If (start + size) is not cacheline aligned and (start & mask) > (end &
mask), the last but one cacheline won't be invalidated as it should.
Fix this by rounding `end' down to the nearest cacheline boundary if
it gets adjusted due to misalignment.
Also flush the write buffer unconditionally -- if the dcache wrote
back a line just before we invalidated it, the dirty data may be
sitting in the write buffer waiting to corrupt our buffer later.
Andi Kleen [Sat, 23 Jun 2007 00:29:23 +0000 (02:29 +0200)]
i386: Make CMPXCHG64 only dependent on PAE
It is only used for PAE kernels in set_64bit.
The problem is that due to a old Windows bug many CPUs need magic MSRs
to enable CMPXCHG64, and we can't do that nicely early enough before
it is potentially used.
But since we only need it in PAE kernels so only force the checking
for CMPXCHG65 with PAE.
Linus Torvalds [Fri, 22 Jun 2007 18:11:33 +0000 (11:11 -0700)]
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney/blackfin-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
Blackfin arch: add proper const volatile to addr argument to the read functions
Blackfin arch: Add definition of dma_mapping_error
Blackfin arch: move cond_syscall() behind __KERNEL__ like all other architectures
Blackfin arch: match kernel startup messaage with new linker script
Blackfin arch: add missing braces around array bfin serial init
Blackfin arch: update printk to use KERN_EMERG and reformat crash output
Blackfin arch: update ANOMALY handling
Linus Torvalds [Fri, 22 Jun 2007 18:10:34 +0000 (11:10 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
IB/mlx4: Correct max_srq_wr returned from mlx4_ib_query_device()
IPoIB/cm: Remove dead definition of struct ipoib_cm_id
IPoIB/cm: Fix interoperability when MTU doesn't match
IPoIB/cm: Initialize RX before moving QP to RTR
IB/umem: Fix possible hang on process exit
Linus Torvalds [Fri, 22 Jun 2007 18:10:05 +0000 (11:10 -0700)]
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
[AF_RXRPC]: Return the number of bytes buffered in rxrpc_send_data()
[IPVS]: Fix state variable on failure to start ipvs threads
[XFRM]: Fix MTU calculation for non-ESP SAs
Arjan van de Ven [Thu, 21 Jun 2007 05:23:21 +0000 (22:23 -0700)]
Allow DEBUG_RODATA and KPROBES to co-exist
Do not mark the kernel text read only if KPROBES is in the kernel;
kprobes needs to hot-patch the kernel text to insert it's
instrumentation.
In this case, only mark the .rodata segment as read only.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Tested-by: S. P. Prasanna <prasanna@in.ibm.com> Cc: Andi Kleen <ak@suse.de> Cc: William Cohen <wcohen@redhat.com> Cc: Ian McDonald <ian.mcdonald@jandi.co.nz> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Thomas Gleixner [Thu, 21 Jun 2007 20:45:15 +0000 (20:45 +0000)]
posix-timers: Prevent softirq starvation by small intervals and SIG_IGN
posix-timers which deliver an ignored signal are currently rearmed in
the timer softirq: This is necessary because the timer needs to be
delivered again when SIG_IGN is removed. This is not a problem, when
the interval is reasonable.
With high resolution timers enabled one might arm a posix timer with a
very small interval and ignore the signal. This might lead to a
softirq starvation when the interval is so small that the timer is
requeued onto the softirq pending list right away.
This problem was pointed out by Jan Kiszka. Thanks Jan !
The correct solution would be to stop the timer, when the signal is
ignored and rearm it when SIG_IGN is removed. Unfortunately this
requires modification in sigaction and involves non trivial sighand
locking. It's too late in the release cycle for such a change.
For now we just keep the timer running and enforce that the timer only
fires every jiffie. This does not break anything as we keep the
overrun counter correct. It adds a little inaccuracy to the
timer_gettime() interface, but...
The more complex change is necessary anyway to fix another short
coming of the current implementation, which I discovered while looking
at this problem: A pending signal is discarded when SIG_IGN is set. In
case that a posixtimer signal is pending then it is discarded as well,
but when SIG_IGN is removed later nothing rearms the timer. This is
not new, it's that way since posix timers have been merged. So nothing
to worry about right now.
I have a working solution to fix all of this, but the impact is too
large for both stable and 2.6.22. I'm going to send it out for review
in the next days.
This should go into 2.6.21.stable as well.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Cc: Jan Kiszka <jan.kiszka@web.de> Cc: Ulrich Drepper <drepper@redhat.com> Cc: Stable Team <stable@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Hugh Dickins [Thu, 21 Jun 2007 22:27:45 +0000 (23:27 +0100)]
page_mapping must avoid slub pages
Nicolas Ferre reports oops from flush_dcache_page() on ARM when using
SLUB: which reuses page->mapping as page->slab. The page_mapping()
function, used by ARM and PA-RISC flush_dcache_page() implementations,
must not confuse SLUB pages with those which have page->mapping set.
Signed-off-by: Hugh Dickins <hugh@veritas.com> Cc: Nicolas Ferre <nicolas.ferre@rfo.atmel.com> Cc: Christoph Lameter <clameter@sgi.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Randolph Chung [Tue, 12 Jun 2007 06:27:32 +0000 (14:27 +0800)]
[PARISC] unwinder improvements
Add special-case handling for "handle_interruption" so that we can rewind
past the interruption. This is useful for seeing what caused a BUG() or
WARN_ON(); otherwise the unwind stops at the interruption.
Helge Deller [Tue, 19 Dec 2006 18:28:33 +0000 (19:28 +0100)]
[PARISC] Handle wrapping in expand_upwards()
Function expand_upwards() did not guarded against wrapping
around to address 0. This fixes the adjtimex02 testcase from
the Linux Test Project on a 32bit PARISC kernel.
[expand_upwards is only used on parisc and ia64; it looks like it does
the right thing on both. --kyle]
Jack Morgenstein [Thu, 21 Jun 2007 20:39:10 +0000 (13:39 -0700)]
IB/mlx4: Correct max_srq_wr returned from mlx4_ib_query_device()
We need to keep a spare entry in the SRQ so that there always is a
next WQE available when posting receives (so that we can tell the
difference between a full queue and an empty queue). So subtract 1
from the value HW gives us before reporting the limit on SRQ entries
to consumers.
Found by Mellanox QA.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il> Signed-off-by: Roland Dreier <rolandd@cisco.com>
IPoIB/cm: Fix interoperability when MTU doesn't match
IPoIB connected mode currently rejects a connection request unless the
supported MTU is >= the local netdevice MTU. This breaks
interoperability with implementations that might have tweaked
IPOIB_CM_MTU, and there's real no longer a reason to do so: this test
is just a leftover from when we did not tweak MTU per-connection. Fix
this by making the test as permissive as possible.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il> Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fix a crasher bug in IPoIB CM: once a QP is in the RTR state, a
receive completion (or even an asynchronous error) might be observed
on this QP, so we have to initialize all of our receive data
structures before moving to the RTR state.
As an optimization (since modify_qp might take a long time), the
jiffies update done when moving RX to the passive_ids list is also
left in place to reduce the chance of the RX being misdetected as
stale.
This fixes bug <https://bugs.openfabrics.org/show_bug.cgi?id=662>.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il> Signed-off-by: Roland Dreier <rolandd@cisco.com>
Roland Dreier [Thu, 21 Jun 2007 18:05:58 +0000 (11:05 -0700)]
IB/umem: Fix possible hang on process exit
If ib_umem_release() is called after ib_uverbs_close() sets context->closing,
then a process can get stuck in a D state, because the code boils down to
if (down_write_trylock(&mm->mmap_sem))
down_write(&mm->mmap_sem);
which is obviously a stupid instant deadlock. Fix the code so that we
only try to take the lock once.
This bug was introduced in commit f7c6a7b5 ("IB/uverbs: Export
ib_umem_get()/ib_umem_release() to modules") which fortunately never
made it into a release, and was reported by Pete Wyckoff <pw@osc.edu>.
Wang Zhenyu [Thu, 21 Jun 2007 05:43:18 +0000 (13:43 +0800)]
[AGPGART] intel_agp: don't load if no IGD and AGP port
After i915 chip, GMCH has no AGP port. Origin bridge driver in device
table will try to access illegal regs like APBASE, APSIZE, etc. This
may cause problem.
So mark them as NULL in the table, we won't load if no IGD got detect
and bridge has no AGP port.
Signed-off-by: Wang Zhenyu <zhenyu.z.wang@intel.com> Signed-off-by: Dave Jones <davej@redhat.com>
Linus Torvalds [Thu, 21 Jun 2007 00:11:44 +0000 (17:11 -0700)]
Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
fix module_param mistake in it821x
ahci: fix PORTS_IMPL override
kerneldoc fix in libata
libata: more NONCQ devices
pata_it821x: (partially) fix DMA in RAID mode
PATA: Add the MCP73/77 support to PATA driver
Stas Sergeev [Wed, 20 Jun 2007 18:42:13 +0000 (22:42 +0400)]
fix module_param mistake in it821x
The attached patch fixes a trivial
mistake in a MODULE_PARAM_DESC of pata_it821x
driver. The parameter name in MODULE_PARAM_DESC
should match the one in module_param_named.
Signed-off-by: Stas Sergeev <stsp@aknet.ru> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Tejun Heo [Tue, 19 Jun 2007 09:52:56 +0000 (18:52 +0900)]
ahci: fix PORTS_IMPL override
If PORTS_IMPL register is zero, ahci initialize it to full mask
corresponding to nr_ports in the CAP register. hpriv->cap, which is
initialized at the end of the function, is incorrectly used as value
of CAP causing ahci to always override PORTS_IMPL to 0x1 if it's zero.
Fix it.
This fixes a bug where early ich6 ahci can only access the first port.
Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Code intended to check DMA status was checking DMA command register.
Moreover firmware seems to "forget" to set DMA capable bit for the
slave device (at least in RAID mode but without ITE RAID volumes) so
check device ID for DMA capable bit when deciding whether to use DMA
and remove DMA status check completely.
Thanks to Pavol Simo for the bugreport and testing the initial fix.
This change unfortunately still doesn't fix DMA in RAID mode (which
works fine with IDE it821x) but Alan is working on the missing pieces
(pata_it821x vs libata EH issues).
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com> Acked-by: Alan Cox <alan@redhat.com> Cc: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Divy Le Ray [Wed, 30 May 2007 17:01:44 +0000 (10:01 -0700)]
cxgb3 - Stop mac RX when changing MTU
Rx traffic needs to be halted when the MTU is changed
to avoid a potential chip hang.
Reset/restore MAC filters around a MTU change.
Also fix the pause frames high materwark setting.
Signed-off-by: Divy Le Ray <divy@chelsio.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Tim Mann [Thu, 14 Jun 2007 20:16:38 +0000 (13:16 -0700)]
forcedeth: use unicast receive mode for WoL
I happened to notice that a system with an NVidia NIC using the
forcedeth driver won't wake-on-LAN if the interface was in promiscuous
mode when you power off. By experiment, it looks like
the hardware needs to have NvRegPacketFilterFlags set to
NVREG_PFF_ALWAYS|NVREG_PFF_MYADDR (i.e., receive unicast packets to my
address) in order for WoL to work.
Jeff Garzik writes: "NVIDIA says the patch looks OK." I didn't venture
to insert a signed-off-by line with his name on it, though.
Signed-off-by: Tim Mann <mann@vmware.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
The recent iucv rework patches re-introduced some unnecessary inlines.
Remove them again.
Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Ursula Braun [Wed, 20 Jun 2007 11:01:30 +0000 (13:01 +0200)]
s390: don't call iucv_path_connect from tasklet context
net/iucv/iucv.c creates the requirement for
iucv_path_connect not to be called from tasklet context anymore.
An extra checking is added in case of a failing netiucv_tx
to fulfil this requirement for netiucv.
Signed-off-by: Ursula Braun <braunu@de.ibm.com> Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Frank Pavlic [Wed, 20 Jun 2007 10:59:14 +0000 (12:59 +0200)]
s390: qeth: wrong packet length in qdio header
Packets Length in qdio header is broken when using
EDDP on Layer2 devices. This leads to skb_under_panic on receiver
system when running on z/VM GuestLAN devices.
Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Ursula Braun [Wed, 20 Jun 2007 10:58:02 +0000 (12:58 +0200)]
s390: avoid inconsistent lock state in qeth
ipv6_regen_rndid in net/ipv6/addrconf.c makes use of "write_lock_bh"
for its inet6_dev->lock. It may run in softirq-context.
qeth makes use of "read_lock" for the same inet6_dev->lock.
To avoid a potential deadlock situation, qeth should make use of
"read_lock_bh" for its usages of inet6_dev->lock.
Signed-off-by: Ursula Braun <braunu@de.ibm.com> Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Ursula Braun [Wed, 20 Jun 2007 10:56:49 +0000 (12:56 +0200)]
s390: qeth driver does not recover
While first recovery continues, the card issues
a STARTLAN command itself. In this case qeth
schedules another recovery. This second
recovery is cancelled because of an already running first recovery.
Stop first recovery in case of 0xe080.
Signed-off-by: Ursula Braun <braunu@de.ibm.com> Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Jay Vosburgh [Thu, 14 Jun 2007 05:11:34 +0000 (22:11 -0700)]
bonding: Fix 802.3ad no carrier on "no partner found" instance
Modify carrier state determination for 802.3ad mode to comply
with section 43.3.9 of IEEE 802.3, which requires that "Links that are
not successful candidates for aggregation (e.g., links that are attached
to other devices that cannot perform aggregation or links that have been
manually configured to be non-aggregatable) are enabled to operate as
individual IEEE 802.3 links."
Bug reported by Laurent Chavey <chavey@google.com>. This patch
is an updated version of his patch that changes the wording of
commentary and adds an update to the driver version.
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com> Signed-off-by: Laurent Chavey <chavey@google.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Jay Vosburgh [Tue, 19 Jun 2007 18:12:12 +0000 (11:12 -0700)]
bonding: Fix use after free in unregister path
The following patch (based on a patch from Stephen Hemminger
<shemminger@linux-foundation.org>) removes use after free conditions in
the unregister path for the bonding master. Without this patch, an
operation of the form "echo -bond0 > /sys/class/net/bonding_masters"
would trigger a NULL pointer dereference in sysfs. I was not able to
induce the failure with the non-sysfs code path, but for consistency I
updated that code as well.
I also did some testing of the bonding /proc file being open
while the bond is being deleted, and didn't see any problems there.
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
It doesn't look like spidernet hardware can really checksum all protocols,
the code looks like it does IPV4 only. If so, it should use NETIF_F_IP_CSUM
instead of NETIF_F_HW_CSUM.
The driver doesn't need it's own get/set for ethtool tx csum, and it
should use the standard ethtool_op_get_link.
Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org> Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linas Vepstas [Mon, 11 Jun 2007 18:29:03 +0000 (13:29 -0500)]
spidernet: turn off descriptor chain end interrupt.
At some point, the transmit descriptor chain end interrupt (TXDCEINT)
was turned on. This is a mistake; and it damages small packet
transmit performance, as it results in a huge storm of interrupts.
Turn it off.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linas Vepstas [Mon, 11 Jun 2007 18:26:50 +0000 (13:26 -0500)]
spidernet: silence the ramfull messages
Although the previous patch resolved issues with hangs when the
RX ram full interrupt is encountered, there are still situations
where lots of RX ramfull interrupts arrive, resulting in a noisy
log in syslog. There is no need for this.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linas Vepstas [Mon, 11 Jun 2007 18:23:42 +0000 (13:23 -0500)]
spidernet: Don't terminate the RX ring
The terminated RX ring will cause trouble during the RX ram full
conditions, leading to a hung driver, as the hardware can't find
the next descr. There is no real reason to terminate the RX ring;
it doesn't make the operation any smooother, and it does
require an extra sync. So don't do it.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linas Vepstas [Mon, 11 Jun 2007 18:21:13 +0000 (13:21 -0500)]
spidernet: Cure RX ram full bug
This patch fixes a rare deadlock that can occur when the kernel
is not able to empty out the RX ring quickly enough. Below follows
a detailed description of the bug and the fix.
As long as the OS can empty out the RX buffers at a rate faster than
the hardware can fill them, there is no problem. If, for some reason,
the OS fails to empty the RX ring fast enough, the hardware GDACTDPA
pointer will catch up to the head, notice the not-empty condition,
ad stop. However, RX packets may still continue arriving on the wire.
The spidernet chip can save some limited number of these in local RAM.
When this local ram fills up, the spider chip will issue an interrupt
indicating this (GHIINT0STS will show ERRINT, and the GRMFLLINT bit
will be set in GHIINT1STS). When te RX ram full condition occurs,
a certain bug/feature is triggered that has to be specially handled.
This section describes the special handling for this condition.
When the OS finally has a chance to run, it will empty out the RX ring.
In particular, it will clear the descriptor on which the hardware had
stopped. However, once the hardware has decided that a certain
descriptor is invalid, it will not restart at that descriptor; instead
it will restart at the next descr. This potentially will lead to a
deadlock condition, as the tail pointer will be pointing at this descr,
which, from the OS point of view, is empty; the OS will be waiting for
this descr to be filled. However, the hardware has skipped this descr,
and is filling the next descrs. Since the OS doesn't see this, there
is a potential deadlock, with the OS waiting for one descr to fill,
while the hardware is waiting for a differen set of descrs to become
empty.
A call to show_rx_chain() at this point indicates the nature of the
problem. A typical print when the network is hung shows the following:
net eth1: Spider RX RAM full, incoming packets might be discarded!
net eth1: Total number of descrs=256
net eth1: Chain tail located at descr=255
net eth1: Chain head is at 255
net eth1: HW curr desc (GDACTDPA) is at 0
net eth1: Have 1 descrs with stat=xa0800000
net eth1: HW next desc (GDACNEXTDA) is at 1
net eth1: Have 127 descrs with stat=x40800101
net eth1: Have 1 descrs with stat=x40800001
net eth1: Have 126 descrs with stat=x40800101
net eth1: Last 1 descrs with stat=xa0800000
Both the tail and head pointers are pointing at descr 255, which is
marked xa... which is "empty". Thus, from the OS point of view, there
is nothing to be done. In particular, there is the implicit assumption
that everything in front of the "empty" descr must surely also be empty,
as explained in the last section. The OS is waiting for descr 255 to
become non-empty, which, in this case, will never happen.
The HW pointer is at descr 0. This descr is marked 0x4.. or "full".
Since its already full, the hardware can do nothing more, and thus has
halted processing. Notice that descrs 0 through 254 are all marked
"full", while descr 254 and 255 are empty. (The "Last 1 descrs" is
descr 254, since tail was at 255.) Thus, the system is deadlocked,
and there can be no forward progress; the OS thinks there's nothing
to do, and the hardware has nowhere to put incoming data.
This bug/feature is worked around with the spider_net_resync_head_ptr()
routine. When the driver receives RX interrupts, but an examination
of the RX chain seems to show it is empty, then it is probable that
the hardware has skipped a descr or two (sometimes dozens under heavy
network conditions). The spider_net_resync_head_ptr() subroutine will
search the ring for the next full descr, and the driver will resume
operations there. Since this will leave "holes" in the ring, there
is also a spider_net_resync_tail_ptr() that will skip over such holes.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linas Vepstas [Mon, 11 Jun 2007 18:17:02 +0000 (13:17 -0500)]
spidernet: null out skb pointer after its been used.
Avoid kernel crash in mm/slab.c due to double-free of pointer.
If the ethernet interface is brought down while there is still
RX traffic in flight, the device shutdown routine can end up
trying to double-free an skb, leading to a crash in mm/slab.c
Avoid the double-free by nulling out the skb pointer.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Linus Torvalds [Wed, 20 Jun 2007 22:39:50 +0000 (15:39 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
firewire: Only set client->iso_context if allocation was successful.
ieee1394: fix to ether1394_tx in ether1394.c
firewire: fix hang after card ejection
Linus Torvalds [Wed, 20 Jun 2007 22:39:24 +0000 (15:39 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
IB/mlx4: Make sure inline data segments don't cross a 64 byte boundary
IB/mlx4: Handle FW command interface rev 3
IB/mlx4: Handle buffer wraparound in __mlx4_ib_cq_clean()
IB/mlx4: Get rid of max_inline_data calculation
IB/mlx4: Handle new FW requirement for send request prefetching
IB/mlx4: Fix warning in rounding up queue sizes
IB/mlx4: Fix handling of wq->tail for send completions
Linus Torvalds [Wed, 20 Jun 2007 21:28:54 +0000 (14:28 -0700)]
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 powermac late initcall to only run on powermac
[POWERPC] PowerPC: Prevent data exception in kernel space (32-bit)
Andi Kleen [Wed, 20 Jun 2007 10:23:36 +0000 (12:23 +0200)]
x86: change_page_attr bandaids
- Disable CLFLUSH again; it is still broken. Always do WBINVD.
- Always flush in the i386 case, not only when there are deferred pages.
These are both brute-force inefficient fixes, to be improved
next release cycle.
The changes to i386 are a little more extensive than strictly
needed (some dead code added), but it is more similar to the x86-64 version
now and the dead code will be used soon.
Andi Kleen [Wed, 20 Jun 2007 10:23:33 +0000 (12:23 +0200)]
x86_64: Quieten Atari keyboard warnings in Kconfig
Not directly related to x86, but I got tired of seeing these warnings on every
kconfig update when building on a non m68k box:
drivers/input/keyboard/Kconfig:170:warning: 'select' used by config symbol 'KEYBOARD_ATARI' refers to undefined symbol 'ATARI_KBD_CORE'
drivers/input/mouse/Kconfig:182:warning: 'select' used by config symbol 'MOUSE_ATARI' refers to undefined symbol 'ATARI_KBD_CORE'
I moved the definition of ATARI_KBD_CORE into drivers/input/keyboard/Kconfig
so it's always seen by Kconfig.
Ralf Baechle [Wed, 20 Jun 2007 21:27:10 +0000 (22:27 +0100)]
[MIPS] Don't drag a platform specific header into generic arch code.
For some platforms it's definitions may conflict. So that's the one-liner.
The rest is 10 square kilometers of collateral damage fixup this include
used to paper over.
This happens because some powermac code unconditionally sets
ppc_md.progress to NULL. This patch makes sure the powermac late
initcall is only run on powermac machines.
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
[POWERPC] PowerPC: Prevent data exception in kernel space (32-bit)
The "is_exec" branch of the protection check in do_page_fault()
didn't do anything on 32-bit PowerPC. So if a userland program
jumps to a page with Linux protection flags "---p", all the tests
happily fall through, and handle_mm_fault() is called, which in
turn calls handle_pte_fault(), which calls update_mmu_cache(),
which goes flush the dcache to a page with no access rights.
Boom.
This fixes it.
Signed-off-by: Segher Boessenkool <segher@kernel.crashing.org> Cc: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: Paul Mackerras <paulus@samba.org>
Li Yang [Mon, 18 Jun 2007 11:29:21 +0000 (19:29 +0800)]
[POWERPC] rheap - eliminates internal fragments caused by alignment
The patch adds fragments caused by rh_alloc_align() back to free list, instead
of allocating the whole chunk of memory. This will greatly improve memory
utilization managed by rheap.
It solves MURAM not enough problem with 3 UCCs enabled on MPC8323.
Signed-off-by: Li Yang <leoli@freescale.com> Acked-by: Joakim Tjernlund <joakim.tjernlund@transmode.se> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Linus Torvalds [Tue, 19 Jun 2007 15:07:08 +0000 (08:07 -0700)]
Merge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6
* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
[S390] Move psw_set_key.
[S390] Add oops_enter()/oops_exit() calls to die().
[S390] Print list of modules on die().
[S390] Fix yet another two section mismatches.
[S390] Fix zfcpdump header
[S390] Missing blank when appending cio_ignore kernel parameter
Heiko Carstens [Tue, 19 Jun 2007 11:10:06 +0000 (13:10 +0200)]
[S390] Move psw_set_key.
Move psw_set_key() from ptrace.h to processor.h which is a more
suitable place for it. In addition the moves makes the function
invisible to user space.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Heiko Carstens [Tue, 19 Jun 2007 11:10:05 +0000 (13:10 +0200)]
[S390] Add oops_enter()/oops_exit() calls to die().
This is mainly to switch off all potentially debugging stuff that
won't report anything useful after an oops happened.
Besided that setting pause_on_oops will work too, but doesn't make
too much sense on s390.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Michael Holzheu [Tue, 19 Jun 2007 11:10:01 +0000 (13:10 +0200)]
[S390] Missing blank when appending cio_ignore kernel parameter
When appending the 'cio_ignore' kernel parameter to the command line, a blank
has to be inserted in order to separate 'cio_ignore' from the preceding kernel
parameters.
Signed-off-by: Michael Holzheu <holzheu@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Neil Horman [Tue, 19 Jun 2007 05:33:20 +0000 (22:33 -0700)]
[IPVS]: Fix state variable on failure to start ipvs threads
ip_vs currently fails to reset its ip_vs_sync_state variable if the
sync thread fails to start properly. The result is that the kernel
will report a running daemon when their actuall is none.
Patrick McHardy [Tue, 19 Jun 2007 05:30:15 +0000 (22:30 -0700)]
[XFRM]: Fix MTU calculation for non-ESP SAs
My IPsec MTU optimization patch introduced a regression in MTU calculation
for non-ESP SAs, the SA's header_len needs to be subtracted from the MTU if
the transform doesn't provide a ->get_mtu() function.
Reported-and-tested-by: Marco Berizzi <pupilla@hotmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
The old snd-powermac driver has some serious refcounting issues when
initialisation fails, which is the case on all new machines with
a layout-id since those are handled by the new snd-aoa driver.
Some of those bugs seem to have been under the radar for some time
(like double pci_dev_put), but one was actually added in 2.6.22 with
Stephen attempt at teaching refcounting to the driver which didn't
do it at all.
This patch fixes both, thus removing all sort of kref errors that
would happen if that driver gets loaded on a G5 machine or a recent
PowerBook due to OF nodes left around with a 0 refcount.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
Paul Mundt [Tue, 19 Jun 2007 03:41:32 +0000 (12:41 +0900)]
sh64: Handle -ERESTART_RESTARTBLOCK for restartable syscalls.
The current implementation only handles -ERESTARTNOHAND, whereas we
also need to handle -ERESTART_RESTARTBLOCK in the handle_signal()
case for restartable system calls. Follows the sh change.
Paul Mundt [Tue, 19 Jun 2007 03:33:21 +0000 (12:33 +0900)]
sh: Handle -ERESTART_RESTARTBLOCK for restartable syscalls.
The current implementation only handles -ERESTARTNOHAND, whereas we
also need to handle -ERESTART_RESTARTBLOCK in the handle_signal()
case for restartable system calls.
As noted by Carl:
This fixes the LTP test nanosleep03 - the current kernel causes
-ERESTART_RESTARTBLOCK to reach user space rather than the correct
-EINTR.
Reported-by: Carl Shaw <shaw.carl@gmail.com> Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Linus Torvalds [Mon, 18 Jun 2007 16:34:40 +0000 (09:34 -0700)]
Fix possible runqueue lock starvation in wait_task_inactive()
Miklos Szeredi reported very long pauses (several seconds, sometimes
more) on his T60 (with a Core2Duo) which he managed to track down to
wait_task_inactive()'s open-coded busy-loop.
He observed that an interrupt on one core tries to acquire the
runqueue-lock but does not succeed in doing so for a very long time -
while wait_task_inactive() on the other core loops waiting for the first
core to deschedule a task (which it wont do while spinning in an
interrupt handler).
This rewrites wait_task_inactive() to do all its waiting optimistically
without any locks taken at all, and then just double-check the end
result with the proper runqueue lock held over just a very short
section. If there were races in the optimistic wait, of a preemption
event scheduled the process away, we simply re-synchronize, and start
over.
/* Preempted away? Yield if so.. */
if (unlikely(array)) {
yield();
goto repeat;
}
Basically, that first "while()" loop is done entirely without any
locking at all (and doesn't check for the case where the target process
might have been preempted away), and so it's possibly "incorrect", but
we don't really care. Both the runqueue used, and the "task_running()"
check might be the wrong tests, but they won't oops - they just mean
that we could possibly get the wrong results due to lack of locking and
exit the loop early in the case of a race condition.
So once we've exited the loop, we then get the proper (and careful) rq
lock, and check the running/runnable state _safely_. And if it turns
out that our quick-and-dirty and unsafe loop was wrong after all, we
just go back and try it all again.
(The patch also adds a lot of comments, which is the actual bulk of it
all, to make it more obvious why we can do these things without holding
the locks).
Thanks to Miklos for all the testing and tracking it down.
Ingo Molnar [Sun, 17 Jun 2007 16:37:45 +0000 (18:37 +0200)]
sched: fix SysRq-N (normalize RT tasks)
Gene Heskett reported the following problem while testing CFS: SysRq-N
is not always effective in normalizing tasks back to SCHED_OTHER.
The reason for that turns out to be the following bug:
- normalize_rt_tasks() uses for_each_process() to iterate through all
tasks in the system. The problem is, this method does not iterate
through all tasks, it iterates through all thread groups.
The proper mechanism to enumerate over all threads is to use a
do_each_thread() + while_each_thread() loop.
Fix signalfd interaction with thread-private signals
Don't let signalfd dequeue private signals off other threads (in the
case of things like SIGILL or SIGSEGV, trying to do so would result
in undefined behaviour on who actually gets the signal, since they
are force unblocked).
projects / linux-2.6 / log