platform/kernel/linux-stable.git
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc
Linus Torvalds [Tue, 11 Sep 2007 17:12:34 +0000 (10:12 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/drzeus/mmc

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc:
  bug in AT91 MCI suspend routines

17 years agobug in AT91 MCI suspend routines
Anti Sullin [Thu, 30 Aug 2007 14:15:16 +0000 (16:15 +0200)]
bug in AT91 MCI suspend routines

This patch fixes a bug in AT91 mmc host driver, that enables the wakeup
from suspend on card detection pin even if the card detect pin is not
available (==0). If not card detection pin is defined, IRQ0 == FIQ gets
enabled and if some activity is present on that pin, the system gets a
FIQ request, that causes a crash.

Signed-off-by: Anti Sullin <anti.sullin@artecdesign.ee>
Signed-off-by: Nicolas Ferre <nicolas.ferre@rfo.atmel.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 11 Sep 2007 14:46:53 +0000 (07:46 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [INET_DIAG]: Fix oops in netlink_rcv_skb
  [IPv6]: Fix NULL pointer dereference in ip6_flush_pending_frames
  [NETFILTER]: Fix/improve deadlock condition on module removal netfilter
  [NETFILTER]: nf_conntrack_ipv4: fix "Frag of proto ..." messages
  [NET] DOC: Update networking/multiqueue.txt with correct information.
  [IPV6]: Freeing alive inet6 address
  [DECNET]: Fix interface address listing regression.
  [IPV4] devinet: show all addresses assigned to interface
  [NET]: Do not dereference iov if length is zero
  [TG3]: Workaround MSI bug on 5714/5780.
  [Bluetooth] Fix parameter list for event filter command
  [Bluetooth] Update security filter for Bluetooth 2.1
  [Bluetooth] Add compat handling for timestamp structure
  [Bluetooth] Add missing stat.byte_rx counter modification

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Tue, 11 Sep 2007 14:46:09 +0000 (07:46 -0700)]
Merge /linux/kernel/git/jejb/scsi-rc-fixes-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] libiscsi: sync up iscsi and scsi eh's access to the connection
  [SCSI] libiscsi: fix null ptr regression when aborting a command with data to transfer
  [SCSI] qla2xxx: Update version number to 8.02.00-k3.
  [SCSI] qla2xxx: Correct mailbox register dump for FWI2 capable ISPs.
  [SCSI] qla2xxx: Correct 8GB iIDMA support.
  [SCSI] qla2xxx: Correct management-server login-state synchronization issue.
  [SCSI] qla2xxx: Don't modify parity bits during ISP25XX restart.
  [SCSI] qla2xxx: Allocate enough space for the full PCI descriptor.
  [SCSI] zfcp: fix the data buffer accessor patch
  [SCSI] zfcp: allocate gid_pn_data objects from gid_pn_cache
  [SCSI] zfcp: fix memory leak

17 years ago[INET_DIAG]: Fix oops in netlink_rcv_skb
Patrick McHardy [Tue, 11 Sep 2007 09:33:28 +0000 (11:33 +0200)]
[INET_DIAG]: Fix oops in netlink_rcv_skb

netlink_run_queue() doesn't handle multiple processes processing the
queue concurrently. Serialize queue processing in inet_diag to fix
a oops in netlink_rcv_skb caused by netlink_run_queue passing a
NULL for the skb.

BUG: unable to handle kernel NULL pointer dereference at virtual address 00000054
[349587.500454]  printing eip:
[349587.500457] c03318ae
[349587.500459] *pde = 00000000
[349587.500464] Oops: 0000 [#1]
[349587.500466] PREEMPT SMP
[349587.500474] Modules linked in: w83627hf hwmon_vid i2c_isa
[349587.500483] CPU:    0
[349587.500485] EIP:    0060:[<c03318ae>]    Not tainted VLI
[349587.500487] EFLAGS: 00010246   (2.6.22.3 #1)
[349587.500499] EIP is at netlink_rcv_skb+0xa/0x7e
[349587.500506] eax: 00000000   ebx: 00000000   ecx: c148d2a0   edx: c0398819
[349587.500510] esi: 00000000   edi: c0398819   ebp: c7a21c8c   esp: c7a21c80
[349587.500517] ds: 007b   es: 007b   fs: 00d8  gs: 0033  ss: 0068
[349587.500521] Process oidentd (pid: 17943, ti=c7a20000 task=cee231c0 task.ti=c7a20000)
[349587.500527] Stack: 00000000 c7a21cac f7c8ba78 c7a21ca4 c0331962 c0398819 f7c8ba00 0000004c
[349587.500542]        f736f000 c7a21cb4 c03988e3 00000001 f7c8ba00 c7a21cc4 c03312a5 0000004c
[349587.500558]        f7c8ba00 c7a21cd4 c0330681 f7c8ba00 e4695280 c7a21d00 c03307c6 7fffffff
[349587.500578] Call Trace:
[349587.500581]  [<c010361a>] show_trace_log_lvl+0x1c/0x33
[349587.500591]  [<c01036d4>] show_stack_log_lvl+0x8d/0xaa
[349587.500595]  [<c010390e>] show_registers+0x1cb/0x321
[349587.500604]  [<c0103bff>] die+0x112/0x1e1
[349587.500607]  [<c01132d2>] do_page_fault+0x229/0x565
[349587.500618]  [<c03c8d3a>] error_code+0x72/0x78
[349587.500625]  [<c0331962>] netlink_run_queue+0x40/0x76
[349587.500632]  [<c03988e3>] inet_diag_rcv+0x1f/0x2c
[349587.500639]  [<c03312a5>] netlink_data_ready+0x57/0x59
[349587.500643]  [<c0330681>] netlink_sendskb+0x24/0x45
[349587.500651]  [<c03307c6>] netlink_unicast+0x100/0x116
[349587.500656]  [<c0330f83>] netlink_sendmsg+0x1c2/0x280
[349587.500664]  [<c02fcce9>] sock_sendmsg+0xba/0xd5
[349587.500671]  [<c02fe4d1>] sys_sendmsg+0x17b/0x1e8
[349587.500676]  [<c02fe92d>] sys_socketcall+0x230/0x24d
[349587.500684]  [<c01028d2>] syscall_call+0x7/0xb
[349587.500691]  =======================
[349587.500693] Code: f0 ff 4e 18 0f 94 c0 84 c0 0f 84 66 ff ff ff 89 f0 e8 86 e2 fc ff e9 5a ff ff ff f0 ff 40 10 eb be 55 89 e5 57 89 d7 56 89 c6 53 <8b> 50 54 83 fa 10 72 55 8b 9e 9c 00 00 00 31 c9 8b 03 83 f8 0f

Reported by Athanasius <link@miggy.org>

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPv6]: Fix NULL pointer dereference in ip6_flush_pending_frames
YOSHIFUJI Hideaki [Tue, 11 Sep 2007 09:31:43 +0000 (11:31 +0200)]
[IPv6]: Fix NULL pointer dereference in ip6_flush_pending_frames

Some of skbs in sk->write_queue do not have skb->dst because
we do not fill skb->dst when we allocate new skb in append_data().

BTW, I think we may not need to (or we should not) increment some stats
when using corking; if 100 sendmsg() (with MSG_MORE) result in 2 packets,
how many should we increment?

If 100, we should set skb->dst for every queued skbs.

If 1 (or 2 (*)), we increment the stats for the first queued skb and
we should just skip incrementing OutDiscards for the rest of queued skbs,
adn we should also impelement this semantics in other places;
e.g., we should increment other stats just once, not 100 times.

*: depends on the place we are discarding the datagram.

I guess should just increment by 1 (or 2).

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NETFILTER]: Fix/improve deadlock condition on module removal netfilter
Neil Horman [Tue, 11 Sep 2007 09:28:26 +0000 (11:28 +0200)]
[NETFILTER]: Fix/improve deadlock condition on module removal netfilter

So I've had a deadlock reported to me.  I've found that the sequence of
events goes like this:

1) process A (modprobe) runs to remove ip_tables.ko

2) process B (iptables-restore) runs and calls setsockopt on a netfilter socket,
increasing the ip_tables socket_ops use count

3) process A acquires a file lock on the file ip_tables.ko, calls remove_module
in the kernel, which in turn executes the ip_tables module cleanup routine,
which calls nf_unregister_sockopt

4) nf_unregister_sockopt, seeing that the use count is non-zero, puts the
calling process into uninterruptible sleep, expecting the process using the
socket option code to wake it up when it exits the kernel

4) the user of the socket option code (process B) in do_ipt_get_ctl, calls
ipt_find_table_lock, which in this case calls request_module to load
ip_tables_nat.ko

5) request_module forks a copy of modprobe (process C) to load the module and
blocks until modprobe exits.

6) Process C. forked by request_module process the dependencies of
ip_tables_nat.ko, of which ip_tables.ko is one.

7) Process C attempts to lock the request module and all its dependencies, it
blocks when it attempts to lock ip_tables.ko (which was previously locked in
step 3)

Theres not really any great permanent solution to this that I can see, but I've
developed a two part solution that corrects the problem

Part 1) Modifies the nf_sockopt registration code so that, instead of using a
use counter internal to the nf_sockopt_ops structure, we instead use a pointer
to the registering modules owner to do module reference counting when nf_sockopt
calls a modules set/get routine.  This prevents the deadlock by preventing set 4
from happening.

Part 2) Enhances the modprobe utilty so that by default it preforms non-blocking
remove operations (the same way rmmod does), and add an option to explicity
request blocking operation.  So if you select blocking operation in modprobe you
can still cause the above deadlock, but only if you explicity try (and since
root can do any old stupid thing it would like....  :)  ).

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NETFILTER]: nf_conntrack_ipv4: fix "Frag of proto ..." messages
Patrick McHardy [Tue, 11 Sep 2007 09:27:01 +0000 (11:27 +0200)]
[NETFILTER]: nf_conntrack_ipv4: fix "Frag of proto ..." messages

Since we're now using a generic tuple decoding function in ICMP
connection tracking, ipv4_get_l4proto() might get called with a
fragmented packet from within an ICMP error. Remove the error
message we used to print when this happens.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/holtmann/bluetooth-2.6
David S. Miller [Tue, 11 Sep 2007 09:15:30 +0000 (11:15 +0200)]
Merge /linux/kernel/git/holtmann/bluetooth-2.6

17 years ago[NET] DOC: Update networking/multiqueue.txt with correct information.
Peter P Waskiewicz Jr [Tue, 11 Sep 2007 09:12:06 +0000 (11:12 +0200)]
[NET] DOC: Update networking/multiqueue.txt with correct information.

Updated the multiqueue.txt document to call out the correct kernel
options to select to enable multiqueue.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPV6]: Freeing alive inet6 address
Denis V. Lunev [Tue, 11 Sep 2007 09:04:49 +0000 (11:04 +0200)]
[IPV6]: Freeing alive inet6 address

From: Denis V. Lunev <den@openvz.org>

addrconf_dad_failure calls addrconf_dad_stop which takes referenced address
and drops the count. So, in6_ifa_put perrformed at out: is extra. This
results in message: "Freeing alive inet6 address" and not released dst entries.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: Alexey Dobriyan <adobriyan@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[DECNET]: Fix interface address listing regression.
Patrick McHardy [Tue, 11 Sep 2007 08:45:15 +0000 (10:45 +0200)]
[DECNET]: Fix interface address listing regression.

Not all are listed, same as the IPV4 devinet bug.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPV4] devinet: show all addresses assigned to interface
Stephen Hemminger [Tue, 11 Sep 2007 08:41:04 +0000 (10:41 +0200)]
[IPV4] devinet: show all addresses assigned to interface

Bug: http://bugzilla.kernel.org/show_bug.cgi?id=8876

Not all ips are shown by "ip addr show" command when IPs number assigned to an
interface is more than 60-80 (in fact it depends on broadcast/label etc
presence on each address).

Steps to reproduce:
It's terribly simple to reproduce:

# for i in $(seq 1 100); do ip ad add 10.0.$i.1/24 dev eth10 ; done
# ip addr show

this will _not_ show all IPs.
Looks like the problem is in netlink/ipv4 message processing.

This is fix from bug submitter, it looks correct.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: Do not dereference iov if length is zero
Herbert Xu [Thu, 6 Sep 2007 13:06:35 +0000 (14:06 +0100)]
[NET]: Do not dereference iov if length is zero

When msg_iovlen is zero we shouldn't try to dereference
msg_iov.  Right now the only thing that tries to do so
is skb_copy_and_csum_datagram_iovec.  Since the total
length should also be zero if msg_iovlen is zero, it's
sufficient to check the total length there and simply
return if it's zero.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TG3]: Workaround MSI bug on 5714/5780.
Michael Chan [Thu, 6 Sep 2007 11:04:29 +0000 (12:04 +0100)]
[TG3]: Workaround MSI bug on 5714/5780.

A hardware bug was revealed after a recent PCI MSI patch was made to
always disable legacy INTX when enabling MSI.  The 5714/5780 chips
will not generate MSI when INTX is disabled, causing MSI failure
messages to be reported, and another patch was made to workaround the
problem by disabling MSI on ServerWorks HT1000 bridge chips commonly
found with the 5714.

We workaround this chip bug by enabling INTX after we enable MSI and
after we resume from suspend.

Update version to 3.81.

This problem was discovered by David Miller.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Acked-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoLinux 2.6.23-rc6 v2.6.23-rc6
Linus Torvalds [Tue, 11 Sep 2007 02:49:55 +0000 (19:49 -0700)]
Linux 2.6.23-rc6

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Tue, 11 Sep 2007 02:22:59 +0000 (19:22 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  libata clear horkage on ata_dev_init()
  [libata, IDE] add new VIA bridge to VIA PATA drivers
  pata_it821x: fix lost interrupt with atapi devices
  Fix broken pata_via cable detection

17 years agolibata clear horkage on ata_dev_init()
Tejun Heo [Mon, 3 Sep 2007 03:20:11 +0000 (12:20 +0900)]
libata clear horkage on ata_dev_init()

dev->horkage should be cleared over device hotunplug/plug.  Clear it
in ata_dev_init().

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years ago[libata, IDE] add new VIA bridge to VIA PATA drivers
Joseph Chan [Tue, 11 Sep 2007 02:06:01 +0000 (22:06 -0400)]
[libata, IDE] add new VIA bridge to VIA PATA drivers

Signed-off-by: Joseph Chan <josephchan@via.com.tw>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoUML: Fix ELF_CORE_COPY_REGS build botch
Jeff Dike [Mon, 10 Sep 2007 15:55:46 +0000 (11:55 -0400)]
UML: Fix ELF_CORE_COPY_REGS build botch

The earlier crash dump fix on x86_64 depended on patches in -mm which
are intended for post-2.6.23.  Without those, it broke the build when
it went into 2.6.23-rc5.

This changes the field references in ELF_CORE_COPY_REGS back to those
still used in mainline.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoi386: fix a hang on stuck nmi watchdog
Daniel Walker [Thu, 6 Sep 2007 14:59:54 +0000 (16:59 +0200)]
i386: fix a hang on stuck nmi watchdog

In the case when an nmi gets stucks the endflag stays equal to zero.
This causes the busy looping on other cpus to continue, even though the
nmi test is done.

On my machine with out the change below the system would hang right
after check_nmi_watchdog().  The change below just sets endflag prior to
checking if the test was successful or not.

Signed-off-by: Daniel Walker <dwalker@mvista.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agox86_64: Remove CLFLUSH in text_poke()
Andi Kleen [Thu, 6 Sep 2007 14:59:52 +0000 (16:59 +0200)]
x86_64: Remove CLFLUSH in text_poke()

The CLFLUSH for the modified code line in text_poke was supposed
to speed up CPU recovery. Unfortunately it seems to cause hangs
on some VIA C3s (at least on VIA Esther Model 10 Stepping 9)
Remove it.

Thanks to Stefan Becker for reporting/testing.

Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoi386: Fix perfctr watchdog on core duo
Stephane Eranian [Thu, 6 Sep 2007 14:59:51 +0000 (16:59 +0200)]
i386: Fix perfctr watchdog on core duo

Fix the NMI watchdog on Intel CoreDuo processor where the kernel would
get stuck during boot.  The issue is related to errata AE49, where the
PERFEVTSEL1 counter does not have a working enable bit.  Thus it is not
possible to use it for NMI.

The patch creates a dedicated wd_ops for CoreDuo which falls back to
using PERFEVTSEL0.  The other Intel processors supporting the
architectural PMU will keep on using PERFEVTSEL1 as this allows other
subsystems, such as perfmon, to use PERFEVTSEL0 for PEBS monitoring in
particular.  Bug initially reported by Daniel Walker.

AK: Added comments

Signed-off-by: Stephane Eranian <eranian@hpl.hp.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoknfsd: Validate filehandle type in fsid_source
Neil Brown [Wed, 5 Sep 2007 21:22:13 +0000 (17:22 -0400)]
knfsd: Validate filehandle type in fsid_source

fsid_source decided where to get the 'fsid' number to
return for a GETATTR based on the type of filehandle.
It can be from the device, from the fsid, or from the
UUID.

It is possible for the filehandle to be inconsistent
with the export information, so make sure the export information
actually has the info implied by the value returned by
fsid_source.

Signed-off-by: Neil Brown <neilb@suse.de>
Cc: "Luiz Fernando N. Capitulino" <lcapitulino@gmail.com>
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoknfsd: Fixed problem with NFS exporting directories which are mounted on.
Neil Brown [Wed, 5 Sep 2007 21:22:12 +0000 (17:22 -0400)]
knfsd: Fixed problem with NFS exporting directories which are mounted on.

Recent changes in NFSd cause a directory which is mounted-on
to not appear properly when the filesystem containing it is exported.

*exp_get* now returns -ENOENT rather than NULL and when
  commit 5d3dbbeaf56d0365ac6b5c0a0da0bd31cc4781e1
removed the NULL checks, it didn't add a check for -ENOENT.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoFix spurious syscall tracing after PTRACE_DETACH + PTRACE_ATTACH
Roland McGrath [Wed, 5 Sep 2007 10:05:56 +0000 (03:05 -0700)]
Fix spurious syscall tracing after PTRACE_DETACH + PTRACE_ATTACH

When PTRACE_SYSCALL was used and then PTRACE_DETACH is used, the
TIF_SYSCALL_TRACE flag is left set on the formerly-traced task.  This
means that when a new tracer comes along and does PTRACE_ATTACH, it's
possible he gets a syscall tracing stop even though he's never used
PTRACE_SYSCALL.  This happens if the task was in the middle of a system
call when the second PTRACE_ATTACH was done.  The symptom is an
unexpected SIGTRAP when the tracer thinks that only SIGSTOP should have
been provoked by his ptrace calls so far.

A few machines already fixed this in ptrace_disable (i386, ia64, m68k).
But all other machines do not, and still have this bug.  On x86_64, this
constitutes a regression in IA32 compatibility support.

Since all machines now use TIF_SYSCALL_TRACE for this, I put the
clearing of TIF_SYSCALL_TRACE in the generic ptrace_detach code rather
than adding it to every other machine's ptrace_disable.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agopata_it821x: fix lost interrupt with atapi devices
Jeff Norden [Tue, 4 Sep 2007 16:07:20 +0000 (11:07 -0500)]
pata_it821x: fix lost interrupt with atapi devices

Fix "lost" interrupt problem when using dma with CD/DVD drives in some
configurations.  This problem can make installing linux from media
impossible for distro's that have switched to libata-only configurations.

The simple fix is to eliminate the use of dma for reading drive status, etc,
by checking the number of bytes to transferred.

This change will only affect the behavior of atapi devices, not disks.
There is more info at http://bugzilla.redhat.com/show_bug.cgi?id=242229
This patch is for 2.6.22.1

Signed-off-by: Jeff Norden <jnorden@math.tntech.edu>
Reviewed-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoFix broken pata_via cable detection
Laurent Riffard [Sun, 2 Sep 2007 19:01:32 +0000 (21:01 +0200)]
Fix broken pata_via cable detection

via_do_set_mode overwrites 80-wire cable detection bits. Let's
preserve them.

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Mon, 10 Sep 2007 21:43:37 +0000 (14:43 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Ocelot: remove remaining bits
  [MIPS] TLB: Fix instruction bitmasks
  [MIPS] R10000: Fix wrong test in dma-default.c
  [MIPS] Provide empty irq_enable_hazard definition for legacy and R1 cores.
  [MIPS] Sibyte: Remove broken dependency on EXPERIMENTAL from SIBYTE_SB1xxx_SOC.
  [MIPS] Kconfig: whitespace cleanup.
  [MIPS] PCI: Set need_domain_info if controller domain index is non-zero.
  [MIPS] BCM1480: Fix computation of interrupt mask address register.
  [MIPS] i8259: Add disable method.
  [MIPS] tty: add the new ioctls and definitions.

17 years agoMerge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
Linus Torvalds [Mon, 10 Sep 2007 21:42:39 +0000 (14:42 -0700)]
Merge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  i2c-algo-bit: Read block data bugfix
  i2c-pxa: Fix adapter number
  i2c-gpio: Fix adapter number

17 years agoMerge branch 'release' of git://lm-sensors.org/kernel/mhoffman/hwmon-2.6
Linus Torvalds [Mon, 10 Sep 2007 21:41:25 +0000 (14:41 -0700)]
Merge branch 'release' of git://lm-sensors.org/kernel/mhoffman/hwmon-2.6

* 'release' of git://lm-sensors.org/kernel/mhoffman/hwmon-2.6:
  hwmon: End of I/O region off-by-one

17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Mon, 10 Sep 2007 21:40:29 +0000 (14:40 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  [CRYPTO] blkcipher: Fix inverted test in blkcipher_get_spot
  [CRYPTO] blkcipher: Fix handling of kmalloc page straddling

17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Mon, 10 Sep 2007 21:38:15 +0000 (14:38 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  firewire: fw-ohci: ignore failure of pci_set_power_state (fix suspend regression)

17 years agoMerge branch 'for-linus' of git://www.linux-m32r.org/git/takata/linux-2.6_dev
Linus Torvalds [Mon, 10 Sep 2007 21:37:20 +0000 (14:37 -0700)]
Merge branch 'for-linus' of git://linux-m32r.org/git/takata/linux-2.6_dev

* 'for-linus' of git://www.linux-m32r.org/git/takata/linux-2.6_dev:
  m32r: Rename STI/CLI macros
  m32r: build fix of entry.S
  m32r: Separate syscall table from entry.S
  m32r: Cosmetic updates of arch/m32r/kernel/entry.S
  m32r: Exit ei_handler directly for no IRQ case or IPI operations
  m32r: Simplify ei_handler code
  m32r: Define symbols to unify platform-dependent ICU checks
  m32r: Move dot.gdbinit files
  m32r: Rearrange platform-dependent codes
  m32r: Add defconfig file for the usrv platform.
  m32r: Update defconfig files for 2.6.23-rc1
  m32r: Move defconfig files to arch/m32r/configs/

17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched
Linus Torvalds [Mon, 10 Sep 2007 21:36:49 +0000 (14:36 -0700)]
Merge git://git./linux/kernel/git/mingo/linux-2.6-sched

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched:
  sched: fix ideal_runtime calculations for reniced tasks
  sched: improve prev_sum_exec_runtime setting
  sched: simplify __check_preempt_curr_fair()
  sched: fix xtensa build warning
  sched: debug: fix sum_exec_runtime clearing
  sched: debug: fix cfs_rq->wait_runtime accounting
  sched: fix niced_granularity() shift
  sched: fix MC/HT scheduler optimization, without breaking the FUZZ logic.

17 years agoMerge branch 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6
Linus Torvalds [Mon, 10 Sep 2007 21:36:37 +0000 (14:36 -0700)]
Merge branch 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6

* 'for-linus' of git://oss.sgi.com:8090/xfs/xfs-2.6:
  [XFS] fix nasty quota hashtable allocation bug
  [XFS] fix sparse shadowed variable warnings
  [XFS] fix ASSERT and ASSERT_ALWAYS
  [XFS] Fix sparse warning in kmem_shake_allow
  [XFS] Fix sparse NULL vs 0 warnings
  [XFS] Set filestreams object timeout to something sane.

17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux...
Linus Torvalds [Mon, 10 Sep 2007 21:35:55 +0000 (14:35 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/hpa/linux-2.6-x86setup

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup:
  [x86 setup] Work around bug in Xen HVM

17 years ago[MIPS] Ocelot: remove remaining bits
Yoichi Yuasa [Thu, 6 Sep 2007 12:32:57 +0000 (21:32 +0900)]
[MIPS] Ocelot: remove remaining bits

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] TLB: Fix instruction bitmasks
Thiemo Seufer [Wed, 5 Sep 2007 11:11:22 +0000 (12:11 +0100)]
[MIPS] TLB: Fix instruction bitmasks

Signed-Off-By: Thiemo Seufer <ths@networkno.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] R10000: Fix wrong test in dma-default.c
Maxime Bizon [Fri, 31 Aug 2007 09:03:14 +0000 (11:03 +0200)]
[MIPS] R10000: Fix wrong test in dma-default.c

Signed-off-by: Maxime Bizon <mbizon@freebox.fr>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Provide empty irq_enable_hazard definition for legacy and R1 cores.
Ralf Baechle [Mon, 3 Sep 2007 14:22:26 +0000 (16:22 +0200)]
[MIPS] Provide empty irq_enable_hazard definition for legacy and R1 cores.

Following a strict interpretation the empty definition of irq_enable_hazard
has always been a bug - but an intentional one because it didn't bite.
This has now changed, for uniprocessor kernels mm/slab.c:do_drain()

[...]
        on_each_cpu(do_drain, cachep, 1, 1);
        check_irq_on();
[...]

may be compiled into a mtc0 c0_status; mfc0 c0_status sequence resulting
in a back-to-back hazard.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Sibyte: Remove broken dependency on EXPERIMENTAL from SIBYTE_SB1xxx_SOC.
Ralf Baechle [Wed, 29 Aug 2007 23:58:53 +0000 (00:58 +0100)]
[MIPS] Sibyte: Remove broken dependency on EXPERIMENTAL from SIBYTE_SB1xxx_SOC.

Otherwise Kconfig will produce a nonsenical .config for a kernel that is
neither 32-bit nor 64-bit.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Kconfig: whitespace cleanup.
Ralf Baechle [Wed, 29 Aug 2007 13:21:45 +0000 (14:21 +0100)]
[MIPS] Kconfig: whitespace cleanup.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] PCI: Set need_domain_info if controller domain index is non-zero.
Ralf Baechle [Wed, 29 Aug 2007 07:34:39 +0000 (08:34 +0100)]
[MIPS] PCI: Set need_domain_info if controller domain index is non-zero.

This fixes this little funny:

bigsur:/proc/bus/pci# ls -l
total 0
dr-xr-xr-x    2 root     root            0 Aug 28 19:31 00
dr-xr-xr-x    2 root     root            0 Aug 28 19:31 00
dr-xr-xr-x    2 root     root            0 Aug 28 19:31 01
dr-xr-xr-x    2 root     root            0 Aug 28 19:31 03
-r--r--r--    1 root     root            0 Aug 28 19:31 devices

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] BCM1480: Fix computation of interrupt mask address register.
Ralf Baechle [Tue, 28 Aug 2007 23:38:13 +0000 (00:38 +0100)]
[MIPS] BCM1480: Fix computation of interrupt mask address register.

  CC      arch/mips/sibyte/bcm1480/irq.o
arch/mips/sibyte/bcm1480/irq.c: In function 'bcm1480_mask_irq':
arch/mips/sibyte/bcm1480/irq.c:112: warning: cast to pointer from integer of different size
arch/mips/sibyte/bcm1480/irq.c:114: warning: cast to pointer from integer of different size
arch/mips/sibyte/bcm1480/irq.c: In function 'bcm1480_unmask_irq':
arch/mips/sibyte/bcm1480/irq.c:130: warning: cast to pointer from integer of different size
arch/mips/sibyte/bcm1480/irq.c:132: warning: cast to pointer from integer of different size

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] i8259: Add disable method.
Kyungmin Park [Fri, 10 Aug 2007 21:00:21 +0000 (14:00 -0700)]
[MIPS] i8259: Add disable method.

After 76d2160147f43f982dfe881404cfde9fd0a9da21, the qemu NE2000 was
frequently producing WATCHDOG timeouts.

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] tty: add the new ioctls and definitions.
Alan Cox [Fri, 10 Aug 2007 21:00:20 +0000 (14:00 -0700)]
[MIPS] tty: add the new ioctls and definitions.

Same as all the others, just put in the constants for the existing kernel
code and termios2 structure

Signed-off-by: Alan Cox <alan@redhat.com>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[POWERPC] ibmebus: Prevent bus_id collisions
Joachim Fenkes [Wed, 29 Aug 2007 16:15:17 +0000 (18:15 +0200)]
[POWERPC] ibmebus: Prevent bus_id collisions

Previously, ibmebus derived a device's bus_id from its location code.
The location code is not guaranteed to be unique, so we might get bus_id
collisions if two devices share the same location code.  The OFDT
full_name, however, is unique, so we use that instead (truncating it
on the left if it is too long).

Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years ago[POWERPC] cell/PS3: Ignore storage devices that are still being probed
Geert Uytterhoeven [Thu, 6 Sep 2007 16:14:57 +0000 (18:14 +0200)]
[POWERPC] cell/PS3: Ignore storage devices that are still being probed

On PS3, A storage device may show up in the repository before the hypervisor
has finished probing:
  - If its type is not yet known, it shows up as PS3_DEV_TYPE_STOR_DUMMY,
  - If its regions are being probed, it shows up as having zero regions.
If any of these happen, consider the device not yet present.  The storage
probe thread will retry later.

This fixes the timing-dependent problem where a kernel booted from FLASH ROM
sometimes cannot find the hard disk.

Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Acked-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years ago[POWERPC] cell/PS3: Always set master run control bit in mfc_sr1_set
Jeremy Kerr [Fri, 7 Sep 2007 08:28:27 +0000 (18:28 +1000)]
[POWERPC] cell/PS3: Always set master run control bit in mfc_sr1_set

At present, running any SPE program on the ps3 will trigger a BUG_ON
when spufs_run_spu tries to clear the master run control bit, as lv1
does not make the master run control available to Linux.

This change makes SPE apps work again by disabling changes to the
master run control on PS3.  Although we don't have the facility to
disable a SPE with supervisor-level privileges, it's better than
hitting the BUG_ON unconditionally.

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Acked-by: Masato Noguchi <Masato.Noguchi@jp.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years ago[POWERPC] cell/PS3: Fix a bug that causes the PS3 to hang on the SPU Class 0 interrupt.
Masato Noguchi [Fri, 7 Sep 2007 08:28:27 +0000 (18:28 +1000)]
[POWERPC] cell/PS3: Fix a bug that causes the PS3 to hang on the SPU Class 0 interrupt.

The Cell BE Architecture spec states that the SPU MFC Class 0 interrupt
is edge-triggered.  The current spu interrupt handler assumes this
behavior and does not clear the interrupt status.

The PS3 hypervisor visualizes all SPU interrupts as level, and on return
from the interrupt handler the hypervisor will deliver a new virtual
interrupt for any unmasked interrupts which for which the status has not
been cleared.  This fix clears the interrupt status in the interrupt
handler.

Signed-off-by: Masato Noguchi <Masato.Noguchi@jp.sony.com>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Acked-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years ago[CRYPTO] blkcipher: Fix inverted test in blkcipher_get_spot
Herbert Xu [Mon, 10 Sep 2007 07:51:11 +0000 (15:51 +0800)]
[CRYPTO] blkcipher: Fix inverted test in blkcipher_get_spot

The previous patch had the conditional inverted.  This patch fixes it
so that we return the original position if it does not straddle a page.

Thanks to Bob Gilligan for spotting this.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
17 years agoi2c-algo-bit: Read block data bugfix
David Brownell [Sun, 9 Sep 2007 20:29:14 +0000 (22:29 +0200)]
i2c-algo-bit: Read block data bugfix

This fixes a bug in the way i2c-algo-bit handles I2C_M_RECV_LEN,
used to implement i2c_smbus_read_block_data().  Previously, in the
absence of PEC (rarely used!) it would NAK the "length" byte:

S addr Rd [A] [length] NA

That prevents the subsequent data bytes from being read:

S addr Rd [A] [length] { A [data] }* NA

The primary fix just reorders two code blocks, so the length used
in the "should I NAK now?" check incorporates the data which it
just read from the slave device.

However, that move also highlighted other fault handling glitches.
This fixes those by abstracting the RX path ack/nak logic, so it
can be used in more than one location.  Also, a few CodingStyle
issues were also resolved.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
17 years agoi2c-pxa: Fix adapter number
Jean Delvare [Sun, 9 Sep 2007 20:29:13 +0000 (22:29 +0200)]
i2c-pxa: Fix adapter number

It turns out that platform_device.id is a "u32" so testing it for being
nonnegative is useless when setting up an i2c adapte.  Instead,
do what the platform_bus code does and test it against the value "-1".

Signed-off-by: Jean Delvare <khali@linux-fr.org>
17 years agoi2c-gpio: Fix adapter number
David Brownell [Sun, 9 Sep 2007 20:29:13 +0000 (22:29 +0200)]
i2c-gpio: Fix adapter number

It turns out that platform_device.id is a "u32" so testing it for being
nonnegative is useless when setting up an i2c bitbang device.  Instead,
do what the platform_bus code does and test it against the value "-1".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
17 years agohwmon: End of I/O region off-by-one
Jean Delvare [Wed, 29 Aug 2007 08:39:57 +0000 (10:39 +0200)]
hwmon: End of I/O region off-by-one

Fix an off-by-one error in the I/O region declaration of two
hardware monitoring drivers (lm78 and w83781d.) We were requesting
one extra port at the end of the region.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
17 years ago[CRYPTO] blkcipher: Fix handling of kmalloc page straddling
Herbert Xu [Sun, 9 Sep 2007 07:45:21 +0000 (08:45 +0100)]
[CRYPTO] blkcipher: Fix handling of kmalloc page straddling

The function blkcipher_get_spot tries to return a buffer of
the specified length that does not straddle a page.  It has
an off-by-one bug so it may advance a page unnecessarily.

What's worse, one of its callers doesn't provide a buffer
that's sufficiently long for this operation.

This patch fixes both problems.  Thanks to Bob Gilligan for
diagnosing this problem and providing a fix.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
17 years ago[Bluetooth] Fix parameter list for event filter command
Marcel Holtmann [Sun, 9 Sep 2007 06:39:49 +0000 (08:39 +0200)]
[Bluetooth] Fix parameter list for event filter command

On device initialization the event filters are cleared. In case of
clearing the filters the extra condition type shall be omitted.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Update security filter for Bluetooth 2.1
Marcel Holtmann [Sun, 9 Sep 2007 06:39:43 +0000 (08:39 +0200)]
[Bluetooth] Update security filter for Bluetooth 2.1

This patch updates the HCI security filter with support for the
Bluetooth 2.1 commands and events.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Add compat handling for timestamp structure
Marcel Holtmann [Sun, 9 Sep 2007 06:39:34 +0000 (08:39 +0200)]
[Bluetooth] Add compat handling for timestamp structure

The timestamp structure needs special handling in case of compat
programs. Use the same wrapping method the network core uses.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Add missing stat.byte_rx counter modification
Marcel Holtmann [Sun, 9 Sep 2007 06:39:27 +0000 (08:39 +0200)]
[Bluetooth] Add missing stat.byte_rx counter modification

With the support for hci_recv_fragment() the call to increase the
stat.byte_rx counter got accidentally removed. This patch fixes it.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years agofirewire: fw-ohci: ignore failure of pci_set_power_state (fix suspend regression)
Stefan Richter [Thu, 6 Sep 2007 07:50:30 +0000 (09:50 +0200)]
firewire: fw-ohci: ignore failure of pci_set_power_state (fix suspend regression)

Fixes (papers over) "Sleep problems with kernels >= 2.6.21 on powerpc",
http://lkml.org/lkml/2007/8/25/155.  The issue is that the FireWire
controller's pci_dev.current_state of iBook G3 and presumably older
PowerBooks is still in PCI_UNKNOWN instead of PCI_D0 when the firewire
driver's .suspend method is called.

Like it was suggested earlier in http://lkml.org/lkml/2006/10/24/13, we
do not fail .suspend anymore if pci_set_power_state failed.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
17 years agom32r: Rename STI/CLI macros
Hirokazu Takata [Mon, 20 Aug 2007 11:53:50 +0000 (20:53 +0900)]
m32r: Rename STI/CLI macros

The names of STI and CLI macros were derived from i386 arch historically,
but their name are incomprehensible.
So, for easy to understand, rename these macros to ENABLE_INTERRUPTS
and DISABLE_INTERRUPTS, respectively.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: build fix of entry.S
Hirokazu Takata [Tue, 21 Aug 2007 03:04:29 +0000 (12:04 +0900)]
m32r: build fix of entry.S

This patch fixes the following compile error:
<--  snip  -->
 ...
  AS      arch/m32r/kernel/entry.o
/home/bunk/linux/kernel-2.6/linux-2.6.23-rc3-mm1/arch/m32r/kernel/entry.S: Assembler messages:
/home/bunk/linux/kernel-2.6/linux-2.6.23-rc3-mm1/arch/m32r/kernel/entry.S:358: Error: bad instruction `addi r0,#(((((0)+(64))+(32))+(32)))'
make[2]: *** [arch/m32r/kernel/entry.o] Error 1
<--  snip  -->

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
Cc: Adrian Bunk <bunk@kernel.org>
17 years agosched: fix ideal_runtime calculations for reniced tasks
Peter Zijlstra [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: fix ideal_runtime calculations for reniced tasks

fix ideal_runtime:

  - do not scale it using niced_granularity()
    it is against sum_exec_delta, so its wall-time, not fair-time.

  - move the whole check into __check_preempt_curr_fair()
    so that wakeup preemption can also benefit from the new logic.

this also results in code size reduction:

   text    data     bss     dec     hex filename
  13391     228    1204   14823    39e7 sched.o.before
  13369     228    1204   14801    39d1 sched.o.after

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: improve prev_sum_exec_runtime setting
Peter Zijlstra [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: improve prev_sum_exec_runtime setting

Second preparatory patch for fix-ideal runtime:

Mark prev_sum_exec_runtime at the beginning of our run, the same spot
that adds our wait period to wait_runtime. This seems a more natural
location to do this, and it also reduces the code a bit:

   text    data     bss     dec     hex filename
  13397     228    1204   14829    39ed sched.o.before
  13391     228    1204   14823    39e7 sched.o.after

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: simplify __check_preempt_curr_fair()
Peter Zijlstra [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: simplify __check_preempt_curr_fair()

Preparatory patch for fix-ideal-runtime:

simplify __check_preempt_curr_fair(): get rid of the integer return.

   text    data     bss     dec     hex filename
  13404     228    1204   14836    39f4 sched.o.before
  13393     228    1204   14825    39e9 sched.o.after

functionality is unchanged.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: fix xtensa build warning
Ingo Molnar [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: fix xtensa build warning

rename RSR to SRR - 'RSR' is already defined on xtensa.

found by Adrian Bunk.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: debug: fix sum_exec_runtime clearing
Ingo Molnar [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: debug: fix sum_exec_runtime clearing

when cleaning sched-stats also clear prev_sum_exec_runtime.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: debug: fix cfs_rq->wait_runtime accounting
Ingo Molnar [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: debug: fix cfs_rq->wait_runtime accounting

the cfs_rq->wait_runtime debug/statistics counter was not maintained
properly - fix this.

this also removes some code:

   text    data     bss     dec     hex filename
  13420     228    1204   14852    3a04 sched.o.before
  13404     228    1204   14836    39f4 sched.o.after

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
17 years agosched: fix niced_granularity() shift
Ingo Molnar [Wed, 5 Sep 2007 12:32:49 +0000 (14:32 +0200)]
sched: fix niced_granularity() shift

fix niced_granularity(). This resulted in under-scheduling for
CPU-bound negative nice level tasks (and this in turn caused
higher than necessary latencies in nice-0 tasks).

Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agosched: fix MC/HT scheduler optimization, without breaking the FUZZ logic.
Suresh Siddha [Wed, 5 Sep 2007 12:32:48 +0000 (14:32 +0200)]
sched: fix MC/HT scheduler optimization, without breaking the FUZZ logic.

First fix the check
if (*imbalance + SCHED_LOAD_SCALE_FUZZ < busiest_load_per_task)
with this
if (*imbalance < busiest_load_per_task)

As the current check is always false for nice 0 tasks (as
SCHED_LOAD_SCALE_FUZZ is same as busiest_load_per_task for nice 0
tasks).

With the above change, imbalance was getting reset to 0 in the corner
case condition, making the FUZZ logic fail. Fix it by not corrupting the
imbalance and change the imbalance, only when it finds that the HT/MC
optimization is needed.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years ago[XFS] fix nasty quota hashtable allocation bug
Eric Sandeen [Thu, 16 Aug 2007 06:49:11 +0000 (16:49 +1000)]
[XFS] fix nasty quota hashtable allocation bug

This git mod: 77e4635ae191774526ed695482a151ac986f3806
converted to a "greedy" allocation interface, but for the quota hashtables
it switched from allocating XFS_QM_HASHSIZE (nr of elements)
xfs_dqhash_t's to allocating only XFS_QM_HASHSIZE *bytes* - quite a lot
smaller! Then when we converted hsize "back" to nr of elements (the
division line) hsize went to 0. This was leading to oopses when running
any quota tests on the Fedora 8 test kernel, but the problem has been
there for almost a year.

SGI-PV: 968837
SGI-Modid: xfs-linux-melb:xfs-kern:29354a

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[XFS] fix sparse shadowed variable warnings
Christoph Hellwig [Thu, 16 Aug 2007 05:38:19 +0000 (15:38 +1000)]
[XFS] fix sparse shadowed variable warnings

- in xfs_probe_cluster rename the inner len to pg_len. There's no harm
  here because the outer len isn't used after the inner len comes into
  existence but it keeps the code clean.
- in xfs_da_do_buf remove the inner i because they don't overlap
  and they are both the same type.

SGI-PV: 968555
SGI-Modid: xfs-linux-melb:xfs-kern:29311a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[XFS] fix ASSERT and ASSERT_ALWAYS
Christoph Hellwig [Thu, 16 Aug 2007 05:38:08 +0000 (15:38 +1000)]
[XFS] fix ASSERT and ASSERT_ALWAYS

- remove the != 0 inside the unlikely in ASSERT_ALWAYS because sparse now
  complains about comparisons between pointers and 0
- add a standalone ASSERT implementation because defining it to
  ASSERT_ALWAYS means the string is expanded before the token passing
  stringification. This way we get the actual content of the
  assertion in the assfail message and don't overflow sparse's
  stringification buffer leading to sparse error messages.

SGI-PV: 968555
SGI-Modid: xfs-linux-melb:xfs-kern:29310a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[XFS] Fix sparse warning in kmem_shake_allow
Christoph Hellwig [Thu, 16 Aug 2007 05:37:57 +0000 (15:37 +1000)]
[XFS] Fix sparse warning in kmem_shake_allow

We can't return a masked result of a __bitwise type. Compare it to 0 first
to keep the behaviour without the warning.

SGI-PV: 968555
SGI-Modid: xfs-linux-melb:xfs-kern:29309a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[XFS] Fix sparse NULL vs 0 warnings
Christoph Hellwig [Thu, 16 Aug 2007 05:37:36 +0000 (15:37 +1000)]
[XFS] Fix sparse NULL vs 0 warnings

Sparse now warns about comparing pointers to 0, so change all instance
where that happens to NULL instead.

SGI-PV: 968555
SGI-Modid: xfs-linux-melb:xfs-kern:29308a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[XFS] Set filestreams object timeout to something sane.
David Chinner [Thu, 16 Aug 2007 05:20:56 +0000 (15:20 +1000)]
[XFS] Set filestreams object timeout to something sane.

SGI-PV: 968554
SGI-Modid: xfs-linux-melb:xfs-kern:29303a

Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
17 years ago[x86 setup] Work around bug in Xen HVM
Christian Ehrhardt [Mon, 3 Sep 2007 18:32:38 +0000 (20:32 +0200)]
[x86 setup] Work around bug in Xen HVM

Apparently XEN does not keep the contents of the 48-bit gdt_48 data
structure that is passed to lgdt in the XEN machine state. Instead it
appears to save the _address_ of the 48-bit descriptor
somewhere. Unfortunately this data happens to reside on the stack and
is probably no longer availiable at the time of the actual protected
mode jump.

This is Xen bug but given that there is a one-line patch to work
around this problem, the linux kernel should probably do this.  My fix
is to make the gdt_48 description in setup_gdt static (in setup_idt
this is already the case). This allows the kernel to boot under
Xen HVM again.

Signed-off-by: Christian Ehrhardt <lk@c--e.de>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 4 Sep 2007 08:04:43 +0000 (01:04 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

17 years agoMerge branch 'for_linus' of git://git.linux-nfs.org/pub/linux/nfs-2.6
Linus Torvalds [Tue, 4 Sep 2007 07:45:54 +0000 (00:45 -0700)]
Merge branch 'for_linus' of git://git.linux-nfs.org/pub/linux/nfs-2.6

17 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Tue, 4 Sep 2007 07:43:57 +0000 (00:43 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild-fix
Linus Torvalds [Tue, 4 Sep 2007 07:42:41 +0000 (00:42 -0700)]
Merge git://git./linux/kernel/git/sam/kbuild-fix

17 years agoMerge git://git.infradead.org/~dwmw2/mtd-2.6.23
Linus Torvalds [Tue, 4 Sep 2007 07:39:50 +0000 (00:39 -0700)]
Merge git://git.infradead.org/~dwmw2/mtd-2.6.23

17 years agom32r: Separate syscall table from entry.S
Hirokazu Takata [Mon, 20 Aug 2007 00:12:46 +0000 (09:12 +0900)]
m32r: Separate syscall table from entry.S

- Separate sys_call_table from arch/m32r/kernel/entry.S and
  move it to arch/m32r/kernel/system_call.S.
- Change sys_call_table section from .data to .rodata.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Cosmetic updates of arch/m32r/kernel/entry.S
Hirokazu Takata [Fri, 17 Aug 2007 15:10:18 +0000 (00:10 +0900)]
m32r: Cosmetic updates of arch/m32r/kernel/entry.S

- Remove unused symbols *_MASK
- Change indentation of comments, etc.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Exit ei_handler directly for no IRQ case or IPI operations
Hirokazu Takata [Fri, 17 Aug 2007 14:40:37 +0000 (23:40 +0900)]
m32r: Exit ei_handler directly for no IRQ case or IPI operations

If no IRQ request is found in the IRQ check of ei_handler,
we can exit directly by jumping "restore_all", instead of via
"ret_from_intr".

This modification is also likely effective for IPI operations,
because scheduler call never happen at the exit of IPIs.

Signed-off-by: Hitoshi Yamamoto <hitoshiy@linux-m32r.org>
Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Simplify ei_handler code
Hirokazu Takata [Fri, 17 Aug 2007 09:11:37 +0000 (18:11 +0900)]
m32r: Simplify ei_handler code

Simplify and clean up messy ei_handler code in arch/m32r/kernel/entry.S.
- Remove ifdef's for CONFIG_CHIP_* configulations.
- Rearrange the M32700 workaround code.
- Remove the messy platform-dependent interrupt check routines and
  consolidate them to common INT0/INT1/INT2 check routines for all
  platforms with cascaded interrupt controllers.

Signed-off-by: Hitoshi Yamamoto <hitoshiy@linux-m32r.org>
Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Define symbols to unify platform-dependent ICU checks
Hirokazu Takata [Fri, 17 Aug 2007 08:22:15 +0000 (17:22 +0900)]
m32r: Define symbols to unify platform-dependent ICU checks

On some m32r platforms, cascaded ICUs are used.
This patch is required to simplify ei_handler and consolidate platform-
dependent ICU check routines.

  platform       ICU/INT1  ICU/INT0  ICU/INT2
 --------------  --------  --------  --------
  m32104ut        o         -         -
  m32700ut        o         o         o
  opsput          o         o         o
  usrv            o         -         -
  (others)        -         -         -

Signed-off-by: Hitoshi Yamamoto <hitoshiy@linux-m32r.org>
Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Move dot.gdbinit files
Hirokazu Takata [Wed, 1 Aug 2007 12:10:11 +0000 (21:10 +0900)]
m32r: Move dot.gdbinit files

Move dot.gdbinit files from arch/m32r/{platforms}/dot.gdbinit*
to arch/m32r/platforms/{platform}/.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Rearrange platform-dependent codes
Hirokazu Takata [Wed, 1 Aug 2007 12:09:31 +0000 (21:09 +0900)]
m32r: Rearrange platform-dependent codes

Rearrange platform-dependent codes from arch/m32r/kernel/*.c
to arch/m32r/platforms/{platform}/.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Add defconfig file for the usrv platform.
Hirokazu Takata [Wed, 1 Aug 2007 12:00:46 +0000 (21:00 +0900)]
m32r: Add defconfig file for the usrv platform.

Add usrv_defcofig file for the usrv (M32R MicroServer) platform.

  platform    defconfig                 Note
 ----------  ----------------------    ---------------------------
  usrv        usrv_defconfig            SMP

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Update defconfig files for 2.6.23-rc1
Hirokazu Takata [Wed, 1 Aug 2007 11:59:20 +0000 (20:59 +0900)]
m32r: Update defconfig files for 2.6.23-rc1

Update defconfig files for 2.6.23-rc1 in arch/m32r/configs/.

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years agom32r: Move defconfig files to arch/m32r/configs/
Hirokazu Takata [Wed, 1 Aug 2007 11:54:38 +0000 (20:54 +0900)]
m32r: Move defconfig files to arch/m32r/configs/

Change defconfig file's location from arch/m32r/{platform}/defconfig*
to arch/m32r/configs/{platform}_defconfig.

Applying this patch, we can use defconfig file for each m32r platform
easily, like other architectures.

  ex. Setup defconfig for cross-building
    $ make ARCH=m32r CROSS_COMPILE=m32r-linux-gnu- {platform}_defconfig.

  platform    defconfig                 Note
 ----------  ----------------------    ---------------------------
  m32104ut    m32104ut_defconfig        MMU-less
  m32700ut    m32700ut.smp_defconfig    SMP
  m32700ut    m32700ut.up_defconfig     UP
  mappi       mappi.smp_defconfig       SMP
  mappi       mappi.up_defconfig        UP
  mappi       mappi.nommu_defconfig     MMU-less
  mappi2      mappi2.opsp_defconfig     FPGA env. (CPU Core: OPSP)
  mappi2      mappi2.vdec2_defconfig    FPGA env. (CPU Core: VDEC2)
  mappi3      imappi3.smp_defconfig     SMP
  oaks32r     oaks32r_defconfig         MMU-less
  opsput      opsput_defconfig          UP

Signed-off-by: Hirokazu Takata <takata@linux-m32r.org>
17 years ago[MTD] Initialise s_flags in get_sb_mtd_aux()
David Howells [Thu, 30 Aug 2007 16:21:19 +0000 (17:21 +0100)]
[MTD] Initialise s_flags in get_sb_mtd_aux()

Initialise s_flags in get_sb_mtd_aux() from the flags parameter.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
17 years ago[JFFS2] fix write deadlock regression
Jason Lunz [Sat, 1 Sep 2007 19:06:03 +0000 (12:06 -0700)]
[JFFS2] fix write deadlock regression

I've bisected the deadlock when many small appends are done on jffs2 down to
this commit:

commit 6fe6900e1e5b6fa9e5c59aa5061f244fe3f467e2
Author: Nick Piggin <npiggin@suse.de>
Date:   Sun May 6 14:49:04 2007 -0700

    mm: make read_cache_page synchronous

    Ensure pages are uptodate after returning from read_cache_page, which allows
    us to cut out most of the filesystem-internal PageUptodate calls.

    I didn't have a great look down the call chains, but this appears to fixes 7
    possible use-before uptodate in hfs, 2 in hfsplus, 1 in jfs, a few in
    ecryptfs, 1 in jffs2, and a possible cleared data overwritten with readpage in
    block2mtd.  All depending on whether the filler is async and/or can return
    with a !uptodate page.

It introduced a wait to read_cache_page, as well as a
read_cache_page_async function equivalent to the old read_cache_page
without any callers.

Switching jffs2_gc_fetch_page to read_cache_page_async for the old
behavior makes the deadlocks go away, but maybe reintroduces the
use-before-uptodate problem? I don't understand the mm/fs interaction
well enough to say.

[It's fine. dwmw2.]

Signed-off-by: Jason Lunz <lunz@falooley.org>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
17 years agoNFS: Fix a write request leak in nfs_invalidate_page()
Trond Myklebust [Tue, 28 Aug 2007 14:29:36 +0000 (10:29 -0400)]
NFS: Fix a write request leak in nfs_invalidate_page()

Ryusuke Konishi says:

The recent truncate_complete_page() clears the dirty flag from a page
before calling a_ops->invalidatepage(),
^^^^^^
static void
truncate_complete_page(struct address_space *mapping, struct page *page)
{
        ...
        cancel_dirty_page(page, PAGE_CACHE_SIZE);  <--- Inserted here at
kernel 2.6.20

        if (PagePrivate(page))
                do_invalidatepage(page, 0);   ---> will call
a_ops->invalidatepage()
        ...
}

and this is disturbing nfs_wb_page_priority() from calling
nfs_writepage_locked() that is expected to handle the pending
request (=nfs_page) associated with the page.

int nfs_wb_page_priority(struct inode *inode, struct page *page, int how)
{
        ...
        if (clear_page_dirty_for_io(page)) {
                ret = nfs_writepage_locked(page, &wbc);
                if (ret < 0)
                        goto out;
        }
        ...
}

Since truncate_complete_page() will get rid of the page after
a_ops->invalidatepage() returns, the request (=nfs_page) associated
with the page becomes a garbage in nfs_inode->nfs_page_tree.
------------------------

Fix this by ensuring that nfs_wb_page_priority() recognises that it may
also need to clear out non-dirty pages that have an nfs_page associated
with them.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: change NFS mount error return when hostname/pathname too long
Chuck Lever [Wed, 29 Aug 2007 21:59:03 +0000 (17:59 -0400)]
NFS: change NFS mount error return when hostname/pathname too long

According to the mount(2) man page, the proper error return code for the
mount(2) system call when the special device name or the mounted-on
directory name is too long is ENAMETOOLONG.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Off-by-one length error in string handling
Chuck Lever [Wed, 29 Aug 2007 21:59:01 +0000 (17:59 -0400)]
NFS: Off-by-one length error in string handling

The hostname was getting truncated in the new text-based NFS mount API.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Return a real error code from mount(2)
Chuck Lever [Wed, 29 Aug 2007 21:58:59 +0000 (17:58 -0400)]
NFS: Return a real error code from mount(2)

Don't filter the return code from the in-kernel rpcbind or NFS mount
clients.  Return the real error code so that callers of the new NFS
text-based mount API can apply a useful retry strategy.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>