Peter Zijlstra [Sat, 20 Sep 2008 21:38:02 +0000 (23:38 +0200)]
sched: wakeup preempt when small overlap
Lin Ming reported a 10% OLTP regression against 2.6.27-rc4.
The difference seems to come from different preemption agressiveness,
which affects the cache footprint of the workload and its effective
cache trashing.
Aggresively preempt a task if its avg overlap is very small, this should
avoid the task going to sleep and find it still running when we schedule
back to it - saving a wakeup.
Reported-by: Lin Ming <ming.m.lin@intel.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
ipv6: Fix OOPS in ip6_dst_lookup_tail().
ipsec: Restore larval states and socket policies in dump
[Bluetooth] Reject L2CAP connections on an insecure ACL link
[Bluetooth] Enforce correct authentication requirements
[Bluetooth] Fix reference counting during ACL config stage
Neil Horman [Tue, 9 Sep 2008 20:51:35 +0000 (13:51 -0700)]
ipv6: Fix OOPS in ip6_dst_lookup_tail().
This fixes kernel bugzilla 11469: "TUN with 1024 neighbours:
ip6_dst_lookup_tail NULL crash"
dst->neighbour is not necessarily hooked up at this point
in the processing path, so blindly dereferencing it is
the wrong thing to do. This NULL check exists in other
similar paths and this case was just an oversight.
Also fix the completely wrong and confusing indentation
here while we're at it.
Based upon a patch by Evgeniy Polyakov.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Merge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
clockevents: remove WARN_ON which was used to gather information
Thomas Gleixner [Tue, 9 Sep 2008 19:38:57 +0000 (21:38 +0200)]
clockevents: remove WARN_ON which was used to gather information
The issue of the endless reprogramming loop due to a too small
min_delta_ns was fixed with the previous updates of the clock events
code, but we had no information about the spread of this problem. I
added a WARN_ON to get automated information via kerneloops.org and to
get some direct reports, which allowed me to analyse the affected
machines.
The WARN_ON has served its purpose and would be annoying for a release
kernel. Remove it and just keep the information about the increase of
the min_delta_ns value.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This is happening because of an improper usage of strcmp() in the
e820 parsing code. The strcmp() always returns !0 and never resets the
value for e820.nr_map and returns an incorrect user-defined map.
Merge branch 'linux-next' of git://git.infradead.org/~dedekind/ubifs-2.6
* 'linux-next' of git://git.infradead.org/~dedekind/ubifs-2.6:
UBIFS: make minimum fanout 3
UBIFS: fix division by zero
UBIFS: amend f_fsid
UBIFS: fill f_fsid
UBIFS: improve statfs reporting even more
UBIFS: introduce LEB overhead
UBIFS: add forgotten gc_idx_lebs component
UBIFS: fix assertion
UBIFS: improve statfs reporting
UBIFS: remove incorrect index space check
UBIFS: push empty flash hack down
UBIFS: do not update min_idx_lebs in stafs
UBIFS: allow for racing between GC and TNC
UBIFS: always read hashed-key nodes under TNC mutex
UBIFS: fix zero-length truncations
James Bottomley [Thu, 4 Sep 2008 01:43:36 +0000 (20:43 -0500)]
lib: Correct printk %pF to work on all architectures
It was introduced by "vsprintf: add support for '%pS' and '%pF' pointer
formats" in commit 0fe1ef24f7bd0020f29ffe287dfdb9ead33ca0b2. However,
the current way its coded doesn't work on parisc64. For two reasons: 1)
parisc isn't in the #ifdef and 2) parisc has a different format for
function descriptors
Make dereference_function_descriptor() more accommodating by allowing
architecture overrides. I put the three overrides (for parisc64, ppc64
and ia64) in arch/kernel/module.c because that's where the kernel
internal linker which knows how to deal with function descriptors sits.
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Acked-by: Tony Luck <tony.luck@intel.com> Acked-by: Kyle McMartin <kyle@mcmartin.ca> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Chris Snook [Tue, 9 Sep 2008 07:26:57 +0000 (03:26 -0400)]
MAINTAINERS: add Atheros maintainer for atlx
Jie Yang at Atheros is getting more directly involved with upstream work
on the atl* drivers. This patch changes the ATL1 entry to ATLX (atl2
support posted to netdev today) and adds him as a maintainer.
update Documentation/filesystems/Locking for 2.6.27 changes
In the 2.6.27 circle ->fasync lost the BKL, and the last remaining
->open variant that takes the BKL is also gone. ->get_sb and ->kill_sb
didn't have BKL forever, so updated the entries while we're at that.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Herbert Xu [Tue, 9 Sep 2008 12:23:37 +0000 (05:23 -0700)]
ipsec: Restore larval states and socket policies in dump
The commit commit 4c563f7669c10a12354b72b518c2287ffc6ebfb3 ("[XFRM]:
Speed up xfrm_policy and xfrm_state walking") inadvertently removed
larval states and socket policies from netlink dumps. This patch
restores them.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
[S390] cio: allow offline processing for disconnected devices
When disconnected ccw devices are removed, the device has to be set
offline, otherwise there will be side effects including a reference
count imbalance. This patch modifies ccw_device_offline to work for
devices in disconnecte/not operational state. ccw_device_offline is
called by cio for devices which are online during device removal.
Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
ssch() has two classes of return codes:
- condition codes (0-3) which need to be translated to Linux
error codes
- Linux error codes (-EIO on exceptions) which should be passed
to the caller (instead of erronously being handled like
condition code 3)
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Jarod Wilson [Tue, 9 Sep 2008 10:38:56 +0000 (12:38 +0200)]
[S390] CVE-2008-1514: prevent ptrace padding area read/write in 31-bit mode
When running a 31-bit ptrace, on either an s390 or s390x kernel,
reads and writes into a padding area in struct user_regs_struct32
will result in a kernel panic.
This is also known as CVE-2008-1514.
Test case available here:
http://sources.redhat.com/cgi-bin/cvsweb.cgi/~checkout~/tests/ptrace-tests/tests/user-area-padding.c?cvsroot=systemtap
Steps to reproduce:
1) wget the above
2) gcc -o user-area-padding-31bit user-area-padding.c -Wall -ggdb2 -D_GNU_SOURCE -m31
3) ./user-area-padding-31bit
<panic>
Test status
-----------
Without patch, both s390 and s390x kernels panic. With patch, the test case,
as well as the gdb testsuite, pass without incident, padding area reads
returning zero, writes ignored.
Nb: original version returned -EINVAL on write attempts, which broke the
gdb test and made the test case slightly unhappy, Jan Kratochvil suggested
the change to return 0 on write attempts.
Signed-off-by: Jarod Wilson <jarod@redhat.com> Tested-by: Jan Kratochvil <jan.kratochvil@redhat.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
[Bluetooth] Reject L2CAP connections on an insecure ACL link
The Security Mode 4 of the Bluetooth 2.1 specification has strict
authentication and encryption requirements. It is the initiators job
to create a secure ACL link. However in case of malicious devices, the
acceptor has to make sure that the ACL is encrypted before allowing
any kind of L2CAP connection. The only exception here is the PSM 1 for
the service discovery protocol, because that is allowed to run on an
insecure ACL link.
Previously it was enough to reject a L2CAP connection during the
connection setup phase, but with Bluetooth 2.1 it is forbidden to
do any L2CAP protocol exchange on an insecure link (except SDP).
The new hci_conn_check_link_mode() function can be used to check the
integrity of an ACL link. This functions also takes care of the cases
where Security Mode 4 is disabled or one of the devices is based on
an older specification.
With the introduction of Security Mode 4 and Simple Pairing from the
Bluetooth 2.1 specification it became mandatory that the initiator
requires authentication and encryption before any L2CAP channel can
be established. The only exception here is PSM 1 for the service
discovery protocol (SDP). It is meant to be used without any encryption
since it contains only public information. This is how Bluetooth 2.0
and before handle connections on PSM 1.
For Bluetooth 2.1 devices the pairing procedure differentiates between
no bonding, general bonding and dedicated bonding. The L2CAP layer
wrongly uses always general bonding when creating new connections, but it
should not do this for SDP connections. In this case the authentication
requirement should be no bonding and the just-works model should be used,
but in case of non-SDP connection it is required to use general bonding.
If the new connection requires man-in-the-middle (MITM) protection, it
also first wrongly creates an unauthenticated link key and then later on
requests an upgrade to an authenticated link key to provide full MITM
protection. With Simple Pairing the link key generation is an expensive
operation (compared to Bluetooth 2.0 and before) and doing this twice
during a connection setup causes a noticeable delay when establishing
a new connection. This should be avoided to not regress from the expected
Bluetooth 2.0 connection times. The authentication requirements are known
up-front and so enforce them.
To fulfill these requirements the hci_connect() function has been extended
with an authentication requirement parameter that will be stored inside
the connection information and can be retrieved by userspace at any
time. This allows the correct IO capabilities exchange and results in
the expected behavior.
[Bluetooth] Fix reference counting during ACL config stage
The ACL config stage keeps holding a reference count on incoming
connections when requesting the extended features. This results in
keeping an ACL link up without any users. The problem here is that
the Bluetooth specification doesn't define an ownership of the ACL
link and thus it can happen that the implementation on the initiator
side doesn't care about disconnecting unused links. In this case the
acceptor needs to take care of this.
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/hskinnemoen/avr32-2.6
* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/hskinnemoen/avr32-2.6:
avr32: pm_standby low-power ram bug fix
avr32: Fix lockup after Java stack underflow in user mode
Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
powerpc: Fix rare boot build breakage
powerpc/spufs: Fix possible scheduling of a context to multiple SPEs
powerpc/spufs: Fix race for a free SPU
powerpc/spufs: Fix multiple get_spu_context()
Merge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
sched: arch_reinit_sched_domains() must destroy domains to force rebuild
sched, cpuset: rework sched domains and CPU hotplug handling (v4)
Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
ahci: RAID mode SATA patch for Intel Ibex Peak DeviceIDs
pata_sil680: remove duplicate pcim_enable_device
libata-sff: kill spurious WARN_ON() in ata_hsm_move()
sata_nv: disable hardreset for generic
ahci: disable PMP for marvell ahcis
sata_mv: add RocketRaid 1720 PCI ID to driver
ahci, pata_marvell: play nicely together
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
bridge: don't allow setting hello time to zero
netns : fix kernel panic in timewait socket destruction
pkt_sched: Fix qdisc state in net_tx_action()
netfilter: nf_conntrack_irc: make sure string is terminated before calling simple_strtoul
netfilter: nf_conntrack_gre: nf_ct_gre_keymap_flush() fixlet
netfilter: nf_conntrack_gre: more locking around keymap list
netfilter: nf_conntrack_sip: de-static helper pointers
Jason Wessel [Mon, 8 Sep 2008 13:53:37 +0000 (14:53 +0100)]
usb: fix null deferences in low level usb serial
The hw interface drivers for the usb serial devices deference the tty
structure to set up the parameters for the initial console. The tty
structure should be passed as a parameter to the set_termios() call.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Chuck Lever [Mon, 8 Sep 2008 15:58:13 +0000 (11:58 -0400)]
NFS: Restore missing hunk in NFS mount option parser
Automounter maps can contain mount options valid for other NFS
implementations but not for Linux. The Linux automounter uses the
mount command's "-s" command line option ("s" for "sloppy") so that
mount requests containing such options are not rejected.
Commit f45663ce5fb30f76a3414ab3ac69f4dd320e760a attempted to address a
known regression with text-based NFS mount option parsing. Unrecognized
mount options would cause mount requests to fail, even if the "-s"
option was used on the mount command line.
Unfortunately, this commit was not complete as submitted. It adds a
new mount option, "sloppy". But it is missing a hunk, so it now allows
NFS mounts with unrecognized mount options, even if the "sloppy" option
is not present. This could be a problem if a required critical mount
option such as "sync" is misspelled, for example, and is considered a
regression from 2.6.26.
This patch restores the missing hunk. Now, the default behavior of
text-based NFS mount options is as before: any unrecognized mount option
will cause the mount to fail.
Please include this in 2.6.27-rc.
Thanks to Neil Brown for reporting this.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Acked-by: J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
and the culprit cause seems to be starting the bridge interface.
In particular, when starting the bridge interface, his scripts
are specifying a hello timer interval of "0".
The bridge hello time can't be safely set to values less than 1
second, otherwise it is possible to end up with a runaway timer.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Lezcano [Mon, 8 Sep 2008 20:17:27 +0000 (13:17 -0700)]
netns : fix kernel panic in timewait socket destruction
How to reproduce ?
- create a network namespace
- use tcp protocol and get timewait socket
- exit the network namespace
- after a moment (when the timewait socket is destroyed), the kernel
panics.
This patch provides a function to purge all timewait sockets related
to a network namespace. The timewait sockets life cycle is not tied with
the network namespace, that means the timewait sockets stay alive while
the network namespace dies. The timewait sockets are for avoiding to
receive a duplicate packet from the network, if the network namespace is
freed, the network stack is removed, so no chance to receive any packets
from the outside world. Furthermore, having a pending destruction timer
on these sockets with a network namespace freed is not safe and will lead
to an oops if the timer callback which try to access data belonging to
the namespace like for example in:
inet_twdr_do_twkill_work
-> NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITED);
Purging the timewait sockets at the network namespace destruction will:
1) speed up memory freeing for the namespace
2) fix kernel panic on asynchronous timewait destruction
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Acked-by: Denis V. Lunev <den@openvz.org> Acked-by: Eric W. Biederman <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net>
On 32-bit, at least the generic nops are fairly reasonable, but the
default nops for 64-bit really look pretty sad, and the P6 nops really do
look better.
So I would suggest perhaps moving the static P6 nop selection into the
CONFIG_X86_64 thing.
The alternative is to just get rid of that static nop selection, and just
have two cases: 32-bit and 64-bit, and just pick obviously safe cases for
them.
Tejun Heo [Fri, 29 Aug 2008 14:27:43 +0000 (16:27 +0200)]
libata-sff: kill spurious WARN_ON() in ata_hsm_move()
On HSM_ST_ERR, ata_hsm_move() triggers WARN_ON() if AC_ERR_DEV or
AC_ERR_HSM is not set. PHY events may trigger HSM_ST_ERR with other
error codes and, with or without it, there just isn't much reason to
do WARN_ON() on it. Even if error code is not set there, core EH
logic won't have any problem dealing with the error condition.
OSDL bz#11065 reports this problem.
Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Tejun Heo [Fri, 29 Aug 2008 14:13:12 +0000 (16:13 +0200)]
sata_nv: disable hardreset for generic
of them being unifying probing, hotplug and EH reset paths uniform.
Previously, broken hardreset could go unnoticed as it wasn't used
during probing but when something goes wrong or after hotplug the
problem will surface and bite hard.
OSDL bug 11195 reports that sata_nv generic flavor falls into this
category. Hardreset itself succeeds but PHY stays offline after
hardreset. I tried longer debounce timing but the result was the
same.
http://bugzilla.kernel.org/show_bug.cgi?id=11195
So, it seems we'll have to drop hardreset from the generic flavor.
Alan Cox [Wed, 3 Sep 2008 13:48:34 +0000 (14:48 +0100)]
ahci, pata_marvell: play nicely together
I've been chasing Jeff about this for months. Jeff added the Marvell
device identifiers to the ahci driver without making the AHCI driver
handle the PATA port. This means a lot of users can't use current
kernels and in most distro cases can't even install.
This has been going on since March 2008 for the 6121 Marvell, and late 2007
for the 6145!!!
This was all pointed out at the time and repeatedly ignored. Bugs assigned
to Jeff about this are ignored also.
To quote Jeff in email
> "Just switch the order of 'ahci' and 'pata_marvell' in
> /etc/modprobe.conf, then use Fedora's tools regenerate the initrd.
> See? It's not rocket science, and the current configuration can be
> easily made to work for Fedora users."
(Which isn't trivial, isn't end user, shouldn't be needed, and as it usually
breaks at install time is in fact impossible)
To quote Jeff in August 2007
> " mv-ahci-pata
> Marvell 6121/6141 PATA support. Needs fixing in the 'PATA controller
> command' area before it is usable, and can go upstream."
Only he add the ids anyway later and caused regressions, adding a further
id in March causing more regresions.
The actual fix for the moment is very simple. If the user has included
the pata_marvell driver let it drive the ports. If they've only selected
for SATA support give them the AHCI driver which will run the port a fraction
faster. Allow the user to control this decision via ahci.marvell_enable as
a module parameter so that distributions can ship 'it works' defaults and
smarter users (or config tools) can then flip it over it desired.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
A make -j20 powerpc kernel build broke a couple of months ago saying:
In file included from arch/powerpc/boot/gunzip_util.h:13,
from arch/powerpc/boot/prpmc2800.c:21:
arch/powerpc/boot/zlib.h:85: error: expected ‘:’, ‘,’, ‘;’, ‘}’ or ‘__attribute__’ before ‘*’ token
arch/powerpc/boot/zlib.h:630: warning: type defaults to ‘int’ in declaration of ‘Byte’
arch/powerpc/boot/zlib.h:630: error: expected ‘;’, ‘,’ or ‘)’ before ‘*’ token
It happened again yesterday: too rare for me to confirm the fix, but
it looks like the list of dependants on gunzip_util.h was incomplete.
Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
Herbert Xu [Mon, 8 Sep 2008 04:29:54 +0000 (14:29 +1000)]
Revert "crypto: camellia - Use kernel-provided bitops, unaligned access helpers"
This reverts commit bd699f2df6dbc2f4cba528fe598bd63a4d3702c5,
which causes camellia to fail the included self-test vectors.
It has also been confirmed that it breaks existing encrypted
disks using camellia.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
net_tx_action() can skip __QDISC_STATE_SCHED bit clearing while qdisc
is neither ran nor rescheduled, which may cause endless loop in
dev_deactivate().
Reported-by: Denys Fedoryshchenko <denys@visp.net.lb> Tested-by: Denys Fedoryshchenko <denys@visp.net.lb> Signed-off-by: Jarek Poplawski <jarkao2@gmail.com> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Mon, 8 Sep 2008 01:21:24 +0000 (18:21 -0700)]
netfilter: nf_conntrack_irc: make sure string is terminated before calling simple_strtoul
Alexey Dobriyan points out:
1. simple_strtoul() silently accepts all characters for given base even
if result won't fit into unsigned long. This is amazing stupidity in
itself, but
2. nf_conntrack_irc helper use simple_strtoul() for DCC request parsing.
Data first copied into 64KB buffer, so theoretically nothing prevents
reading past the end of it, since data comes from network given 1).
This is not actually a problem currently since we're guaranteed to have
a 0 byte in skb_shared_info or in the buffer the data is copied to, but
to make this more robust, make sure the string is actually terminated.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
It does "kfree(list_head)" which looks wrong because entity that was
allocated is definitely not list_head.
However, this all works because list_head is first item in
struct nf_ct_gre_keymap.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
netfilter: nf_conntrack_gre: more locking around keymap list
gre_keymap_list should be protected in all places.
(unless I'm misreading something)
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Helper's ->help hook can run concurrently with itself, so iterating over
SIP helpers with static pointer won't work reliably.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Andre Detsch [Thu, 4 Sep 2008 21:16:27 +0000 (21:16 +0000)]
powerpc/spufs: Fix possible scheduling of a context to multiple SPEs
We currently have a race when scheduling a context to a SPE -
after we have found a runnable context in spusched_tick, the same
context may have been scheduled by spu_activate().
This may result in a panic if we try to unschedule a context that has
been freed in the meantime.
This change exits spu_schedule() if the context has already been
scheduled, so we don't end up scheduling it twice.
Signed-off-by: Andre Detsch <adetsch@br.ibm.com> Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: cpu_init(): fix memory leak when using CPU hotplug
x86: pda_init(): fix memory leak when using CPU hotplug
x86, xen: Use native_pte_flags instead of native_pte_val for .pte_flags
x86: move mtrr cpu cap setting early in early_init_xxxx
x86: delay early cpu initialization until cpuid is done
x86: use X86_FEATURE_NOPL in alternatives
x86: add NOPL as a synthetic CPU feature bit
x86: boot: stub out unimplemented CPU feature words
x86: cpu_init(): fix memory leak when using CPU hotplug
Exception stacks are allocated each time a CPU is set online.
But the allocated space is never freed. Thus with one CPU hotplug
offline/online cycle there is a memory leak of 24K (6 pages) for
a CPU.
Fix is to allocate exception stacks only once -- when the CPU is
set online for the first time.
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: akpm@linux-foundation.org Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86: pda_init(): fix memory leak when using CPU hotplug
pda->irqstackptr is allocated whenever a CPU is set online.
But it is never freed. This results in a memory leak of 16K
for each CPU offline/online cycle.
Fix is to allocate pda->irqstackptr only once.
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com> Cc: akpm@linux-foundation.org Signed-off-by: Ingo Molnar <mingo@elte.hu>
Max Krasnyansky [Fri, 29 Aug 2008 20:11:41 +0000 (13:11 -0700)]
sched: arch_reinit_sched_domains() must destroy domains to force rebuild
What I realized recently is that calling rebuild_sched_domains() in
arch_reinit_sched_domains() by itself is not enough when cpusets are enabled.
partition_sched_domains() code is trying to avoid unnecessary domain rebuilds
and will not actually rebuild anything if new domain masks match the old ones.
What this means is that doing
echo 1 > /sys/devices/system/cpu/sched_mc_power_savings
on a system with cpusets enabled will not take affect untill something changes
in the cpuset setup (ie new sets created or deleted).
This patch fixes restore correct behaviour where domains must be rebuilt in
order to enable MC powersaving flags.
Test on quad-core Core2 box with both CONFIG_CPUSETS and !CONFIG_CPUSETS.
Also tested on dual-core Core2 laptop. Lockdep is happy and things are working
as expected.
Manfred Spraul [Sun, 31 Aug 2008 17:58:49 +0000 (19:58 +0200)]
kernel/cpu.c: Move the CPU_DYING notifiers
When a cpu is taken offline, the CPU_DYING notifiers are called on the
dying cpu. According to <linux/notifiers.h>, the cpu should be "not
running any task, not handling interrupts, soon dead".
For the current implementation, this is not true:
- __cpu_disable can fail. If it fails, then the cpu will remain alive
and happy.
- At least on x86, __cpu_disable() briefly enables the local interrupts
to handle any outstanding interrupts.
What about moving CPU_DYING down a few lines, behind the __cpu_disable()
line?
There are only two CPU_DYING handlers in the kernel right now: one in
kvm, one in the scheduler. Both should work with the patch applied
[and: I'm not sure if either one handles a failing __cpu_disable()]
The patch survives simple offlining a cpu. kvm untested due to lack
of a test setup.
Yinghai Lu [Thu, 4 Sep 2008 19:09:43 +0000 (21:09 +0200)]
x86: move mtrr cpu cap setting early in early_init_xxxx
Krzysztof Helt found MTRR is not detected on k6-2
root cause:
we moved mtrr_bp_init() early for mtrr trimming,
and in early_detect we only read the CPU capability from cpuid,
so some cpu doesn't have that bit in cpuid.
So we need to add early_init_xxxx to preset those bit before mtrr_bp_init
for those earlier cpus.
this patch is for v2.6.27
Reported-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
sched: fix __load_balance_iterator() for cfq with only one task
The __load_balance_iterator() returns a NULL when there's only one
sched_entity which is a task. It is caused by the following code-path.
/* Skip over entities that are not tasks */
do {
se = list_entry(next, struct sched_entity, group_node);
next = next->next;
} while (next != &cfs_rq->tasks && !entity_is_task(se));
if (next == &cfs_rq->tasks)
return NULL;
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This will return NULL even when se is a task.
As a side-effect, there was a regression in sched_mc behavior since 2.6.25,
since iter_move_one_task() when it calls load_balance_start_fair(),
would not get any tasks to move!
Fix this by checking if the last entity was a task or not.
Signed-off-by: Gautham R Shenoy <ego@in.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
The current check for monotonicity is way too weak: Andreas Mohr reports (
http://lkml.org/lkml/2008/8/10/77 ) that on one of his test systems the
current check only triggers in 50% of all cases, leading to catastrophic
timer behaviour. To fix this issue, expand the check for monotonicity by
doing ten consecutive tests instead of one.
clocksource, acpi_pm.c: use proper read function also in errata mode
On all hardware (some Intel ICH4, PIIX4 and PIIX4E chipsets) affected by a
hardware errata there's about a 4.2% chance that initialization of the
ACPI PMTMR fails. On those chipsets, we need to read out the timer value
at least three times to get a correct result, for every once in a while
(i.e. within a 3 ns window every 69.8 ns) the read returns a bogus
result. During normal operation we work around this issue, but during
initialization reading a bogus value may lead to -EINVAL even though the
hardware is usable.
ntp: fix calculation of the next jiffie to trigger RTC sync
We have a bug in the calculation of the next jiffie to trigger the RTC
synchronisation. The aim here is to run sync_cmos_clock() as close as
possible to the middle of a second. Which means we want this function to
be called less than or equal to half a jiffie away from when now.tv_nsec
equals 5e8 (500000000).
If this is not the case for a given call to the function, for this purpose
instead of updating the RTC we calculate the offset in nanoseconds to the
next point in time where now.tv_nsec will be equal 5e8. The calculated
offset is then converted to jiffies as these are the unit used by the
timer.
Hovewer timespec_to_jiffies() used here uses a ceil()-type rounding mode,
where the resulting value is rounded up. As a result the range of
now.tv_nsec when the timer will trigger is from 5e8 to 5e8 + TICK_NSEC
rather than the desired 5e8 - TICK_NSEC / 2 to 5e8 + TICK_NSEC / 2.
As a result if for example sync_cmos_clock() happens to be called at the
time when now.tv_nsec is between 5e8 + TICK_NSEC / 2 and 5e8 to 5e8 +
TICK_NSEC, it will simply be rescheduled HZ jiffies later, falling in the
same range of now.tv_nsec again. Similarly for cases offsetted by an
integer multiple of TICK_NSEC.
This change addresses the problem by subtracting TICK_NSEC / 2 from the
nanosecond offset to the next point in time where now.tv_nsec will be
equal 5e8, effectively shifting the following rounding in
timespec_to_jiffies() so that it produces a rounded-to-nearest result.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
This patch provides an ARM implementation of ioremap_wc().
We use different page table attributes depending on which CPU we
are running on:
- Non-XScale ARMv5 and earlier systems: The ARMv5 ARM documents four
possible mapping types (CB=00/01/10/11). We can't use any of the
cached memory types (CB=10/11), since that breaks coherency with
peripheral devices. Both CB=00 and CB=01 are suitable for _wc, and
CB=01 (Uncached/Buffered) allows the hardware more freedom than
CB=00, so we'll use that.
(The ARMv5 ARM seems to suggest that CB=01 is allowed to delay stores
but isn't allowed to merge them, but there is no other mapping type
we can use that allows the hardware to delay and merge stores, so
we'll go with CB=01.)
- XScale v1/v2 (ARMv5): same as the ARMv5 case above, with the slight
difference that on these platforms, CB=01 actually _does_ allow
merging stores. (If you want noncoalescing bufferable behavior
on Xscale v1/v2, you need to use XCB=101.)
- Xscale v3 (ARMv5) and ARMv6+: on these systems, we use TEXCB=00100
mappings (Inner/Outer Uncacheable in xsc3 parlance, Uncached Normal
in ARMv6 parlance).
The ARMv6 ARM explicitly says that any accesses to Normal memory can
be merged, which makes Normal memory more suitable for _wc mappings
than Device or Strongly Ordered memory, as the latter two mapping
types are guaranteed to maintain transaction number, size and order.
We use the Uncached variety of Normal mappings for the same reason
that we can't use C=1 mappings on ARMv5.
The xsc3 Architecture Specification documents TEXCB=00100 as being
Uncacheable and allowing coalescing of writes, which is also just
what we need.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
CONFIG_AC97_BUS is used from both sound and ucb1400 drivers.
The recent change in Kconfig introduced the exclusive dependency on
CONFIG_SOUND, and disabled the ucb1400 build without sound.
This patch makes CONFIG_AC97_BUS independent.
Signed-off-by: Takashi Iwai <tiwai@suse.de> Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
Thomas Gleixner [Sat, 6 Sep 2008 01:06:08 +0000 (03:06 +0200)]
x86: HPET: read back compare register before reading counter
After fixing the u32 thinko I sill had occasional hickups on ATI chipsets
with small deltas. There seems to be a delay between writing the compare
register and the transffer to the internal register which triggers the
interrupt. Reading back the value makes sure, that it hit the internal
match register befor we compare against the counter value.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Thomas Gleixner [Sat, 6 Sep 2008 01:01:45 +0000 (03:01 +0200)]
clockevents: broadcast fixup possible waiters
Until the C1E patches arrived there where no users of periodic broadcast
before switching to oneshot mode. Now we need to trigger a possible
waiter for a periodic broadcast when switching to oneshot mode.
Otherwise we can starve them for ever.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
H. Peter Anvin [Tue, 19 Aug 2008 00:39:32 +0000 (17:39 -0700)]
x86: add NOPL as a synthetic CPU feature bit
The long noops ("NOPL") are supposed to be detected by family >= 6.
Unfortunately, several non-Intel x86 implementations, both hardware
and software, don't obey this dictum. Instead, probe for NOPL
directly by executing a NOPL instruction and see if we get #UD.
H. Peter Anvin [Fri, 15 Aug 2008 22:36:31 +0000 (15:36 -0700)]
x86: boot: stub out unimplemented CPU feature words
The CPU feature detection code in the boot code is somewhat minimal,
and doesn't include all possible CPUID words. In particular, it
doesn't contain the code for CPU feature words 2 (Transmeta),
3 (Linux-specific), 5 (VIA), or 7 (scattered). Zero them out, so we
can still set those bits as known at compile time; in particular, this
allows creating a Linux-specific NOPL flag and have it required (and
therefore resolvable at compile time) in 64-bit mode.
Andrew Morton [Fri, 5 Sep 2008 21:00:24 +0000 (14:00 -0700)]
drivers/mmc/card/block.c: fix refcount leak in mmc_block_open()
mmc_block_open() increments md->usage although it returns with -EROFS when
default mounting a MMC/SD card with write protect switch on. This
reference counting bug prevents /dev/mmcblkX from being released on card
removal, and situation worsen with reinsertion until the minor number
range runs out.
Reported-by: <sasin@solomon-systech.com> Acked-by: Pierre Ossman <drzeus-list@drzeus.cx> Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
atmel_lcdfb: fix oops in rmmod when framebuffer fails to register
If framebuffer registration failed in platform driver ->probe() callback,
dev_get_drvdata() points to freed memory region, but ->remove() function
try to use it and the following oops occurs:
Fix regression tracked as http://bugzilla.kernel.org/show_bug.cgi?id=11361
and caused by commit f735a2a1a4f2a0f5cd823ce323e82675990469e2 ("[netdrvr]
forcedeth: setup wake-on-lan before shutting down") that makes network
adapters integrated into the NVidia MCP55 chipsets fail to work in kexeced
kernels. The problem appears to be that if the adapter is put into D3_hot
during ->shutdown(), it cannot be brought back into D0 after kexec (ref.
http://marc.info/?l=linux-kernel&m=121900062814967&w=4). Therefore, only
put forcedeth into D3 during ->shutdown() if the system is to be powered
off.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Tested-by: Yinghai Lu <yhlu.kernel@gmail.com> Cc: Ayaz Abdulla <aabdulla@nvidia.com> Acked-by: Jeff Garzik <jeff@garzik.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Li Zefan [Fri, 5 Sep 2008 21:00:18 +0000 (14:00 -0700)]
res_counter: fix off-by-one bug in setting limit
I found we can no longer set limit to 0 with 2.6.27-rcX:
# mount -t cgroup -omemory xxx /mnt
# mkdir /mnt/0
# echo 0 > /mnt/0/memory.limit_in_bytes
bash: echo: write error: Device or resource busy
It turned out 'limit' can't be set to 'usage', which is wrong IMO.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com> Acked-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
sched: fix process time monotonicity
sched_clock: fix NOHZ interaction
* git://git.infradead.org/~dwmw2/dwmw2-2.6.27:
Revert "[ARM] use the new byteorder headers"
Fix conditional export of kvh.h and a.out.h to userspace.
[MTD] [NAND] tmio_nand: fix base address programming
Merge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb
* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb: (98 commits)
V4L/DVB (8881): gspca: After 'while (retry--) {...}', retry will be -1 but not 0.
V4L/DVB (8880): PATCH: Fix parents on some webcam drivers
V4L/DVB (8877): b2c2 and bt8xx: udelay to mdelay
V4L/DVB (8876): budget: udelay changed to mdelay
V4L/DVB (8874): gspca: Adjust hstart for sn9c103/ov7630 and update usb-id's.
V4L/DVB (8873): gspca: Bad image offset with rev012a of spca561 and adjust exposure.
V4L/DVB (8872): gspca: Bad image format and offset with rev072a of spca561.
V4L/DVB (8870): gspca: Fix dark room problem with sonixb.
V4L/DVB (8869): gspca: Move the Sonix webcams with TAS5110C1B from sn9c102 to gspca.
V4L/DVB (8868): gspca: Support for vga modes with sif sensors in sonixb.
V4L/DVB (8844): dabusb_fpga_download(): fix a memory leak
V4L/DVB (8843): tda10048_firmware_upload(): fix a memory leak
V4L/DVB (8842): vivi_release(): fix use-after-free
V4L/DVB (8840): dib0700: add basic support for Hauppauge Nova-TD-500 (84xxx)
V4L/DVB (8839): dib0700: add comment to identify 35th USB id pair
V4L/DVB (8837): dvb: fix I2C adapters name size
V4L/DVB (8835): gspca: Same pixfmt as the sn9c102 driver and raw Bayer added in sonixb.
V4L/DVB (8834): gspca: Have a bigger buffer for sn9c10x compressed images.
V4L/DVB (8833): gspca: Cleanup the sonixb code.
V4L/DVB (8832): gspca: Bad pixelformat of vc0321 webcams.
...
Merge branch 'release-2.6.27' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-acpi-2.6
* 'release-2.6.27' of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-acpi-2.6:
acer-wmi: remove debugfs entries upon unloading
ACPI: Avoid bogus timeout about SMbus check
fujitsu-laptop: fix regression for P8010 in 2.6.27-rc
ACPI: Make Len Brown the ACPI maintainer again
ACPI: thinkpad-acpi: wan radio control is not experimental
PNPACPI: ignore the producer/consumer bit for extended IRQ descriptors
acpi: add checking for NULL early param
ACPI: Fix typo in "Disable MWAIT via DMI on broken Compal board"
ACPI: Fix now signed module parameter.
ACPI: Change package length error to warning
ACPI: Fix now signed module parameter.
Atsushi Nemoto [Tue, 26 Aug 2008 13:34:57 +0000 (22:34 +0900)]
[MIPS] Probe initrd header only if explicitly specified
Currently init_initrd() probes initrd header at the last page of kernel
image, but it is valid only if addinitrd was used. If addinitrd was not
used, the area contains garbage so probing there might misdetect initrd
header (magic number is not strictly robust).
This patch introduces CONFIG_PROBE_INITRD_HEADER to explicitly enable this
probing.
Commmit 59e39ecd933ba49eb6efe84cbfa5597a6c9ef18a ("Fix WARNING: at
kernel/smp.c:290") introduced local_flush_icache_range but lacks
initialization for some TX39 case.