at enable time; rather than assuming device state is unknown at
enable time (which will prevent us from configuring things like MSI),
we read the current state out of the device if possible, fixing some
network and gfx related bugs in the process.”

-Martin Schwidefsky is back with a rather big patchset for s390, issued
for -rc2 , Thomas Gleixner fixed the futex tree, aiming for -rc2, Ingo Molnar
stepped in with a few scheduler fixes and Rafael J. Wisocki had a pull request
for the power management sub-branch of the suspend tree.

-The latest news for this week is Benjamin Herrenschmidt’s group of powerpc
fixes; and with that, all I can say is have a pleasant weekend and take care!

Yours truly, etc.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s