kernel weekly news – 12.03.2011

Posted: March 12, 2011 in kernel

Howdy, y’all! Welcome to this week’s edition of KWN!

-Arnaldo Carvalho de Melo is again hard at work and announcing fixes for
the perf/core tree, while Sage Weil has a bunch of ceph fixes :
There are a couple of patches reverting the broken NULL d_parent
workaround, a couple patches disabling the “complete” flag behavior that
depended on it, and some bugfixes. Those include socket error handling
fixups due to a switch to WQ_NON_REENTRANT workqueue behavior, a fix for
short returns from get_user_pages, and a trivial fix to prevent the magic
.snap directory from appearing within snapshots.

I should have sent these sooner. Proper testing this time around got off
to a slow start due to the d_parent regression and confusion about Nick’s

-Takashi Iwai also is a busy bee , giving updates for the sound subsystem,
Greg Kroah-Hartman announces the release of (changelog here;a=summary ),
also Greg updated the 2.6.32 series (.32) (you have the corresponding link here:;a=summary ),
Kukjin Kim has Samsung fixes for -rc8 and …………..

-Linus announces the release of 2.6.38.-rc8 :

 I would have been ok with releasing this as the final 38, but I'm
going to be partially unreachable for several days next week, so I
felt there was no point in opening the merge window yet.

Also, we do have regressions. Some of them hopefully fixed here, but
it won't hurt to give it another week.

Not everything here is strictly a regression: the i_nlink fixes are
long-standing races (very unlikely ones, admittedly), and the alpha
updates just convert the irq chip descriptions so that we can enable
GENERIC_HARDIRQS_NO_DEPRECATED on alpha too. But most everything is
pretty simple, and the diffstat is fairly flat with nothing huge that
stands out.

40% drivers, 35% arch updates (alpha and arm), with the rest being fs,
net and some block layer updates.

So please keep testing,


-Al Viro has ” more vfs fixes”, as his e-mail is titled, and has the following body,
before the log/stat :

 Several fixes for Nick's RCU stuff; open() without O_CREAT
can be tricked into creating files, ->d_compare() in /proc/sys is trying
and failing to be RCU-safe (strictly speaking not a regression, but only
because it will be kicked out of RCU mode by ->permission() refusing
to accept IPERM_FLAG_RCU) and spurious errors in open() and vfs_path_lookup()
since we forget to set nd->inode on the LOOKUP_REVAL attempt.  More fixes
tomorrow...  Please, pull from for-linus 

-Philipp Reisner has updates for the DRBD tree (from the
webpage : “DRBD® refers to block devices designed as a building block to form high availability
(HA) clusters. This is done by mirroring a whole block device via an assigned network. 
DRBD can be understood as network based raid-1.”), which is part of mainline since 2.6.33,
John Stultz updates the RTC (actually, it’s only some cleanups, but an update nevertheless)
and Benjamin Herrenschmidt has regression fixes for powerpc.

-Arnaldo Carvalho de Melo announces perf/core fixes and various improvements, while
Ingo Molnar has some perf updates himself (/urgent), as well as a single scheduler fix and
3 x86 fixes; Mauro Carvalho Chehab has media_fixes updates and David Miller has quite a
hefty set of updates to the networking tree:

 I think all the major known regressions are cured and we should be
ready to go for 2.6.38-final

1) Fix regression in IPV6 route lookups, cures BZ 29252 and 30462

2) We use ifa_address where we mean ifa_local in ipv4 stack,
   reported by Julian Anastasov.

3) pktgen time reporting units are wrong, fix from Daniel Turull

4) IPV6=m/BRIDGE=y results in broken build, fix from Randy Dunlap.

5) Network modloading security fix needs to handle ip6 tunnel case,
   fix from Stephen Hemminger.

6) bnx2x driver fixes from Dmitry Kravkov and Eilon Greenstein.

7) smsc911x drops full sized VLAN packets erroneously, fix from Göran

8) Fix Makefile logic for entering net/ipv6 directory for the case
   where we are only building {exthdrs,addrconf}_core.o Fix from
   Thomas Graf.

9) Multi-threaded signal handling is botched because we use plain
   mutex_lock() to synchronize readers in recvmsg(), change to use
   mutex_lock_interruptible().  Fix from Rainer Weikusat.

10) Bonding driver state machine locking doesn't cover enough code,
    fix from Nils Carlson.

12) Fix BUG_ON trigger in RDS stack, fix from Neil Horman.

13) Multicase handling fixes to r6040 driver from Shawn Lin.

Please pull, thanks a lot! 

-So another week ends, this is the latest news before closing, here
they are :
-Takashi Iwai – sound fixes for 2.6.38, mostly stable-fixes.
-Ian Campbell of Citrix has a git patchset for xen:

 Changes since last time:
      * correct return value of xen_irq_from_pirq
      * WARN if a pirq cannot be allocated for a legacy IRQ
      * Updated checking comment of "xen: events: do not workaround
        too-small nr_irqs"

The following series makes a few cleanups to the Xen IRQ infrastructure.
The most important thing is that it removes the need to know about
nr_irqs and in particular the reliance on nr_irqs being static.

Apart from being generally a good thing this is needed because in 2.6.39
nr_irqs will be able to grow dynamically, specifically e7bcecb7b1d2
"genirq: Make nr_irqs runtime expandable" from tip/core/irq is targeted
at 2.6.39.

Dynamically growing nr_irqs also allows us to remove the workaround
which eats into GSI space if a dynamic IRQ cannot be allocated.

There is no ideal sequencing of this series vs e7bcecb7b1d2 (most should
have gone in before, but the penultimate patch really needed to be
simultaneous) so I haven't bothered to try and pull anything from tip
into this branch -- it should all be resolved during the merge window
and bisection won't be too broken since the "eat into GSI space"
workaround only appears to be needed on a small number of older
platforms (qemu being the main exception).

I have tested:
      * Domain 0 on real h/w and under qemu
      * PV guest, including migration and passthrough of both VF and PF.
      * PVHVM guest, including migration and passthrough of both VF and

The git pull is a branch on top of konrad/stable/irq.cleanup. However
there is an interaction with konrad/devel/xen-pciback-0.4.driver (the
addition of the domid parameter) so for convenience I have also produced
an irq-pciback branch at the same location which has
konrad/devel/xen-pciback-0.4.driver merged into this branch.

Note that this series obsoletes an older patcho f mine "xen: events:
mark cpu_evtchn_mask_p as __refdata" by virtue of removing the code in

That’s about it, enjoy your weekend!

  1. Pla says:

    Very nice summary! Why don’t you ask your feed to be added to

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s