Holger Schurig [Fri, 23 May 2008 10:16:51 +0000 (12:16 +0200)]
libertas: use lbs_pr_XX instead of printk
... because lbs_pr_XXX prefixes all messages with "libertas: "
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Holger Schurig [Fri, 23 May 2008 08:18:26 +0000 (10:18 +0200)]
libertas: speeds up downloading of CF firmware
Keep the timeout the same (1000*500 == 100000 * 5), but take shorter
naps. Makes downloading the firmware slightly faster.
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Holger Schurig [Fri, 23 May 2008 08:07:56 +0000 (10:07 +0200)]
libertas: don't spin_unlock_irq() twice
priv->driver_lock has already been unlocked some lines above. This patch
fixes the sparse warning:
drivers/net/wireless/libertas/main.c:792:6: warning: context problem in 'lbs_thread': '_spin_unlock_irq' expected different context
drivers/net/wireless/libertas/main.c:792:6: context 'lock': wanted >= 1, got 0
Signed-of-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Thu, 29 May 2008 08:38:53 +0000 (10:38 +0200)]
mac80211: clean up skb reallocation code
This cleans up the skb reallocation code to avoid problems with
skb->truesize, not resize an skb twice for a single output path
because we didn't expand it enough during the first copy and also
removes the code to further expand it during crypto operations
which will no longer be necessary.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Daniel Walker [Thu, 22 May 2008 07:00:03 +0000 (00:00 -0700)]
ps3: gelic: updown_lock semaphore to mutex
Signed-off-by: Daniel Walker <dwalker@mvista.com> Acked-by: Masakazu Mokuno <mokuno@sm.sony.co.jp> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Daniel Walker [Thu, 22 May 2008 07:00:02 +0000 (00:00 -0700)]
ps3: gelic: assoc_stat_lock semaphore to mutex
Signed-off-by: Daniel Walker <dwalker@mvista.com> Acked-by: Masakazu Mokuno <mokuno@sm.sony.co.jp> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Daniel Walker [Thu, 22 May 2008 07:00:01 +0000 (00:00 -0700)]
ps3: gelic: scan_lock semaphore to mutex
Signed-off-by: Daniel Walker <dwalker@mvista.com> Acked-by: Masakazu Mokuno <mokuno@sm.sony.co.jp> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Tue, 20 May 2008 10:10:49 +0000 (12:10 +0200)]
b43: enable mesh
This patch enables b43 to do mesh networking, tested against my zd1211rw
dongle.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Reviewed-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Alan Cox [Tue, 3 Jun 2008 14:18:54 +0000 (15:18 +0100)]
serial_core: uart_set_ldisc infrastructure
The tty layer provides a callback that is used when the line discipline
is changed. Some hardware uses this to configure hardware specific
features such as IrDA mode on serial ports. Unfortunately the serial
layer does not provide this feature or pass it down to drivers.
Blackfin used to hack around this by rewriting the tty ops, but those are
now properly shared and const so the hack fails. Instead provide the
proper operations.
This change plus a follow up from the Blackfin guys is needed to avoid
blackfin losing features in this release.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Takashi Iwai [Tue, 3 Jun 2008 12:46:34 +0000 (14:46 +0200)]
[ALSA] hda - Fix resume of auto-config mode with Realtek codecs
The auto-config mode of Realtek ALC codecs has a bug since 2.6.25
that it cannot resume properly. The problem was the wrong assignment
of init_hook that overrides the whole initialization.
Yinghai Lu [Mon, 2 Jun 2008 06:53:50 +0000 (23:53 -0700)]
x86: clean up max_pfn_mapped usage - 32-bit
on 32-bit in head_32.S after initial page table is done, we get initial
max_pfn_mapped, and then kernel_physical_mapping_init will give us
a final one.
We need to use that to make sure find_e820_area will get valid addresses
for boot_map and for NODE_DATA(0) on numa32.
XEN PV and lguest may need to assign max_pfn_mapped too.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Yinghai Lu [Sun, 1 Jun 2008 20:17:38 +0000 (13:17 -0700)]
x86: update mptable
make mptable to be consistent with acpi routing, so we could:
1. kexec kernel with acpi=off
2. work around BIOSes where acpi routing is working, but mptable is
not right, so can use kernel/kexec to start other OSes that don't have
good acpi support.
command line: update_mptable
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Yinghai Lu [Sun, 1 Jun 2008 05:52:47 +0000 (22:52 -0700)]
x86, 32-bit: change propagate_e820_map() back to find_max_pfn()
we don't need to call memory_present that early.
numa and sparse will call memory_present later and might
even fail, it will call memory_present for the full range.
also for sparse it will call alloc_bootmem ... before we set up bootmem.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Yinghai Lu [Sun, 1 Jun 2008 05:53:47 +0000 (22:53 -0700)]
x86: set node_remap_size[0] in fallback path
... otherwise alloc_remap will not get node_mem_map from kva area, and
alloc_node_mem_map has to alloc_bootmem_node to get mem_map.
It will use two low address copies ...
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Timur Tabi [Mon, 2 Jun 2008 20:09:35 +0000 (15:09 -0500)]
[POWERPC] fsl: Update fsl_soc to use cell-index property of I2C nodes
Currently, fsl_i2c_of_init() uses the order of the I2C adapter nodes in the
device tree to enumerate the I2C adapters. Instead, let's check for the
cell-index property and use it if it exists.
This is handy for device drivers that need to identify the I2C adapters by
specific numbers. The Freescale MPC8610 ASoC V2 sound drivers are an example.
Signed-off-by: Timur Tabi <timur@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Anton Vorontsov [Sun, 1 Jun 2008 09:49:32 +0000 (11:49 +0200)]
mmc_spi: mmc_spi.h should include linux/interrupts.h
Since mmc_spi.h uses irqreturn_t type, it should include appropriate
header, otherwise build will break if users didn't include it (some of
them do not use interrupts).
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Mon, 2 Jun 2008 22:25:27 +0000 (15:25 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
8250 Serial Driver: revert extra IRQ flag definition patch
Blackfin arch: update anomaly headers from toolchain trunk
Blackfin arch: Remove bad and usless code
Blackfin arch: Fix bug - set corret SSEL and IRQ to enable AD7877 on BF527
Blackfin arch: Fix typo. it should be _outsw_8
Blackfin arch: Cleanup no functional changes
On machines with more than one exception level any system register that
might be modified by the "normal" exception level needs to be saved and
restored on taking a higher level exception. We already are saving
and restoring ESR and DEAR.
For critical level add SRR0/1.
For debug level add CSRR0/1 and SRR0/1.
For machine check level add DSRR0/1, CSRR0/1, and SRR0/1.
On FSL Book-E parts we always save/restore the MAS registers for critical,
debug, and machine check level exceptions. On 44x we always save/restore
the MMUCR.
Additionally, we save and restore the ksp_limit since we have to adjust it
for each exception level.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org> Acked-by: Paul Mackerras <paulus@samba.org>
Kumar Gala [Wed, 30 Apr 2008 09:17:22 +0000 (04:17 -0500)]
[POWERPC] Rework EXC_LEVEL_EXCEPTION_PROLOG code
* Cleanup the code a bit my allocating an INT_FRAME on our exception
stack there by make references go from GPR11-INT_FRAME_SIZE(r8) to
just GPR11(r8)
* simplify {lvl}_transfer_to_handler code by moving the copying of the
temp registers we use if we come from user space into the PROLOG
* If the exception came from kernel mode copy thread_info flags,
preempt, and task pointer from the process thread_info.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org> Acked-by: Paul Mackerras <paulus@samba.org>
Kumar Gala [Wed, 30 Apr 2008 08:49:55 +0000 (03:49 -0500)]
[POWERPC] Move to runtime allocated exception stacks
For the additonal exception levels (critical, debug, machine check) on
40x/book-e we were using "static" allocations of the stack in the
associated head.S.
Move to a runtime allocation to make the code a bit easier to read as
we mimic how we handle IRQ stacks. Its also a bit easier to setup the
stack with a "dummy" thread_info in C code.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org> Acked-by: Paul Mackerras <paulus@samba.org>
Kumar Gala [Fri, 30 May 2008 18:43:43 +0000 (13:43 -0500)]
[POWERPC] 85xx: Add next-level-cache property
Added next-level-cache to the L1 and a reference to the new L2 label.
This is per the ePAPR 0.94 spec. Since we are't really dependent on this
today we aren't supporting the "legacy" l2-cache phandle that is specified
in the PPC v2.1 OF Binding spec.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Kumar Gala [Fri, 30 May 2008 17:12:26 +0000 (12:12 -0500)]
[POWERPC] Cleanup mpic nodes in .dts
Removed clock-frequency, big-endian, and built-in props as they aren't
specified anywhere. Also added compatible = "chrp,open-pic" in the
places it was missing.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Timur Tabi [Thu, 15 May 2008 22:46:10 +0000 (17:46 -0500)]
[POWERPC] fsl: Add CS4270 i2c data to fsl_soc.c
The i2c_devices[] array in fsl_soc.c lists all the I2C nodes that are supported
on Freescale boards. Add an entry for the Cirrus Logic CS4270 so that a
new-style CS4270 driver will work.
Signed-off-by: Timur Tabi <timur@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Jason Jin [Fri, 23 May 2008 08:32:46 +0000 (16:32 +0800)]
[POWERPC] fsl: PCIe MSI support for 83xx/85xx/86xx processors.
This MSI driver can be used on 83xx/85xx/86xx board.
In this driver, virtual interrupt host and chip were
setup. There are 256 MSI interrupts in this host, Every 32
MSI interrupts cascaded to one IPIC/MPIC interrupt.
The chip was treated as edge sensitive and some necessary
functions were setup for this chip.
Before using the MSI interrupt, PCI/PCIE device need to
ask for a MSI interrupt in the 256 MSI interrupts. A 256bit
bitmap show which MSI interrupt was used, reserve bit in
the bitmap can be used to force the device use some designate
MSI interrupt in the 256 MSI interrupts. Sometimes this is useful
for testing the all the MSI interrupts. The msi-available-ranges
property in the dts file was used for this purpose.
Signed-off-by: Jason Jin <Jason.jin@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Russell King [Mon, 2 Jun 2008 19:38:15 +0000 (20:38 +0100)]
[ARM] pxa: fix tosa.c build error
Work around:
arch/arm/mach-pxa/tosa.c: In function `tosa_poweroff':
arch/arm/mach-pxa/tosa.c:470: error: `GPIO_OUT' undeclared (first use in this function)
arch/arm/mach-pxa/tosa.c:470: error: (Each undeclared identifier is reported only once
arch/arm/mach-pxa/tosa.c:470: error: for each function it appears in.)
The proper fix exists in the PXA branch of my kernel git tree, which
will be pushed during the next merge window.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Dmitry Baryshkov [Sat, 31 May 2008 15:14:48 +0000 (16:14 +0100)]
[ARM] 5071/2: Drop PXA_SHARPSL_25x/27x case from PXA Kconfig.
As nothing in the code references to the PXA_SHARPSL_25x/27x,
we can drop that Kconfig case and permit all-zaurus builds.
Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com> Acked-by: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Dmitry Baryshkov [Sat, 31 May 2008 15:17:32 +0000 (16:17 +0100)]
[ARM] 5073/1: spitz_pm: don't register devices on non-spitz machines
Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com> Acked-by: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Dmitry Baryshkov [Sat, 31 May 2008 15:16:54 +0000 (16:16 +0100)]
[ARM] 5072/1: corgi_pm: don't register devices on non-corgi machines
Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com> Acked-by: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Russell King [Sun, 18 May 2008 13:59:36 +0000 (14:59 +0100)]
[ARM] pxa: separate out power manager and clock registers
The power manager and core clock registers aren't present in PXA3
CPUs. Move them out of pxa-regs.h into pxa2xx-regs.h, and include
pxa2xx-regs.h where necessary.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Nicolas Ferre [Tue, 8 Apr 2008 12:59:18 +0000 (13:59 +0100)]
[ARM] 4940/1: AT91: UDPHS driver: SAM9RL board and cpu integration.
Adds support for the USB High Speed Device Port on the AT91SAM9RL
system on chip. The AT91SAM9RL uses the same UDPHS IP as the AVR32 and
the AT91CAP9 (atmel_usba_udc driver).
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com> Acked-by: Andrew Victor <linux@maxim.org.za> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Stelian Pop [Sat, 5 Apr 2008 20:16:15 +0000 (21:16 +0100)]
[ARM] 4935/1: AT91CAP9: enable RTC-on-RTT in defconfig.
Update the help text for RTC_DRV_AT91SAM9 to mention that the
option apply to AT91CAP9 processors too, and enable it in the
defconfig.
Signed-off-by: Stelian Pop <stelian@popies.net> Signed-off-by: Andrew Victor <linux@maxim.org.za> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Stelian Pop [Sat, 5 Apr 2008 20:15:25 +0000 (21:15 +0100)]
[ARM] 4934/1: AT91CAP9 UDPHS driver: board and cpu integration.
This is patch 2 of 2 adding support for the USB High Speed Device Port
on the AT91CAP9 system on chip. The AT91CAP9 uses the same UDPHS IP
as the AVR32 and the AT91SAM9RL.
This patch declares the UDPHS ressources in the at91cap9 (cpu and
adk board) files, wires up the atmel_usba_udc driver to them,
and activates the driver in the defconfig.
Signed-off-by: Stelian Pop <stelian@popies.net> Acked-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Acked-by: Andrew Victor <linux@maxim.org.za> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This is patch 1 of 2 adding support for the USB High Speed Device Port
on the AT91CAP9 system on chip. The AT91CAP9 uses the same UDPHS IP
as the AVR32 and the AT91SAM9RL.
This patch makes the generic AT91 adaptations, mainly dealing with
the addition of the UDPHS UTMI clock.
Signed-off-by: Stelian Pop <stelian@popies.net> Acked-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Acked-by: Andrew Victor <linux@maxim.org.za> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Andrew Victor [Sat, 24 May 2008 16:06:45 +0000 (17:06 +0100)]
[ARM] 5056/1: [AT91] Cleanup YL9200 board file
Cleanup the YL9200 board-support file.
Other things fixed are:
- Use new-style UART initialization
- Register all LEDs as gpio_leds.
- NOR Flash error noted in comments fixed by increasing YL9200_FLASH_SIZE
- The only I2C device is the AT24C eeprom.
- Setup of NWAIT pin and programming of SMC controller for the LCD/VGA.
- Configure touchscreen interrupt pin.
Also adding the board to the KConfig and Makefile.
Signed-off-by: Andrew Victor <linux@maxim.org.za> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Russell King [Thu, 24 Apr 2008 14:28:11 +0000 (15:28 +0100)]
[ARM] pxa: avoid kfreeing static data if platform device fails to register
When a dynamically allocated platform device is 'put', the platform
device's platform_data is kfree'd. This is bad if it's pointing at
static data. Use the provided function to register platform data
for these devices.
This also means we can mark the pcmcia ops structures as __initdata.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
cm_x270 and mainstone both register their PCMCIA devices using the same
name, resulting in a warning message from the kernel. Avoid this by
making the cm_x270 and mainstone PCMCIA initialisation conditional on
the machine type we're running on.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Russell King [Thu, 24 Apr 2008 14:13:36 +0000 (15:13 +0100)]
[ARM] pxa: don't register lpd270 cpld_irq sysdev if !lpd270
Don't register the LPD270 cpld_irq system device when we're not running
on a LPD270 machine - "cpld_irq" is also registered (separately) by
Lubbock and Mainstone.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Dmitry Baryshkov [Sat, 31 May 2008 15:18:11 +0000 (16:18 +0100)]
[ARM] 5074/1: fix warning: missing terminating ' character
Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com> Acked-by: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
surinder [Wed, 28 May 2008 08:51:16 +0000 (09:51 +0100)]
[ARM] 5067/1: _raw_write_can_lock macro bugfix
The current __raw_write_can_lock macro tests whether the lock can be
locked by checking if it is equal to 0x80000000, whereas the lock
should be lockable if its value is 0 i.e. unlocked state is
represented by 0. Hence the macro should test the value of lock
against 0 and not 0x80000000.
Signed-off-by: Surinder Pal Singh <srplsnh@gmail.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Philipp Zabel [Fri, 30 May 2008 17:53:55 +0000 (18:53 +0100)]
[ARM] 5070/1: pxa: add GPIO104_PSKTSEL to pxa27x MFP configuration
PSKTSEL can be routed to GPIO pin 104. This configuration is used by
HP iPAQ hx4700.
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com> Acked-by: Jrgen Schindele <linux@schindele.name> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add the IORESOURCE_IRQ_HIGHEDGE to the DM9000 IRQ resource
to stop the driver itself complaining it was not given
any flags to use.
Signed-off-by: Mike Rapoport <mike@compulab.co.il> Acked-by: Eric Miao <eric.miao@marvell.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Add the IORESOURCE_IRQ_HIGHEDGE to the DM9000 IRQ resource
to stop the driver itself complaining it was not given
any flags to use.
Signed-off-by: Mike Rapoport <mike@compulab.co.il> Acked-by: Eric Miao <eric.miao@marvell.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Abhishek Sagar [Sat, 31 May 2008 08:54:02 +0000 (14:24 +0530)]
ftrace: export kretprobe_trampoline for function tracer
Follow suit from kprobe implementations on other archs and make kretprobe_trampoline non-static. Ftrace implmentation (more specifically, kernel/trace/trace.c) requires access to it (see-> http://kerneltrap.org/mailarchive/linux-kernel/2008/5/27/1955234).
On resume, the vcpu timer modes will not be restored. The timer
infrastructure doesn't do this for us, since it assumes the cpus
are offline. We can just poke the other vcpus into the right mode
directly though.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
If we're using vcpu_info mapping, then make sure its restored on all
processors before relasing them from stop_machine.
The only complication is that if this fails, we can't continue because
we've already made assumptions that the mapping is available (baked in
calls to the _direct versions of the functions, for example).
Fortunately this can only happen with a 32-bit hypervisor, which may
possibly run out of mapping space. On a 64-bit hypervisor, this is a
non-issue.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
xen: avoid hypercalls when updating unpinned pud/pmd
When operating on an unpinned pagetable (ie, one under construction or
destruction), it isn't necessary to use a hypercall to update a
pud/pmd entry. Jan Beulich observed that a similar optimisation
avoided many thousands of hypercalls while doing a kernel build.
One tricky part is that early in the kernel boot there's no page
structure, so we can't check to see if the page is pinned. In that
case, we just always use the hypercall.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: Jan Beulich <jbeulich@novell.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Pavel Machek [Thu, 29 May 2008 07:30:21 +0000 (00:30 -0700)]
suspend-vs-iommu: prevent suspend if we could not resume
iommu/gart support misses suspend/resume code, which can do bad stuff,
including memory corruption on resume. Prevent system suspend in case we
would be unable to resume.
Signed-off-by: Pavel Machek <pavel@suse.cz> Tested-by: Patrick <ragamuffin@datacomm.ch> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Jack Steiner [Wed, 28 May 2008 14:51:18 +0000 (09:51 -0500)]
x86, uv: update macros used by UV platform
Update the UV address macros to better describe the
fields of UV physical addresses. Improve comments
in the header files. Add additional MMR definitions.
Signed-off-by: Jack Steiner <steiner@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Steven Rostedt [Wed, 28 May 2008 00:48:37 +0000 (20:48 -0400)]
ftrace: user update and disable dynamic ftrace daemon
In dynamic ftrace, the mcount function starts off pointing to a stub
function that just returns.
On start up, the call to the stub is modified to point to a "record_ip"
function. The job of the record_ip function is to add the function to
a pre-allocated hash list. If the function is already there, it simply is
ignored, otherwise it is added to the list.
Later, a ftraced daemon wakes up and calls kstop_machine if any functions
have been recorded, and changes the calls to the recorded functions to
a simple nop. If no functions were recorded, the daemon goes back to sleep.
The daemon wakes up once a second to see if it needs to update any newly
recorded functions into nops. Usually it does not, but if a lot of code
has been executed for the first time in the kernel, the ftraced daemon
will call kstop_machine to update those into nops.
The problem currently is that there's no way to stop the daemon from doing
this, and it can cause unneeded latencies (800us which for some is bothersome).
This patch adds a new file /debugfs/tracing/ftraced_enabled. If the daemon
is active, reading this will return "enabled\n" and "disabled\n" when the
daemon is not running. To disable the daemon, the user can echo "0" or
"disable" into this file, and "1" or "enable" to re-enable the daemon.
Since the daemon is used to convert the functions into nops to increase
the performance of the system, I also added that anytime something is
written into the ftraced_enabled file, kstop_machine will run if there
are new functions that have been detected that need to be converted.
This way the user can disable the daemon but still be able to control the
conversion of the mcount calls to nops by simply,
"echo 0 > /debugfs/tracing/ftraced_enabled"
when they need to do more conversions.
To see the number of converted functions:
"cat /debugfs/tracing/dyn_ftrace_total_info"
Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Vegard Nossum [Wed, 28 May 2008 07:46:19 +0000 (09:46 +0200)]
x86: break mutual header inclusion
This breaks up the mutual inclusion between headers ptrace.h and vm86.h
by moving some small part of vm86.h which is needed by ptrace.h into
processor-flags.h.
We also try to move #include lines to the top.
This has been compile tested on x86_32 and x86_64 defconfig, and run
through 'make headers_check'.
Abhishek Sagar [Tue, 27 May 2008 18:33:18 +0000 (00:03 +0530)]
ftrace: distinguish kretprobe'd functions in trace logs
Tracing functions via ftrace which have a kretprobe installed on them, can produce misleading output in their trace logs. E.g, consider the correct trace of the following sequence:
But if irq_enter() has a kretprobe installed on it, the return value stored on the stack at each invocation is modified to divert the return to a kprobe trampoline function called kretprobe_trampoline(). So with this the trace would (currently) look like:
Now this is quite misleading to the end user, as it suggests something that didn't actually happen. So just to avoid such misinterpretations, the inlined patch aims to output such a log as:
Jason Wessel [Tue, 27 May 2008 17:23:29 +0000 (12:23 -0500)]
softlockup: fix NMI hangs due to lock race - 2.6.26-rc regression
The touch_nmi_watchdog() routine on x86 ultimately calls
touch_softlockup_watchdog(). The problem is that to touch the
softlockup watchdog, the cpu_clock code has to be called which could
involve multiple cpu locks and can lead to a hard hang if one of the
locks is held by a processor that is not going to return anytime soon
(such as could be the case with kgdb or perhaps even with some other
kind of exception).
This patch causes the public version of the
touch_softlockup_watchdog() to defer the cpu clock access to a later
point.
The test case for this problem is to use the following kernel config
options:
It should be noted that kgdb test suite and these options were not
available until 2.6.26-rc2, so it was necessary to patch the kgdb
test suite during the bisection.
I would consider this patch a regression fix because the problem first
appeared in commit 27ec4407790d075c325e1f4da0a19c56953cce23 when some
logic was added to try to periodically sync the clocks. It was
possible to work around this particular problem by simply not
performing the sync anytime the system was in a critical context.
This was ok until commit 3e51f33fcc7f55e6df25d15b55ed10c8b4da84cd,
which added config option CONFIG_HAVE_UNSTABLE_SCHED_CLOCK and some
multi-cpu locks to sync the clocks. It became clear that accessing
this code from an nmi was the source of the lockups. Avoiding the
access to the low level clock code from an code inside the NMI
processing also fixed the problem with the 27ec44... commit.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Linus Torvalds [Tue, 27 May 2008 16:47:13 +0000 (09:47 -0700)]
x86: MMIO and gcc re-ordering issue
On Tue, 27 May 2008, Linus Torvalds wrote:
>
> Expecting people to fix up all drivers is simply not going to happen. And
> serializing things shouldn't be *that* expensive. People who cannot take
> the expense can continue to use the magic __raw_writel() etc stuff.
Of course, for non-x86, you kind of have to expect drivers to be
well-behaved, so non-x86 can probably avoid this simply because there are
less relevant drivers involved.
Here's a UNTESTED patch for x86 that may or may not compile and work, and
which serializes (on a compiler level) the IO accesses against regular
memory accesses.
__read[bwlq]()/__write[bwlq]() are not serialized with a :"memory"
barrier, although since they still use "asm volatile" I suspect that i
practice they are probably serial too. Did not look very closely at any
generated code (only did a trivial test to see that the code looks
*roughly* correct).
x86: pci-dma.c: use __GFP_NO_OOM instead of __GFP_NORETRY
On Wed, 2008-05-28 at 04:47 +0200, Andi Kleen wrote:
> > So... why not just remove the setting of __GFP_NORETRY? Why is it
> > wrong to oom-kill things in this case?
>
> When the 16MB zone overflows (which can be common in some workloads)
> calling the OOM killer is pretty useless because it has barely any
> real user data [only exception would be the "only 16MB" case Alan
> mentioned]. Killing random processes in this case is bad.
>
> I think for 16MB __GFP_NORETRY is ok because there should be
> nothing freeable in there so looping is useless. Only exception would be the
> "only 16MB total" case again but I'm not sure 2.6 supports that at all
> on x86.
>
> On the other hand d_a_c() does more allocations than just 16MB, especially
> on 64bit and the other zones need different strategies.
projects / linux-2.6-omap-h63xx.git / log