openwrt/staging/blogic.git
14 years agopowerpc/fsl-booke: fix the case where we are not in the first page
Sebastian Andrzej Siewior [Sun, 4 Apr 2010 20:19:01 +0000 (22:19 +0200)]
powerpc/fsl-booke: fix the case where we are not in the first page

During boot we change the mapping a few times until we have a "defined"
mapping. During this procedure a small 4KiB mapping is created and after
that one a 64MiB. Currently the offset of the 4KiB page in that we run
is zero because the complete startup up code is in first page which
starts at RPN zero.
If the code is recycled and moved to another location then its execution
will fail because the start address in 64 MiB mapping is computed
wrongly. It does not consider the offset to the page from the begin of
the memory.
This patch fixes this. Usually (system boot) r25 is zero so this does
not change anything unless the code is recycled.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/85xx: Enable support for ports 3 and 4 on 8548 CDS
Andy Fleming [Tue, 30 Mar 2010 01:39:08 +0000 (20:39 -0500)]
powerpc/85xx: Enable support for ports 3 and 4 on 8548 CDS

I believe support was disabled due to issues with earlier versions of
the board/processor.  At worst, adding the ports back into the device
tree should result in enabling ports that don't work on older systems,
so the default should be to enable them.

Signed-off-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/fsl-booke: Add hibernation support for FSL BookE processors
Anton Vorontsov [Mon, 17 May 2010 18:56:52 +0000 (22:56 +0400)]
powerpc/fsl-booke: Add hibernation support for FSL BookE processors

This is started as swsusp_32.S modifications, but the amount of #ifdefs
made the whole file horribly unreadable, so let's put the support into
its own separate file.

The code should be relatively easy to modify to support 44x BookEs as
well, but since I don't have any 44x to test, let's confine the code to
FSL BookE. (The only FSL-specific part so far is 'flush_dcache_L1'.)

Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Acked-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/e500mc: Implement machine check handler.
Scott Wood [Thu, 8 Apr 2010 05:38:22 +0000 (00:38 -0500)]
powerpc/e500mc: Implement machine check handler.

Most of the MSCR bit assigments are different in e500mc versus
e500, and they are now write-one-to-clear.

Some e500mc machine check conditions are made recoverable (as long as
they aren't stuck on), most notably L1 instruction cache parity errors.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc: Remove unused 'protect4gb' boot parameter
FUJITA Tomonori [Sun, 9 May 2010 17:39:05 +0000 (17:39 +0000)]
powerpc: Remove unused 'protect4gb' boot parameter

'protect4gb' boot parameter was introduced to avoid allocating dma
space acrossing 4GB boundary in 2007 (the commit
569975591c5530fdc9c7a3c45122e5e46f075a74).

In 2008, the IOMMU was fixed to use the boundary_mask parameter per
device properly. So 'protect4gb' workaround was removed (the
383af9525bb27f927511874f6306247ec13f1c28). But somehow I messed the
'protect4gb' boot parameter that was used to enable the
workaround.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Build-in e1000e for pseries & ppc64_defconfig
Michael Neuling [Sun, 18 Apr 2010 20:05:41 +0000 (20:05 +0000)]
powerpc: Build-in e1000e for pseries & ppc64_defconfig

The e1000e device is becoming more common these days, so let's just
build it in for pseries & ppc64_defconfig.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/pseries: Make request_ras_irqs() available to other pseries code
Mark Nelson [Tue, 18 May 2010 22:51:00 +0000 (22:51 +0000)]
powerpc/pseries: Make request_ras_irqs() available to other pseries code

At the moment only the RAS code uses event-sources interrupts (for EPOW
events and internal errors) so request_ras_irqs() (which actually requests
the event-sources interrupts) is found in ras.c and is static.

We want to be able to use event-sources interrupts in other pseries code,
so let's rename request_ras_irqs() to request_event_sources_irqs() and
move it to event_sources.c.

This will be used in an upcoming patch that adds support for IO Event
interrupts that come through as event sources.

Signed-off-by: Mark Nelson <markn@au1.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/numa: Use ibm,architecture-vec-5 to detect form 1 affinity
Anton Blanchard [Sun, 16 May 2010 20:28:35 +0000 (20:28 +0000)]
powerpc/numa: Use ibm,architecture-vec-5 to detect form 1 affinity

I've been told that the architected way to determine we are in form 1
affinity mode is by reading the ibm,architecture-vec-5 property which
mirrors the layout of the fifth vector of the ibm,client-architecture
structure.

Eventually we may want to parse the ibm,architecture-vec-5 and create
FW_FEATURE_* bits.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/numa: Set a smaller value for RECLAIM_DISTANCE to enable zone reclaim
Anton Blanchard [Sun, 16 May 2010 20:19:56 +0000 (20:19 +0000)]
powerpc/numa: Set a smaller value for RECLAIM_DISTANCE to enable zone reclaim

I noticed /proc/sys/vm/zone_reclaim_mode was 0 on a ppc64 NUMA box. It gets
enabled via this:

        /*
         * If another node is sufficiently far away then it is better
         * to reclaim pages in a zone before going off node.
         */
        if (distance > RECLAIM_DISTANCE)
                zone_reclaim_mode = 1;

Since we use the default value of 20 for REMOTE_DISTANCE and 20 for
RECLAIM_DISTANCE it never kicks in.

The local to remote bandwidth ratios can be quite large on System p
machines so it makes sense for us to reclaim clean pagecache locally before
going off node.

The patch below sets a smaller value for RECLAIM_DISTANCE and thus enables
zone reclaim.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Use smt_snooze_delay=-1 to always busy loop
Anton Blanchard [Sun, 16 May 2010 20:02:39 +0000 (20:02 +0000)]
powerpc: Use smt_snooze_delay=-1 to always busy loop

Right now if we want to busy loop and not give up any time to the hypervisor
we put a very large value into smt_snooze_delay. This is sometimes useful
when running a single partition and you want to avoid any latencies due
to the hypervisor or CPU power state transitions. While this works, it's a bit
ugly - how big a number is enough now we have NO_HZ and can be idle for a very
long time.

The patch below makes smt_snooze_delay signed, and a negative value means loop
forever:

echo -1 > /sys/devices/system/cpu/cpu0/smt_snooze_delay

This change shouldn't affect the existing userspace tools (eg ppc64_cpu), but
I'm cc-ing Nathan just to be sure.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Remove check of ibm,smt-snooze-delay OF property
Anton Blanchard [Sun, 16 May 2010 20:01:28 +0000 (20:01 +0000)]
powerpc: Remove check of ibm,smt-snooze-delay OF property

I'm not sure why we have code for parsing an ibm,smt-snooze-delay OF
property. Since we have a smt-snooze-delay= boot option and we can
also set it at runtime via sysfs, it should be safe to get rid of
this code.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/kdump: Fix race in kdump shutdown
Michael Neuling [Thu, 13 May 2010 19:40:11 +0000 (19:40 +0000)]
powerpc/kdump: Fix race in kdump shutdown

When we are crashing, the crashing/primary CPU IPIs the secondaries to
turn off IRQs, go into real mode and wait in kexec_wait.  While this
is happening, the primary tears down all the MMU maps.  Unfortunately
the primary doesn't check to make sure the secondaries have entered
real mode before doing this.

On PHYP machines, the secondaries can take a long time shutting down
the IRQ controller as RTAS calls are need.  These RTAS calls need to
be serialised which resilts in the secondaries contending in
lock_rtas() and hence taking a long time to shut down.

We've hit this on large POWER7 machines, where some secondaries are
still waiting in lock_rtas(), when the primary tears down the HPTEs.

This patch makes sure all secondaries are in real mode before the
primary tears down the MMU.  It uses the new kexec_state entry in the
paca.  It times out if the secondaries don't reach real mode after
10sec.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/kexec: Fix race in kexec shutdown
Michael Neuling [Thu, 13 May 2010 19:40:11 +0000 (19:40 +0000)]
powerpc/kexec: Fix race in kexec shutdown

In kexec_prepare_cpus, the primary CPU IPIs the secondary CPUs to
kexec_smp_down().  kexec_smp_down() calls kexec_smp_wait() which sets
the hw_cpu_id() to -1.  The primary does this while leaving IRQs on
which means the primary can take a timer interrupt which can lead to
the IPIing one of the secondary CPUs (say, for a scheduler re-balance)
but since the secondary CPU now has a hw_cpu_id = -1, we IPI CPU
-1... Kaboom!

We are hitting this case regularly on POWER7 machines.

There is also a second race, where the primary will tear down the MMU
mappings before knowing the secondaries have entered real mode.

Also, the secondaries are clearing out any pending IPIs before
guaranteeing that no more will be received.

This changes kexec_prepare_cpus() so that we turn off IRQs in the
primary CPU much earlier.  It adds a paca flag to say that the
secondaries have entered the kexec_smp_down() IPI and turned off IRQs,
rather than overloading hw_cpu_id with -1.  This new paca flag is
again used to in indicate when the secondaries has entered real mode.

It also ensures that all CPUs have their IRQs off before we clear out
any pending IPI requests (in kexec_cpu_down()) to ensure there are no
trailing IPIs left unacknowledged.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/kexec: Speedup kexec hash PTE tear down
Michael Neuling [Mon, 10 May 2010 20:28:26 +0000 (20:28 +0000)]
powerpc/kexec: Speedup kexec hash PTE tear down

Currently for kexec the PTE tear down on 1TB segment systems normally
requires 3 hcalls for each PTE removal. On a machine with 32GB of
memory it can take around a minute to remove all the PTEs.

This optimises the path so that we only remove PTEs that are valid.
It also uses the read 4 PTEs at once HCALL.  For the common case where
a PTEs is invalid in a 1TB segment, this turns the 3 HCALLs per PTE
down to 1 HCALL per 4 PTEs.

This gives an > 10x speedup in kexec times on PHYP, taking a 32GB
machine from around 1 minute down to a few seconds.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/pseries: Add hcall to read 4 ptes at a time in real mode
Michael Neuling [Mon, 10 May 2010 20:28:26 +0000 (20:28 +0000)]
powerpc/pseries: Add hcall to read 4 ptes at a time in real mode

This adds plpar_pte_read_4_raw() which can be used read 4 PTEs from
PHYP at a time, while in real mode.

It also creates a new hcall9 which can be used in real mode.  It's the
same as plpar_hcall9 but minus the tracing hcall statistics which may
require variables outside the RMO.

Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Use more accurate limit for first segment memory allocations
Anton Blanchard [Mon, 10 May 2010 18:59:18 +0000 (18:59 +0000)]
powerpc: Use more accurate limit for first segment memory allocations

Author: Milton Miller <miltonm@bga.com>

On large machines we are running out of room below 256MB. In some cases we
only need to ensure the allocation is in the first segment, which may be
256MB or 1TB.

Add slb0_limit and use it to specify the upper limit for the irqstack and
emergency stacks.

On a large ppc64 box, this fixes a panic at boot when the crashkernel=
option is specified (previously we would run out of memory below 256MB).

Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/kdump: Use chip->shutdown to disable IRQs
Anton Blanchard [Mon, 10 May 2010 16:27:38 +0000 (16:27 +0000)]
powerpc/kdump: Use chip->shutdown to disable IRQs

I saw this in a kdump kernel:

IOMMU table initialized, virtual merging enabled
Interrupt 155954 (real) is invalid, disabling it.
Interrupt 155953 (real) is invalid, disabling it.

ie we took some spurious interrupts. default_machine_crash_shutdown tries
to disable all interrupt sources but uses chip->disable which maps to
the default action of:

static void default_disable(unsigned int irq)
{
}

If we use chip->shutdown, then we actually mask the IRQ:

static void default_shutdown(unsigned int irq)
{
        struct irq_desc *desc = irq_to_desc(irq);

        desc->chip->mask(irq);
        desc->status |= IRQ_MASKED;
}

Not sure why we don't implement a ->disable action for xics.c, or why
default_disable doesn't mask the interrupt.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/kdump: CPUs assume the context of the oopsing CPU
Anton Blanchard [Mon, 10 May 2010 16:25:51 +0000 (16:25 +0000)]
powerpc/kdump: CPUs assume the context of the oopsing CPU

We wrap the crash_shutdown_handles[] calls with longjmp/setjmp, so if any
of them fault we can recover. The problem is we add a hook to the debugger
fault handler hook which calls longjmp unconditionally.

This first part of kdump is run before we marshall the other CPUs, so there
is a very good chance some CPU on the box is going to page fault. And when
it does it hits the longjmp code and assumes the context of the oopsing CPU.
The machine gets very confused when it has 10 CPUs all with the same stack,
all thinking they have the same CPU id. I get even more confused trying
to debug it.

The patch below adds crash_shutdown_cpu and uses it to specify which cpu is
in the protected region. Since it can only be -1 or the oopsing CPU, we don't
need to use memory barriers since it is only valid on the local CPU - no other
CPU will ever see a value that matches it's local CPU id.

Eventually we should switch the order and marshall all CPUs before doing the
crash_shutdown_handles[] calls, but that is a bigger fix.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/crashdump: Do not fail on NULL pointer dereferencing
Maxim Uvarov [Tue, 11 May 2010 05:41:08 +0000 (05:41 +0000)]
powerpc/crashdump: Do not fail on NULL pointer dereferencing

Signed-off-by: Maxim Uvarov <muvarov@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/eeh: Fix oops when probing in early boot
Anton Blanchard [Mon, 10 May 2010 15:38:56 +0000 (15:38 +0000)]
powerpc/eeh: Fix oops when probing in early boot

If we take an EEH error early enough, we oops:

Call Trace:
[c000000010483770] [c000000000013ee4] .show_stack+0xd8/0x218 (unreliable)
[c000000010483850] [c000000000658940] .dump_stack+0x28/0x3c
[c0000000104838d0] [c000000000057a68] .eeh_dn_check_failure+0x2b8/0x304
[c000000010483990] [c0000000000259c8] .rtas_read_config+0x120/0x168
[c000000010483a40] [c000000000025af4] .rtas_pci_read_config+0xe4/0x124
[c000000010483af0] [c00000000037af18] .pci_bus_read_config_word+0xac/0x104
[c000000010483bc0] [c0000000008fec98] .pcibios_allocate_resources+0x7c/0x220
[c000000010483c90] [c0000000008feed8] .pcibios_resource_survey+0x9c/0x418
[c000000010483d80] [c0000000008fea10] .pcibios_init+0xbc/0xf4
[c000000010483e20] [c000000000009844] .do_one_initcall+0x98/0x1d8
[c000000010483ed0] [c0000000008f0560] .kernel_init+0x228/0x2e8
[c000000010483f90] [c000000000031a08] .kernel_thread+0x54/0x70
EEH: Detected PCI bus error on device <null>
EEH: This PCI device has failed 1 times in the last hour:
EEH: location=U78A5.001.WIH8464-P1 driver= pci addr=0001:00:01.0
EEH: of node=/pci@800000020000209/usb@1
EEH: PCI device/vendor: 00351033
EEH: PCI cmd/status register: 12100146

Unable to handle kernel paging request for data at address 0x00000468
Oops: Kernel access of bad area, sig: 11 [#1]
....
NIP [c000000000057610] .rtas_set_slot_reset+0x38/0x10c
LR [c000000000058724] .eeh_reset_device+0x5c/0x124
Call Trace:
[c00000000bc6bd00] [c00000000005a0e0] .pcibios_remove_pci_devices+0x7c/0xb0 (unreliable)
[c00000000bc6bd90] [c000000000058724] .eeh_reset_device+0x5c/0x124
[c00000000bc6be40] [c0000000000589c0] .handle_eeh_events+0x1d4/0x39c
[c00000000bc6bf00] [c000000000059124] .eeh_event_handler+0xf0/0x188
[c00000000bc6bf90] [c000000000031a08] .kernel_thread+0x54/0x70

We called rtas_set_slot_reset while scanning the bus and before the pci_dn
to pcidev mapping has been created. Since we only need the pcidev to work
out the type of reset and that only gets set after the module for the
device loads, lets just do a hot reset if the pcidev is NULL.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Linas Vepstas <linasvepstas@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/pci: Check devices status property when scanning OF tree
Sonny Rao [Mon, 10 May 2010 15:13:41 +0000 (15:13 +0000)]
powerpc/pci: Check devices status property when scanning OF tree

We ran into an issue where it looks like we're not properly ignoring a
pci device with a non-good status property when we walk the device tree
and instanciate the Linux side PCI devices.

However, the EEH init code does look for the property and disables EEH
on these devices. This leaves us in an inconsistent where we are poking
at a supposedly bad piece of hardware and RTAS will block our config
cycles because EEH isn't enabled anyway.

Signed-of-by: Sonny Rao <sonnyrao@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/vio: Switch VIO Bus PM to use generic helpers
Brian King [Fri, 14 May 2010 12:04:41 +0000 (12:04 +0000)]
powerpc/vio: Switch VIO Bus PM to use generic helpers

Switch to use the generic power management helpers.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Avoid bad relocations in iSeries code
Paul Mackerras [Thu, 6 May 2010 19:46:50 +0000 (19:46 +0000)]
powerpc: Avoid bad relocations in iSeries code

Subrata Modak reported that building a CONFIG_RELOCATABLE kernel with
CONFIG_ISERIES enabled gives the following warnings:

WARNING: 4 bad relocations
c00000000007216e R_PPC64_ADDR16_HIGHEST  __ksymtab+0x00000000009dcec8
c000000000072172 R_PPC64_ADDR16_HIGHER  __ksymtab+0x00000000009dcec8
c00000000007217a R_PPC64_ADDR16_HI  __ksymtab+0x00000000009dcec8
c00000000007217e R_PPC64_ADDR16_LO  __ksymtab+0x00000000009dcec8

The reason is that decrementer_iSeries_masked is using
LOAD_REG_IMMEDIATE to get the address of a kernel symbol, which
creates relocations that aren't handled by the kernel relocator code.

Instead of reading the tb_ticks_per_jiffy variable, we can just set
the decrementer to its maximum value (0x7fffffff) and that will work
just as well.  In fact timer_interrupt sets the decrementer to that
value initially anyway, and we are sure to get into timer_interrupt
once interrupts are reenabled because we store 1 to the decrementer
interrupt flag in the lppaca (LPPACADECRINT(r12) here).

Reported-by: Subrata Modak <subrata@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Use common cpu_die (fixes SMP+SUSPEND build)
Milton Miller [Wed, 19 May 2010 02:56:29 +0000 (02:56 +0000)]
powerpc: Use common cpu_die (fixes SMP+SUSPEND build)

Configuring a powerpc 32 bit kernel for both SMP and SUSPEND turns on
CPU_HOTPLUG to enable disable_nonboot_cpus to be called by the common
suspend code.  Previously the definition of cpu_die for ppc32 was in
the powermac platform code, causing it to be undefined if that platform
as not selected.

arch/powerpc/kernel/built-in.o: In function 'cpu_idle':
arch/powerpc/kernel/idle.c:98: undefined reference to 'cpu_die'

Move the code from setup_64 to smp.c and rename the power mac
versions to their specific names.

Note that this does not setup the cpu_die pointers in either
smp_ops (request a given cpu die) or ppc_md (make this cpu die),
for other platforms but there are generic versions in smp.c.

Reported-by: Matt Sealey <matt@genesi-usa.com>
Reported-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Fix string library functions
Andreas Schwab [Tue, 18 May 2010 08:15:21 +0000 (08:15 +0000)]
powerpc: Fix string library functions

The powerpc strncmp implementation does not correctly handle a zero
length, despite the claim in 0119536cd314ef95553604208c25bc35581f7f0a
(Add hand-coded assembly strcmp).

Additionally, all the length arguments are size_t, not int, so use
PPC_LCMPI and eq instead of cmpwi and le throughout.

Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/rtasd: Don't start event scan if scan rate is zero
Michael Ellerman [Wed, 19 May 2010 02:12:32 +0000 (02:12 +0000)]
powerpc/rtasd: Don't start event scan if scan rate is zero

There appear to be Pegasos systems which have the rtas-event-scan
RTAS tokens, but on which the event scan always fails. They also
have an event-scan-rate property containing 0, which means call
event scan 0 times per minute.

So interpret a scan rate of 0 to mean don't scan at all. This fixes
the problem on the Pegasos machines and makes sense as well.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/fsl-booke: Move loadcam_entry back to asm code to fix SMP ftrace
Kumar Gala [Thu, 13 May 2010 19:38:21 +0000 (14:38 -0500)]
powerpc/fsl-booke: Move loadcam_entry back to asm code to fix SMP ftrace

When we build with ftrace enabled its possible that loadcam_entry would
have used the stack pointer (even though the code doesn't need it).  We
call loadcam_entry in __secondary_start before the stack is setup.  To
ensure that loadcam_entry doesn't use the stack pointer the easiest
solution is to just have it in asm code.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/fsl-booke: Fix InstructionTLBError execute permission check
Li Yang [Fri, 7 May 2010 08:38:34 +0000 (16:38 +0800)]
powerpc/fsl-booke: Fix InstructionTLBError execute permission check

In CONFIG_PTE_64BIT the PTE format has unique permission bits for user
and supervisor execute.  However on !CONFIG_PTE_64BIT we overload the
supervisor bit to imply user execute with _PAGE_USER set.  This allows
us to use the same permission check mask for user or supervisor code on
!CONFIG_PTE_64BIT.

However, on CONFIG_PTE_64BIT we map _PAGE_EXEC to _PAGE_BAP_UX so we
need a different permission mask based on the fault coming from a kernel
address or user space.

Without unique permission masks we see issues like the following with
modules:

Unable to handle kernel paging request for instruction fetch
Faulting instruction address: 0xf938d040
Oops: Kernel access of bad area, sig: 11 [#1]

Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Jin Qing <b24347@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/8610: add probing for individual DMA channels, not just DMA controllers
Timur Tabi [Mon, 3 May 2010 21:54:15 +0000 (16:54 -0500)]
powerpc/8610: add probing for individual DMA channels, not just DMA controllers

A future version of the MPC8610 HPCD's ASoC DMA driver will probe on individual
DMA channel nodes, so the DMA controller nodes' compatible string must be listed
in mpc8610_ids[] for the probe to work.

Also remove the "gianfar" compatible from mpc8610_ids[], since there is no
gianfar (or any other networking device) on the 8610.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/83xx: Add MCU LEDs support for MPC837xRDB and MPC8315RDB boards
Anton Vorontsov [Sun, 2 May 2010 17:27:05 +0000 (21:27 +0400)]
powerpc/83xx: Add MCU LEDs support for MPC837xRDB and MPC8315RDB boards

There are two front-panel LEDs on MPC837xRDB and MPC8315RDB boards: PWR
and HDD. After adding appropriate nodes we can program these LEDs from
kernel and user space.

Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/85xx: Fix P1020RDB boot hang due USB2
Anton Vorontsov [Thu, 22 Apr 2010 15:44:47 +0000 (19:44 +0400)]
powerpc/85xx: Fix P1020RDB boot hang due USB2

Since USB2 is shared with local bus, either local bus or USB2
should be disabled. By default U-Boot enables local bus, so we
have to disable USB2, otherwise kernel hangs:

 ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
 fsl-ehci fsl-ehci.0: Freescale On-Chip EHCI Host Controller
 fsl-ehci fsl-ehci.0: new USB bus registered, assigned bus number 1
 fsl-ehci fsl-ehci.0: irq 28, io base 0xffe22000
 fsl-ehci fsl-ehci.0: USB 2.0 started, EHCI 1.00
 hub 1-0:1.0: USB hub found
 hub 1-0:1.0: 1 port detected
 fsl-ehci fsl-ehci.1: Freescale On-Chip EHCI Host Controller
 fsl-ehci fsl-ehci.1: new USB bus registered, assigned bus number 2
 <hangs here>

Note that U-Boot doesn't clear 'status' property when it enables
USB2, so we have to comment out the whole node.

To enable USB2, one can issue
'setenv hwconfig usb2:dr_mode=<host|peripheral>' command at the
U-Boot prompt.

Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc/85xx: Add eTSEC 2.0 support for P1020RDB boards
Anton Vorontsov [Thu, 15 Apr 2010 18:36:31 +0000 (22:36 +0400)]
powerpc/85xx: Add eTSEC 2.0 support for P1020RDB boards

This patch adds support for eTSEC 2.0 as found in P1020.
The changes include introduction of the group nodes for
the etsec nodes.

Signed-off-by: Sandeep Gopalpet <sandeep.kumar@freescale.com>
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc: remove tls_ssl_stream descriptor type capability in sec3.3 node
Kim Phillips [Fri, 23 Apr 2010 00:25:13 +0000 (19:25 -0500)]
powerpc: remove tls_ssl_stream descriptor type capability in sec3.3 node

Technically, whilst SEC v3.3 h/w honours the tls_ssl_stream descriptor
type, it lacks the ARC4 algorithm execution unit required to be able
to execute anything meaningful with it.  Change the node to agree with
the documentation that declares that the sec3.3 really doesn't have such
a descriptor type.

Reported-by: Haiying Wang <Haiying.Wang@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agopowerpc: Add interrupt support to mpc8xxx_gpio
Peter Korsgaard [Thu, 7 Jan 2010 16:57:46 +0000 (17:57 +0100)]
powerpc: Add interrupt support to mpc8xxx_gpio

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Acked-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoMerge commit 'origin/master' into next
Benjamin Herrenschmidt [Fri, 7 May 2010 01:29:25 +0000 (11:29 +1000)]
Merge commit 'origin/master' into next

14 years agopowerpc/cpumask: Add DEBUG_PER_CPU_MAPS option
Anton Blanchard [Mon, 26 Apr 2010 15:32:46 +0000 (15:32 +0000)]
powerpc/cpumask: Add DEBUG_PER_CPU_MAPS option

Enable the DEBUG_PER_CPU_MAPS option so we can look for problems with
cpumasks .

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert mpic driver to new cpumask API
Benjamin Herrenschmidt [Thu, 6 May 2010 08:01:46 +0000 (18:01 +1000)]
powerpc/cpumask: Convert mpic driver to new cpumask API

Convert to the new cpumask API.

irq_choose_cpu can be simplified by using cpumask_next and cpumask_first.

smp_mpic_message_pass was doing open coded cpumask manipulation and passing an
int for a cpumask into mpic_send_ipi. Since mpic_send_ipi is only used
locally, make it static and convert it to take a cpumask. This allows us
to clean up the mess in smp_mpic_message_pass.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Update some comments
Anton Blanchard [Mon, 26 Apr 2010 15:32:44 +0000 (15:32 +0000)]
powerpc/cpumask: Update some comments

Since the *_map cpumask variants are deprecated, change the comments to
instead refer to *_mask.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert NUMA code to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:43 +0000 (15:32 +0000)]
powerpc/cpumask: Convert NUMA code to new cpumask API

Convert NUMA code to new cpumask API. We shift the node to cpumask
setup code until after we complete bootmem allocation so we can
dynamically allocate the cpumasks.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert hotplug-cpu code to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:42 +0000 (15:32 +0000)]
powerpc/cpumask: Convert hotplug-cpu code to new cpumask API

Convert hotplug-cpu code to new cpumask API.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Dynamically allocate cpu_sibling_map and cpu_core_map cpumasks
Anton Blanchard [Mon, 26 Apr 2010 15:32:41 +0000 (15:32 +0000)]
powerpc/cpumask: Dynamically allocate cpu_sibling_map and cpu_core_map cpumasks

Dynamically allocate cpu_sibling_map and cpu_core_map cpumasks.

We don't need to set_cpu_online() the boot cpu in smp_prepare_boot_cpu,
init/main.c does it for us.

We also postpone setting of the boot cpu in cpu_sibling_map and cpu_core_map
until when the memory allocator is available (smp_prepare_cpus), similar
to x86.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert /proc/cpuinfo to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:40 +0000 (15:32 +0000)]
powerpc/cpumask: Convert /proc/cpuinfo to new cpumask API

Use new cpumask API in /proc/cpuinfo code.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Refactor /proc/cpuinfo code
Anton Blanchard [Mon, 26 Apr 2010 15:32:39 +0000 (15:32 +0000)]
powerpc/cpumask: Refactor /proc/cpuinfo code

This separates the per cpu output from the summary output at the end of the
file, making it easier to convert to the new cpumask API in a subsequent
patch.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert xics driver to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:38 +0000 (15:32 +0000)]
powerpc/cpumask: Convert xics driver to new cpumask API

Use the new cpumask API and add some comments to clarify how get_irq_server
works.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert pseries SMP code to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:37 +0000 (15:32 +0000)]
powerpc/cpumask: Convert pseries SMP code to new cpumask API

Use new cpumask functions in pseries SMP startup code.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert iseries SMP code to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:36 +0000 (15:32 +0000)]
powerpc/cpumask: Convert iseries SMP code to new cpumask API

Use new cpumask functions in iseries SMP startup code.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert fixup_irqs to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:35 +0000 (15:32 +0000)]
powerpc/cpumask: Convert fixup_irqs to new cpumask API

Use new cpumask_* functions, and dynamically allocate cpumask in fixup_irqs.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert smp_cpus_done to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:34 +0000 (15:32 +0000)]
powerpc/cpumask: Convert smp_cpus_done to new cpumask API

Use the new cpumask_* functions and dynamically allocate the cpumask in
smp_cpus_done.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Convert rtasd to new cpumask API
Anton Blanchard [Mon, 26 Apr 2010 15:32:33 +0000 (15:32 +0000)]
powerpc/cpumask: Convert rtasd to new cpumask API

Use cpumask_first, cpumask_next in rtasd code.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/cpumask: Use cpu_online_mask
Anton Blanchard [Mon, 26 Apr 2010 15:32:32 +0000 (15:32 +0000)]
powerpc/cpumask: Use cpu_online_mask

Change &cpu_online_map to cpu_online_mask.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Invoke oom-killer from page fault
Benjamin Herrenschmidt [Thu, 6 May 2010 07:15:58 +0000 (17:15 +1000)]
powerpc: Invoke oom-killer from page fault

As explained in commit 1c0fe6e3bd, we want to call the architecture independent
oom killer when getting an unexplained OOM from handle_mm_fault, rather than
simply killing current.

Cc: linuxppc-dev@ozlabs.org
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: linux-arch@vger.kernel.org
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/mm: Track backing pages allocated by vmemmap_populate()
Mark Nelson [Wed, 21 Apr 2010 16:21:03 +0000 (16:21 +0000)]
powerpc/mm: Track backing pages allocated by vmemmap_populate()

We need to keep track of the backing pages that get allocated by
vmemmap_populate() so that when we use kdump, the dump-capture kernel knows
where these pages are.

We use a simple linked list of structures that contain the physical address
of the backing page and corresponding virtual address to track the backing
pages.
To save space, we just use a pointer to the next struct vmemmap_backing. We
can also do this because we never remove nodes.  We call the pointer "list"
to be compatible with changes made to the crash utility.

vmemmap_populate() is called either at boot-time or on a memory hotplug
operation. We don't have to worry about the boot-time calls because they
will be inherently single-threaded, and for a memory hotplug operation
vmemmap_populate() is called through:
sparse_add_one_section()
            |
            V
kmalloc_section_memmap()
            |
            V
sparse_mem_map_populate()
            |
            V
vmemmap_populate()
and in sparse_add_one_section() we're protected by pgdat_resize_lock().
So, we don't need a spinlock to protect the vmemmap_list.

We allocate space for the vmemmap_backing structs by allocating whole pages
in vmemmap_list_alloc() and then handing out chunks of this to
vmemmap_list_populate().

This means that we waste at most just under one page, but this keeps the code
is simple.

Signed-off-by: Mark Nelson <markn@au1.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agoi2c/ibm-iic: Drop NO_IRQ
Wolfram Sang [Thu, 1 Apr 2010 14:17:01 +0000 (14:17 +0000)]
i2c/ibm-iic: Drop NO_IRQ

Drop NO_IRQ as 0 is the preferred way to describe 'no irq'
(http://lkml.org/lkml/2005/11/21/221). This change is safe, as the driver is
only used on powerpc, where NO_IRQ is 0 anyhow.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Sean MacLennan <smaclennan@pikatech.com>
Cc: Ben Dooks <ben-linux@fluff.org>
Acked-by: Sean MacLennan <smaclennan@pikatech.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agoi2c/cpm: Drop NO_IRQ
Wolfram Sang [Thu, 1 Apr 2010 14:17:00 +0000 (14:17 +0000)]
i2c/cpm: Drop NO_IRQ

Drop NO_IRQ as 0 is the preferred way to describe 'no irq'
(http://lkml.org/lkml/2005/11/21/221). This change is safe, as the driver is
only used on powerpc, where NO_IRQ is 0 anyhow.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Jochen Friedrich <jochen@scram.de>
Cc: Ben Dooks <ben-linux@fluff.org>
Acked-by: Jochen Friedrich <jochen@scram.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agoi2c/mpc: Drop NO_IRQ
Wolfram Sang [Thu, 1 Apr 2010 14:16:59 +0000 (14:16 +0000)]
i2c/mpc: Drop NO_IRQ

Drop NO_IRQ as 0 is the preferred way to describe 'no irq'
(http://lkml.org/lkml/2005/11/21/221). This change is safe, as the driver is
only used on powerpc, where NO_IRQ is 0 anyhow.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Correct parport interrupt parsing
Martyn Welch [Mon, 26 Apr 2010 22:50:21 +0000 (22:50 +0000)]
powerpc: Correct parport interrupt parsing

Currently the parsing of the device tree in
arch/powerpc/include/asm/parport.h assumes that the interrupt provided in
the parallel port node is a valid virtual irq. The values for the
interrupts provided in the device tree should have meaning in the context
of the driver for the specific interrupt controller to which the interrupt
is connected and irq_of_parse_and_map() should be used to determine the
correct virtual irq.

Signed-off-by: Martyn Welch <martyn.welch@ge.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc: Fix CONFIG_DEBUG_PAGEALLOC on 603/e300
Benjamin Herrenschmidt [Tue, 27 Apr 2010 21:22:55 +0000 (21:22 +0000)]
powerpc: Fix CONFIG_DEBUG_PAGEALLOC on 603/e300

So we tried to speed things up a bit using flush_hash_pages() directly
but that falls over on 603 of course meaning we fail to flush the TLB
properly and we may even end up having it corrupt memory randomly by
accessing a hash table that doesn't exist.

This removes the "optimization" by always going through flush_tlb_page()
for now at least.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/pseries: Only call start-cpu when a CPU is stopped
Michael Neuling [Wed, 28 Apr 2010 13:39:41 +0000 (13:39 +0000)]
powerpc/pseries: Only call start-cpu when a CPU is stopped

Currently we always call start-cpu irrespective of if the CPU is
stopped or not. Unfortunatley on POWER7, firmware seems to not like
start-cpu being called when a cpu already been started.  This was not
the case on POWER6 and earlier.

This patch checks to see if the CPU is stopped or not via an
query-cpu-stopped-state call, and only calls start-cpu on CPUs which
are stopped.

This fixes a bug with kexec on POWER7 on PHYP where only the primary
thread would make it to the second kernel.

Reported-by: Ankita Garg <ankita@linux.vnet.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agopowerpc/pseries: Make query_cpu_stopped callable outside hotplug cpu
Michael Neuling [Wed, 28 Apr 2010 13:39:41 +0000 (13:39 +0000)]
powerpc/pseries: Make query_cpu_stopped callable outside hotplug cpu

This moves query_cpu_stopped() out of the hotplug cpu code and into
smp.c so it can called in other places and renames it to
smp_query_cpu_stopped().

It also cleans up the return values by adding some #defines

Cc: <stable@kernel.org>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
14 years agoMerge branch 'zerolen' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6
Linus Torvalds [Wed, 5 May 2010 22:48:13 +0000 (15:48 -0700)]
Merge branch 'zerolen' of git://git./linux/kernel/git/jgarzik/misc-2.6

* 'zerolen' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/misc-2.6:
  [MTD] Remove zero-length files mtdbdi.c and internal.ho

14 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Wed, 5 May 2010 22:47:57 +0000 (15:47 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  pata_pcmcia / ide-cs: Fix bad hashes for Transcend and kingston IDs
  libata: Fix several inaccuracies in developer's guide

14 years ago[MTD] Remove zero-length files mtdbdi.c and internal.ho
Jeff Garzik [Wed, 5 May 2010 19:25:12 +0000 (15:25 -0400)]
[MTD] Remove zero-length files mtdbdi.c and internal.ho

Both were "removed" in commit a33eb6b91034c95b9b08576f68be170f995b2c7d.

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agopata_pcmcia / ide-cs: Fix bad hashes for Transcend and kingston IDs
Kristoffer Ericson [Sun, 2 May 2010 19:48:24 +0000 (21:48 +0200)]
pata_pcmcia / ide-cs: Fix bad hashes for Transcend and kingston IDs

This patch fixes the bad hashes for one Kingston and one Transcend card.
Thanks to komuro for pointing this out.

Signed-off-by: Kristoffer Ericson <kristoffer.ericson@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: Fix several inaccuracies in developer's guide
Sergei Shtylyov [Wed, 5 May 2010 13:27:10 +0000 (17:27 +0400)]
libata: Fix several inaccuracies in developer's guide

Commit 6bfff31e77cfa1b13490337e5a4dbaa3407e83ac (libata: kill probe_ent
and related helpers) killed ata_device_add() but didn't remove references
to it from the libata developer's guide.

Commits 9363c3825ea9ad76561eb48a395349dd29211ed6 (libata: rename SFF
functions) and 5682ed33aae05d10a25c95633ef9d9c062825888 (libata: rename
SFF port ops) renamed the taskfile access methods but didn't update the
developer's guide.  Commit c9f75b04ed5ed65a058d18a8a8dda50632a96de8
(libata: kill ata_noop_dev_select()) didn't update the developer's
guide as well.

The guide also refers to the long gone ata_pio_data_xfer_noirq(),
ata_pio_data_xfer(), and ata_mmio_data_xfer() -- replace those by
the modern ata_sff_data_xfer_noirq(), ata_sff_data_xfer(), and
ata_sff_data_xfer32().

Also, remove the reference to non-existant ata_port_stop()...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agoMerge branch 'slab-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penber...
Linus Torvalds [Wed, 5 May 2010 18:18:16 +0000 (11:18 -0700)]
Merge branch 'slab-for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'slab-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slub: Fix bad boundary check in init_kmem_cache_nodes()

14 years agoslub: Fix bad boundary check in init_kmem_cache_nodes()
Zhang, Yanmin [Thu, 1 Apr 2010 09:32:30 +0000 (17:32 +0800)]
slub: Fix bad boundary check in init_kmem_cache_nodes()

Function init_kmem_cache_nodes is incorrect when checking upper limitation of
kmalloc_caches. The breakage was introduced by commit
91efd773c74bb26b5409c85ad755d536448e229c ("dma kmalloc handling fixes").

Acked-by: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
14 years agopowerpc/4xx: Add optional "reset_type" property to control reboot via dts
Stefan Roese [Tue, 27 Apr 2010 22:13:34 +0000 (22:13 +0000)]
powerpc/4xx: Add optional "reset_type" property to control reboot via dts

By setting "reset_type" to one of the following values, the default
software reset mechanism may be overidden. Here the possible values of
"reset_type":

  1 - PPC4xx core reset
  2 - PPC4xx chip reset
  3 - PPC4xx system reset (default)

This will be used by a new PPC440SPe board port, which needs a "chip
reset" instead of the default "system reset" to be asserted.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Wed, 5 May 2010 16:06:24 +0000 (09:06 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  KEYS: call_sbin_request_key() must write lock keyrings before modifying them
  KEYS: Use RCU dereference wrappers in keyring key type code
  KEYS: find_keyring_by_name() can gain access to a freed keyring

14 years agopowerpc/47x: defconfig for 476 on the iss 4xx simulator
Dave Kleikamp [Fri, 5 Mar 2010 10:43:35 +0000 (10:43 +0000)]
powerpc/47x: defconfig for 476 on the iss 4xx simulator

A defconfig for the IBM ISS 476 simulator

Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agopowerpc/4xx: Simple platform for the ISS 4xx simulator
Torez Smith [Fri, 5 Mar 2010 10:45:54 +0000 (10:45 +0000)]
powerpc/4xx: Simple platform for the ISS 4xx simulator

This is a trivial 4xx plaform that uses the new simple bsp from
Josh and is handy to use in simulators such as ISS or even Mambo
who don't properly implement most of the actual devices in the
SoC but really only the core.

Signed-off-by: Torez Smith <lnxtorez@linux.vnet.ibm.com>
Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq
Linus Torvalds [Wed, 5 May 2010 14:56:36 +0000 (07:56 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tj/wq

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: flush_delayed_work: keep the original workqueue for re-queueing

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 5 May 2010 14:55:07 +0000 (07:55 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  FEC: Fix kernel panic in fec_set_mac_address.
  ipv6: Fix default multicast hops setting.
  net: ep93xx_eth stops receiving packets
  drivers/net/phy: micrel phy driver
  dm9601: fix phy/eeprom write routine
  ppp_generic: handle non-linear skbs when passing them to pppd
  ppp_generic: pull 2 bytes so that PPP_PROTO(skb) is valid
  net: fix compile error due to double return type in SOCK_DEBUG
  net/usb: initiate sync sequence in sierra_net.c driver
  net/usb: remove default in Kconfig for sierra_net driver
  r8169: Fix rtl8169_rx_interrupt()
  e1000e: Fix oops caused by ASPM patch.
  net/sb1250: register mdio bus in probe
  sctp: Fix skb_over_panic resulting from multiple invalid parameter errors (CVE-2010-1173) (v4)
  p54pci: fix bugs in p54p_check_tx_ring

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Wed, 5 May 2010 14:54:22 +0000 (07:54 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: hda: Fix 0 dB for Packard Bell models using Conexant CX20549 (Venice)
  ALSA: hda - Add quirk for Dell Inspiron 19T using a Conexant CX20582
  ALSA: take tu->qlock with irqs disabled
  ALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite P500-PSPGSC-01800T
  ALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite Pro T130-15F
  ALSA: hda - fix array indexing while creating inputs for Cirrus codecs
  ALSA: es968: fix wrong PnP dma index

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Wed, 5 May 2010 14:53:18 +0000 (07:53 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: joydev - allow binding to button-only devices
  Input: elantech - ignore high bits in the position coordinates
  Input: elantech - allow forcing Elantech protocol
  Input: elantech - fix firmware version check
  Input: ati_remote - add some missing devices from lirc_atiusb
  Input: eeti_ts - cancel pending work when going to suspend
  Input: Add support of Synaptics Clickpad device
  Revert "Input: ALPS - add signature for HP Pavilion dm3 laptops"
  Input: psmouse - ignore parity error for basic protocols

14 years agoraid6: fix recovery performance regression
Dan Williams [Wed, 5 May 2010 03:41:56 +0000 (20:41 -0700)]
raid6: fix recovery performance regression

The raid6 recovery code should immediately drop back to the optimized
synchronous path when a p+q dma resource is not available.  Otherwise we
run the non-optimized/multi-pass async code in sync mode.

Verified with raid6test (NDISKS=255)

Applies to kernels >= 2.6.32.

Cc: <stable@kernel.org>
Acked-by: NeilBrown <neilb@suse.de>
Reported-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agopowerpc/476: Add isync after loading mmu and debug spr's
Dave Kleikamp [Fri, 5 Mar 2010 10:43:24 +0000 (10:43 +0000)]
powerpc/476: Add isync after loading mmu and debug spr's

476 requires an isync after loading MMU and debug related SPR's.  Some of
these are in performance-critical paths and may need to be optimized, but
initially, we're playing it safe.

Signed-off-by: Torez Smith <lnxtorez@linux.vnet.ibm.com>
Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agoKEYS: call_sbin_request_key() must write lock keyrings before modifying them
David Howells [Fri, 30 Apr 2010 13:32:23 +0000 (14:32 +0100)]
KEYS: call_sbin_request_key() must write lock keyrings before modifying them

call_sbin_request_key() creates a keyring and then attempts to insert a link to
the authorisation key into that keyring, but does so without holding a write
lock on the keyring semaphore.

It will normally get away with this because it hasn't told anyone that the
keyring exists yet.  The new keyring, however, has had its serial number
published, which means it can be accessed directly by that handle.

This was found by a previous patch that adds RCU lockdep checks to the code
that reads the keyring payload pointer, which includes a check that the keyring
semaphore is actually locked.

Without this patch, the following command:

keyctl request2 user b a @s

will provoke the following lockdep warning is displayed in dmesg:

===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
security/keys/keyring.c:727 invoked rcu_dereference_check() without protection!

other info that might help us debug this:

rcu_scheduler_active = 1, debug_locks = 0
2 locks held by keyctl/2076:
 #0:  (key_types_sem){.+.+.+}, at: [<ffffffff811a5b29>] key_type_lookup+0x1c/0x71
 #1:  (keyring_serialise_link_sem){+.+.+.}, at: [<ffffffff811a6d1e>] __key_link+0x4d/0x3c5

stack backtrace:
Pid: 2076, comm: keyctl Not tainted 2.6.34-rc6-cachefs #54
Call Trace:
 [<ffffffff81051fdc>] lockdep_rcu_dereference+0xaa/0xb2
 [<ffffffff811a6d1e>] ? __key_link+0x4d/0x3c5
 [<ffffffff811a6e6f>] __key_link+0x19e/0x3c5
 [<ffffffff811a5952>] ? __key_instantiate_and_link+0xb1/0xdc
 [<ffffffff811a59bf>] ? key_instantiate_and_link+0x42/0x5f
 [<ffffffff811aa0dc>] call_sbin_request_key+0xe7/0x33b
 [<ffffffff8139376a>] ? mutex_unlock+0x9/0xb
 [<ffffffff811a5952>] ? __key_instantiate_and_link+0xb1/0xdc
 [<ffffffff811a59bf>] ? key_instantiate_and_link+0x42/0x5f
 [<ffffffff811aa6fa>] ? request_key_auth_new+0x1c2/0x23c
 [<ffffffff810aaf15>] ? cache_alloc_debugcheck_after+0x108/0x173
 [<ffffffff811a9d00>] ? request_key_and_link+0x146/0x300
 [<ffffffff810ac568>] ? kmem_cache_alloc+0xe1/0x118
 [<ffffffff811a9e45>] request_key_and_link+0x28b/0x300
 [<ffffffff811a89ac>] sys_request_key+0xf7/0x14a
 [<ffffffff81052c0b>] ? trace_hardirqs_on_caller+0x10c/0x130
 [<ffffffff81394fb9>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff81001eeb>] system_call_fastpath+0x16/0x1b

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: Use RCU dereference wrappers in keyring key type code
David Howells [Fri, 30 Apr 2010 13:32:18 +0000 (14:32 +0100)]
KEYS: Use RCU dereference wrappers in keyring key type code

The keyring key type code should use RCU dereference wrappers, even when it
holds the keyring's key semaphore.

Reported-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: find_keyring_by_name() can gain access to a freed keyring
Toshiyuki Okajima [Fri, 30 Apr 2010 13:32:13 +0000 (14:32 +0100)]
KEYS: find_keyring_by_name() can gain access to a freed keyring

find_keyring_by_name() can gain access to a keyring that has had its reference
count reduced to zero, and is thus ready to be freed.  This then allows the
dead keyring to be brought back into use whilst it is being destroyed.

The following timeline illustrates the process:

|(cleaner)                           (user)
|
| free_user(user)                    sys_keyctl()
|  |                                  |
|  key_put(user->session_keyring)     keyctl_get_keyring_ID()
|  || //=> keyring->usage = 0        |
|  |schedule_work(&key_cleanup_task)   lookup_user_key()
|  ||                                   |
|  kmem_cache_free(,user)               |
|  .                                    |[KEY_SPEC_USER_KEYRING]
|  .                                    install_user_keyrings()
|  .                                    ||
| key_cleanup() [<= worker_thread()]    ||
|  |                                    ||
|  [spin_lock(&key_serial_lock)]        |[mutex_lock(&key_user_keyr..mutex)]
|  |                                    ||
|  atomic_read() == 0                   ||
|  |{ rb_ease(&key->serial_node,) }     ||
|  |                                    ||
|  [spin_unlock(&key_serial_lock)]      |find_keyring_by_name()
|  |                                    |||
|  keyring_destroy(keyring)             ||[read_lock(&keyring_name_lock)]
|  ||                                   |||
|  |[write_lock(&keyring_name_lock)]    ||atomic_inc(&keyring->usage)
|  |.                                   ||| *** GET freeing keyring ***
|  |.                                   ||[read_unlock(&keyring_name_lock)]
|  ||                                   ||
|  |list_del()                          |[mutex_unlock(&key_user_k..mutex)]
|  ||                                   |
|  |[write_unlock(&keyring_name_lock)]  ** INVALID keyring is returned **
|  |                                    .
|  kmem_cache_free(,keyring)            .
|                                       .
|                                       atomic_dec(&keyring->usage)
v                                         *** DESTROYED ***
TIME

If CONFIG_SLUB_DEBUG=y then we may see the following message generated:

=============================================================================
BUG key_jar: Poison overwritten
-----------------------------------------------------------------------------

INFO: 0xffff880197a7e200-0xffff880197a7e200. First byte 0x6a instead of 0x6b
INFO: Allocated in key_alloc+0x10b/0x35f age=25 cpu=1 pid=5086
INFO: Freed in key_cleanup+0xd0/0xd5 age=12 cpu=1 pid=10
INFO: Slab 0xffffea000592cb90 objects=16 used=2 fp=0xffff880197a7e200 flags=0x200000000000c3
INFO: Object 0xffff880197a7e200 @offset=512 fp=0xffff880197a7e300

Bytes b4 0xffff880197a7e1f0:  5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a 5a ZZZZZZZZZZZZZZZZ
  Object 0xffff880197a7e200:  6a 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b jkkkkkkkkkkkkkkk

Alternatively, we may see a system panic happen, such as:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000001
IP: [<ffffffff810e61a3>] kmem_cache_alloc+0x5b/0xe9
PGD 6b2b4067 PUD 6a80d067 PMD 0
Oops: 0000 [#1] SMP
last sysfs file: /sys/kernel/kexec_crash_loaded
CPU 1
...
Pid: 31245, comm: su Not tainted 2.6.34-rc5-nofixed-nodebug #2 D2089/PRIMERGY
RIP: 0010:[<ffffffff810e61a3>]  [<ffffffff810e61a3>] kmem_cache_alloc+0x5b/0xe9
RSP: 0018:ffff88006af3bd98  EFLAGS: 00010002
RAX: 0000000000000000 RBX: 0000000000000001 RCX: ffff88007d19900b
RDX: 0000000100000000 RSI: 00000000000080d0 RDI: ffffffff81828430
RBP: ffffffff81828430 R08: ffff88000a293750 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000100000 R12: 00000000000080d0
R13: 00000000000080d0 R14: 0000000000000296 R15: ffffffff810f20ce
FS:  00007f97116bc700(0000) GS:ffff88000a280000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000001 CR3: 000000006a91c000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process su (pid: 31245, threadinfo ffff88006af3a000, task ffff8800374414c0)
Stack:
 0000000512e0958e 0000000000008000 ffff880037f8d180 0000000000000001
 0000000000000000 0000000000008001 ffff88007d199000 ffffffff810f20ce
 0000000000008000 ffff88006af3be48 0000000000000024 ffffffff810face3
Call Trace:
 [<ffffffff810f20ce>] ? get_empty_filp+0x70/0x12f
 [<ffffffff810face3>] ? do_filp_open+0x145/0x590
 [<ffffffff810ce208>] ? tlb_finish_mmu+0x2a/0x33
 [<ffffffff810ce43c>] ? unmap_region+0xd3/0xe2
 [<ffffffff810e4393>] ? virt_to_head_page+0x9/0x2d
 [<ffffffff81103916>] ? alloc_fd+0x69/0x10e
 [<ffffffff810ef4ed>] ? do_sys_open+0x56/0xfc
 [<ffffffff81008a02>] ? system_call_fastpath+0x16/0x1b
Code: 0f 1f 44 00 00 49 89 c6 fa 66 0f 1f 44 00 00 65 4c 8b 04 25 60 e8 00 00 48 8b 45 00 49 01 c0 49 8b 18 48 85 db 74 0d 48 63 45 18 <48> 8b 04 03 49 89 00 eb 14 4c 89 f9 83 ca ff 44 89 e6 48 89 ef
RIP  [<ffffffff810e61a3>] kmem_cache_alloc+0x5b/0xe9

This problem is that find_keyring_by_name does not confirm that the keyring is
valid before accepting it.

Skipping keyrings that have been reduced to a zero count seems the way to go.
To this end, use atomic_inc_not_zero() to increment the usage count and skip
the candidate keyring if that returns false.

The following script _may_ cause the bug to happen, but there's no guarantee
as the window of opportunity is small:

#!/bin/sh
LOOP=100000
USER=dummy_user
/bin/su -c "exit;" $USER || { /usr/sbin/adduser -m $USER; add=1; }
for ((i=0; i<LOOP; i++))
do
/bin/su -c "echo '$i' > /dev/null" $USER
done
(( add == 1 )) && /usr/sbin/userdel -r $USER
exit

Note that the nominated user must not be in use.

An alternative way of testing this may be:

for ((i=0; i<100000; i++))
do
keyctl session foo /bin/true || break
done >&/dev/null

as that uses a keyring named "foo" rather than relying on the user and
user-session named keyrings.

Reported-by: Toshiyuki Okajima <toshi.okajima@jp.fujitsu.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: Toshiyuki Okajima <toshi.okajima@jp.fujitsu.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agopowerpc/476: add machine check handler for 47x core
Dave Kleikamp [Fri, 5 Mar 2010 03:43:18 +0000 (03:43 +0000)]
powerpc/476: add machine check handler for 47x core

The 47x core's MCSR varies from 44x, so it needs it's own machine check
handler.

Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agopowerpc/47x: Base ppc476 support
Dave Kleikamp [Fri, 5 Mar 2010 10:43:12 +0000 (10:43 +0000)]
powerpc/47x: Base ppc476 support

This patch adds the base support for the 476 processor.  The code was
primarily written by Ben Herrenschmidt and Torez Smith, but I've been
maintaining it for a while.

The goal is to have a single binary that will run on 44x and 47x, but
we still have some details to work out.  The biggest is that the L1 cache
line size differs on the two platforms, but it's currently a compile-time
option.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Torez Smith <lnxtorez@linux.vnet.ibm.com>
Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agopowerpc/44x: break out cpu init code into stand-alone function
Dave Kleikamp [Fri, 5 Mar 2010 10:43:07 +0000 (10:43 +0000)]
powerpc/44x: break out cpu init code into stand-alone function

The 47x platform supports multiple cores and shares code with 44x.
Break out code that is common for initializing the primary and secondary
cpus into a function which can be called for both.

Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agopowerpc/booke: Add Stack Marking support to Booke Exception Prolog
Torez Smith [Fri, 5 Mar 2010 10:43:01 +0000 (10:43 +0000)]
powerpc/booke: Add Stack Marking support to Booke Exception Prolog

This patch adds a marker to the exception stack frame to aid in debugging.
It's already inserted on other platforms and xmon recognizes it and
identifies exception frames when showing stack traces.

Signed-off-by: Torez Smith <lnxtorez@linux.vnet.ibm.com>
Signed-off-by: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
14 years agoMerge branch 'fix/hda' into for-linus
Takashi Iwai [Wed, 5 May 2010 08:08:30 +0000 (10:08 +0200)]
Merge branch 'fix/hda' into for-linus

14 years agoALSA: hda: Fix 0 dB for Packard Bell models using Conexant CX20549 (Venice)
Daniel T Chen [Wed, 28 Apr 2010 22:00:11 +0000 (18:00 -0400)]
ALSA: hda: Fix 0 dB for Packard Bell models using Conexant CX20549 (Venice)

BugLink: https://launchpad.net/bugs/541802
The OR's hardware distorts at PCM 100% because it does not correspond to
0 dB. Fix this in patch_cxt5045() for all Packard Bell models.

Reported-by: Valombre
Cc: <stable@kernel.org>
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoALSA: hda - Add quirk for Dell Inspiron 19T using a Conexant CX20582
Anisse Astier [Wed, 28 Apr 2010 16:05:06 +0000 (18:05 +0200)]
ALSA: hda - Add quirk for Dell Inspiron 19T using a Conexant CX20582

Add a quirk for all-in-one computer Dell Inspiron One 19 Touch to have proper
HP and Mic support.

Signed-off-by: Anisse Astier <anisse@astier.eu>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoALSA: take tu->qlock with irqs disabled
Dan Carpenter [Wed, 28 Apr 2010 08:29:14 +0000 (10:29 +0200)]
ALSA: take tu->qlock with irqs disabled

We should disable irqs when we take the tu->qlock because it is used in
the irq handler.  The only place that doesn't is
snd_timer_user_ccallback().  Most of the time snd_timer_user_ccallback()
is called with interrupts disabled but the the first ti->ccallback()
call in snd_timer_notify1() has interrupts enabled.

This was caught by lockdep which generates the following message:

> =================================
> [ INFO: inconsistent lock state ]
> 2.6.34-rc5 #5
> ---------------------------------
> inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
> dolphin/4003 [HC1[1]:SC0[0]:HE0:SE1] takes:
> (&(&tu->qlock)->rlock){?.+...}, at: [<f84ec472>] snd_timer_user_tinterrupt+0x28/0x132 [snd_timer]
> {HARDIRQ-ON-W} state was registered at:
>   [<c1048de9>] __lock_acquire+0x654/0x1482
>   [<c1049c73>] lock_acquire+0x5c/0x73
>   [<c125ac3e>] _raw_spin_lock+0x25/0x34
>   [<f84ec370>] snd_timer_user_ccallback+0x55/0x95 [snd_timer]
>   [<f84ecc4b>] snd_timer_notify1+0x53/0xca [snd_timer]

Reported-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoFEC: Fix kernel panic in fec_set_mac_address.
Mattias Walström [Wed, 5 May 2010 07:55:48 +0000 (00:55 -0700)]
FEC: Fix kernel panic in fec_set_mac_address.

Fix memory corruption that sometimes result in kernel panic.

Signed-off-by: Mattias Walström <mattias@vmlinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite P500-PSPGSC-01800T
Daniel T Chen [Wed, 5 May 2010 02:07:58 +0000 (22:07 -0400)]
ALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite P500-PSPGSC-01800T

BugLink: https://launchpad.net/bugs/549267
The OR verified that using the olpc-xo-1_5 model quirk allows the
headphones to be audible when inserted into the jack. Capture was
also verified to work correctly.

Reported-by: Richard Gagne
Tested-by: Richard Gagne
Cc: <stable@kernel.org>
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite Pro T130-15F
Daniel T Chen [Tue, 4 May 2010 00:39:31 +0000 (20:39 -0400)]
ALSA: hda: Use olpc-xo-1_5 quirk for Toshiba Satellite Pro T130-15F

BugLink: https://launchpad.net/bugs/573284
The OR verified that using the olpc-xo-1_5 model quirk allows the
headphones to be audible when inserted into the jack. Capture was
also verified to work correctly.

Reported-by: Andy Couldrake <acouldrake@googlemail.com>
Tested-by: Andy Couldrake <acouldrake@googlemail.com>
Cc: <stable@kernel.org>
Signed-off-by: Daniel T Chen <crimsun@ubuntu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoALSA: hda - fix array indexing while creating inputs for Cirrus codecs
Brian J. Tarricone [Mon, 3 May 2010 00:32:10 +0000 (17:32 -0700)]
ALSA: hda - fix array indexing while creating inputs for Cirrus codecs

This fixes a problem where cards show up as only having a single mixer
element, suppressing all sound output.

Signed-off-by: Brian J. Tarricone <brian@tarricone.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agoMerge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 5 May 2010 02:08:12 +0000 (19:08 -0700)]
Merge branch 'drm-linus' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms/legacy: only enable load detection property on DVI-I
  drm/radeon/kms: fix panel scaling adjusted mode setup
  drivers/gpu/drm/drm_sysfs.c: sysfs files error handling
  drivers/gpu/drm/radeon/radeon_atombios.c: range check issues
  gpu: vga_switcheroo, fix lock imbalance
  drivers/gpu/drm/drm_memory.c: fix check for end of loop
  drivers/gpu/drm/via/via_video.c: fix off by one issue
  drm/radeon/kms/agp The wrong AGP chipset can cause a NULL pointer dereference
  drm/radeon/kms: r300 fix CS checker to allow zbuffer-only fastfill

14 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 5 May 2010 02:07:35 +0000 (19:07 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/x86/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-tip:
  powernow-k8: Fix frequency reporting
  x86: Fix parse_reservetop() build failure on certain configs
  x86: Fix NULL pointer access in irq_force_complete_move() for Xen guests
  x86: Fix 'reservetop=' functionality

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Wed, 5 May 2010 01:49:34 +0000 (18:49 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  KEYS: Fix RCU handling in key_gc_keyring()
  KEYS: Fix an RCU warning in the reading of user keys

14 years agoKEYS: Fix RCU handling in key_gc_keyring()
David Howells [Tue, 4 May 2010 13:16:10 +0000 (14:16 +0100)]
KEYS: Fix RCU handling in key_gc_keyring()

key_gc_keyring() needs to either hold the RCU read lock or hold the keyring
semaphore if it's going to scan the keyring's list.  Given that it only needs
to read the key list, and it's doing so under a spinlock, the RCU read lock is
the thing to use.

Furthermore, the RCU check added in e7b0a61b7929632d36cf052d9e2820ef0a9c1bfe is
incorrect as holding the spinlock on key_serial_lock is not grounds for
assuming a keyring's pointer list can be read safely.  Instead, a simple
rcu_dereference() inside of the previously mentioned RCU read lock is what we
want.

Reported-by: Serge E. Hallyn <serue@us.ibm.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agoKEYS: Fix an RCU warning in the reading of user keys
David Howells [Fri, 30 Apr 2010 13:32:08 +0000 (14:32 +0100)]
KEYS: Fix an RCU warning in the reading of user keys

Fix an RCU warning in the reading of user keys:

===================================================
[ INFO: suspicious rcu_dereference_check() usage. ]
---------------------------------------------------
security/keys/user_defined.c:202 invoked rcu_dereference_check() without protection!

other info that might help us debug this:

rcu_scheduler_active = 1, debug_locks = 0
1 lock held by keyctl/3637:
 #0:  (&key->sem){+++++.}, at: [<ffffffff811a80ae>] keyctl_read_key+0x9c/0xcf

stack backtrace:
Pid: 3637, comm: keyctl Not tainted 2.6.34-rc5-cachefs #18
Call Trace:
 [<ffffffff81051f6c>] lockdep_rcu_dereference+0xaa/0xb2
 [<ffffffff811aa55f>] user_read+0x47/0x91
 [<ffffffff811a80be>] keyctl_read_key+0xac/0xcf
 [<ffffffff811a8a06>] sys_keyctl+0x75/0xb7
 [<ffffffff81001eeb>] system_call_fastpath+0x16/0x1b

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
14 years agodrm/radeon/kms/legacy: only enable load detection property on DVI-I
Alex Deucher [Fri, 30 Apr 2010 16:37:31 +0000 (12:37 -0400)]
drm/radeon/kms/legacy: only enable load detection property on DVI-I

DVI-D doesn't have analog.  This matches the avivo behavior.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix panel scaling adjusted mode setup
Alex Deucher [Fri, 30 Apr 2010 16:00:44 +0000 (12:00 -0400)]
drm/radeon/kms: fix panel scaling adjusted mode setup

This should duplicate exactly what the ddx does for both
legacy and avivo.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2
Linus Torvalds [Tue, 4 May 2010 23:33:18 +0000 (16:33 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/jlbec/ocfs2

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
  ocfs2: Avoid a gcc warning in ocfs2_wipe_inode().
  ocfs2: Avoid direct write if we fall back to buffered I/O
  ocfs2_dlmfs: Fix math error when reading LVB.
  ocfs2: Update VFS inode's id info after reflink.
  ocfs2: potential ERR_PTR dereference on error paths
  ocfs2: Add directory entry later in ocfs2_symlink() and ocfs2_mknod()
  ocfs2: use OCFS2_INODE_SKIP_ORPHAN_DIR in ocfs2_mknod error path
  ocfs2: use OCFS2_INODE_SKIP_ORPHAN_DIR in ocfs2_symlink error path
  ocfs2: add OCFS2_INODE_SKIP_ORPHAN_DIR flag and honor it in the inode wipe code
  ocfs2: Reset status if we want to restart file extension.
  ocfs2: Compute metaecc for superblocks during online resize.
  ocfs2: Check the owner of a lockres inside the spinlock
  ocfs2: one more warning fix in ocfs2_file_aio_write(), v2
  ocfs2_dlmfs: User DLM_* when decoding file open flags.

14 years agoFix the x86_64 implementation of call_rwsem_wait()
David Howells [Tue, 4 May 2010 12:42:53 +0000 (13:42 +0100)]
Fix the x86_64 implementation of call_rwsem_wait()

The x86_64 call_rwsem_wait() treats the active state counter part of the
R/W semaphore state as being 16-bit when it's actually 32-bit (it's half
of the 64-bit state).  It should do "decl %edx" not "decw %dx".

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>