Krzysztof Helt [Tue, 31 Mar 2009 22:25:17 +0000 (15:25 -0700)]
cirrusfb: fix threshold register mask for Laguna chips
Fix threshold register mask for Laguna chips otherwise some 8bpp modes are
garbled after selecting a 24bpp mode.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:16 +0000 (15:25 -0700)]
cirrusfb: GD5434 (aka SD64) support fixed
Fix handling of the Cirrus Logic GD5434 chip. Distinguish this chip from the
GD5430. It allows detecting memory size for both models correctly.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:15 +0000 (15:25 -0700)]
cirrusfb: set MCLK in one place
A memory clock (MCLK) is set at various places. Move the setting into one
place.
Set the MCLK only for Zorro cards as the x86 cards should be initialized by
BIOS.
Improve handling of the GD5434 (SD64).
Kill one annoying debug output "virtual offset: ...".
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:15 +0000 (15:25 -0700)]
cirrusfb: add accelerator constant
Add an accelerator constant so almost all Cirrus are recognized as
accelerators by the fbset command.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:14 +0000 (15:25 -0700)]
cirrusfb: fix clock doubling
Cirrus' Alpine and Picasso4 chips uses DAC clock doubling to achieve full
range of pixclock frequencies.
[akpm@linux-foundation.org: fix spelling, use usual comment layout] Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:13 +0000 (15:25 -0700)]
cirrusfb: use 24bpp instead of 32bpp
The 32bpp is supported only on the latest Cirrus Logic chips. Use the
24bpp which is supported at least since Alpine chips (GD543x).
Change 32bpp mode setting to 24bpp mode. Change acceleration as well.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:13 +0000 (15:25 -0700)]
cirrusfb: GD5446 fixes
Various fixes to make Cirrus GD5446 chip work.
Another Cirrus chip works with the cirrusfb. The gd5446 seems very
similar to Alpine chips.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:12 +0000 (15:25 -0700)]
cirrusfb: fix error paths in cirrusfb_xxx_register()
Balance iomap and iounmap and alloc and free calls in case of error druing
device register (probing).
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:11 +0000 (15:25 -0700)]
cirrusfb: add imageblit function
Add hardware color expansion (imageblit) function. It roughly doubles
scrolling speed of my Alpine card (GD5430).
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:11 +0000 (15:25 -0700)]
cirrusfb: acceleration improvements
- Fix color expansion for 16bpp and 32bpp modes in the
cirrusfb_RectFill().
- Make a function with a common blitter code (cirrusfb_set_blitter).
- Add fb_sync function to allow a higher layer synchronize with the
blitter.
- Kill one redundant blitter reset.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:10 +0000 (15:25 -0700)]
cirrusfb: various Laguna fixes
- The Laguna GD5465 (AGP) has one register more than non-AGP chips.
Recognize the AGP version and write a tile control register only on the
AGP version. Tested only on an AGP card.
- Bump up RAMDAC frequencies after X11 code. This allow to drive a flat
panel resolution 1680x1050 at 16bpp from the 4MB card.
- Fix exit path in the cirrusfb_pci_register() in case of error.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:09 +0000 (15:25 -0700)]
cirrusfb: check_var improvements
Break cirrusfb_decode_var() function into two parts:
cirrusfb_check_pixclock() which can be called from the
cirrusfb_check_var() aand merge rest into the cirrusfb_set_par_foo().
This allows rejecting modes with too high pixclock before before any
change to hardware state (and a console is messed up).
Also, fix RGB field's lengths for 8bpp modes to correct ones so X11 works
with fbdev driver with cirrusfb.
Kill some redundant function calls or register loads.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:08 +0000 (15:25 -0700)]
cirrusfb: Laguna chipset 8bpp fix
Fix 8bpp mode by adding handling of the Laguna chipsets to various places
and stop trashing a HDR register which probably does not exist on the
Laguna.
Fix compilation warnings about uninitialized variables also.
Finally, all 8bpp, 16bpp and 32bpp modes work on the Laguna chipset.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:08 +0000 (15:25 -0700)]
cirrusfb: various improvements
Various improvements to the code:
- kill a structure with only one
field: multiplexing and use the
field directly
- move the cirrusfb_ops structure
down the file to kill forward
declarations
- move cirrusfb_init() to kill
forward declaration
- kill register loads done already
in the init_vgachip()
- kill assigments done by higher
layer in the cirrusfb_pan_display()
- do not overwrite line pitch bit in
the CL_CRT1D register
- kill btype variables if they were
used only once or twice
- add cpu_relax() in the busy waiting
loop
The fix to the CL_CRT1D register handling makess the 1024x768 32bpp mode
work. Previously, only lower resolution modes have worked with 32bpp.
Signed-off-by: Krzysztof Helt <krzysztof.h1@poczta.fm> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:07 +0000 (15:25 -0700)]
cirrusfb: use 5-6-5 RGB for 16bpp mode
Use the 5-6-5 RGB mode instead of the 5-5-5 mode at 16bpp depth.
It fixes colors in the 16bpp modes on Cirrus Laguna chips.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Cc: Geert Uytterhoeven <geert.uytterhoeven@gmail.com> Cc: Arthur Marsh <arthur.marsh@internode.on.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:06 +0000 (15:25 -0700)]
cirrusfb: do not calculate line length twice
A line length is calculated twice: first in the cirrusfb_decode_var() then
in the cirrusfb_set_par_foo().
Use the first calculated value. A nice side effect is that 32bpp mode
works now.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Cc: Geert Uytterhoeven <geert.uytterhoeven@gmail.com> Cc: Arthur Marsh <arthur.marsh@internode.on.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Krzysztof Helt [Tue, 31 Mar 2009 22:25:05 +0000 (15:25 -0700)]
cirrusfb: add mmio registers for Laguna chipsets
The Laguna chipsets use special registers which are available through the
mmio area. The cirrusfb driver does not use memory mapped registers for
the PCI cards.
Add the memory mapped area for Laguna chipsets and add basic usage of the
special Laguna registers after SVGALIB code.
This gives readable console at 16bpp on the GD-5465 (Laguna AGP). The
8bpp and 32bpp depths are still broken.
Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ian Kent [Tue, 31 Mar 2009 22:24:45 +0000 (15:24 -0700)]
autofs4: fix kernel includes
autofs_dev-ioctl.h is included by both the kernel module and user space tools
and it includes two kernel header files. Compiles work if the kernel headers
are installed but fail otherwise.
Signed-off-by: Ian Kent <raven@themaw.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ian Kent [Tue, 31 Mar 2009 22:24:45 +0000 (15:24 -0700)]
autofs4: fix lookup deadlock
A deadlock can occur when user space uses a signal (autofs version 4 uses
SIGCHLD for this) to effect expire completion.
The order of events is:
Expire process completes, but before being able to send SIGCHLD to it's parent
...
Another process walks onto a different mount point and drops the directory
inode semaphore prior to sending the request to the daemon as it must ...
A third process does an lstat on on the expired mount point causing it to wait
on expire completion (unfortunately) holding the directory semaphore.
The mount request then arrives at the daemon which does an lstat and,
deadlock.
For some time I was concerned about releasing the directory semaphore around
the expire wait in autofs4_lookup as well as for the mount call back. I
finally realized that the last round of changes in this function made the
expiring dentry and the lookup dentry separate and distinct so the check and
possible wait can be done anywhere prior to the mount call back. This patch
moves the check to just before the mount call back and inside the directory
inode mutex release.
Signed-off-by: Ian Kent <raven@themaw.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ian Kent [Tue, 31 Mar 2009 22:24:43 +0000 (15:24 -0700)]
autofs4: cleanup expire code duplication
A significant portion of the autofs_dev_ioctl_expire() and
autofs4_expire_multi() functions is duplicated code. This patch cleans that
up.
Signed-off-by: Ian Kent <raven@themaw.net> Signed-off-by: Jeff Moyer <jmoyer@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:40 +0000 (15:24 -0700)]
powerpc/fsl_soc: isolate legacy fsl_spi support to mpc832x_rdb boards
The advantages of this:
- Don't encourage legacy support;
- Less external symbols, less code to compile-in for !MPC832x_RDB
platforms.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:39 +0000 (15:24 -0700)]
powerpc/83xx: add mmc-spi support via the device tree for MPC8323E-RDB
- Add gpio-controller node to manage QE GPIO Bank D;
- Add mmc-spi node;
- Modify board file so that it won't use legacy SPI support with the new
device trees.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:38 +0000 (15:24 -0700)]
powerpc: add mmc-spi-slot bindings
The bindings describes a case where MMC/SD/SDIO slot directly connected to
a SPI bus. Such setups are widely used on embedded PowerPC boards.
The patch also adds the mmc-spi-slot entry to the OpenFirmware modalias
table.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:37 +0000 (15:24 -0700)]
spi_mpc83xx: add OF platform driver bindings
Implement full support for OF SPI bindings. Now the driver can manage its
own chip selects without any help from the board files and/or fsl_soc
constructors.
The "legacy" code is well isolated and could be removed as time goes by.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:36 +0000 (15:24 -0700)]
spi_mpc83xx: rework chip selects handling
The main purpose of this patch is to pass 'struct spi_device' to the chip
select handling routines. This is needed so that we could implement
full-fledged OpenFirmware support for this driver.
While at it, also:
- Replace two {de,activate}_cs routines by single cs_contol().
- Don't duplicate platform data callbacks in mpc83xx_spi struct.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Anton Vorontsov [Tue, 31 Mar 2009 22:24:35 +0000 (15:24 -0700)]
spi_mpc83xx: fix sparse warnings
The patch fixes following sparse warnings:
CHECK spi_mpc83xx.c
spi_mpc83xx.c:145:1: warning: symbol 'mpc83xx_spi_rx_buf_u8' was not declared. Should it be static?
spi_mpc83xx.c:146:1: warning: symbol 'mpc83xx_spi_rx_buf_u16' was not declared. Should it be static?
spi_mpc83xx.c:147:1: warning: symbol 'mpc83xx_spi_rx_buf_u32' was not declared. Should it be static?
spi_mpc83xx.c:148:1: warning: symbol 'mpc83xx_spi_tx_buf_u8' was not declared. Should it be static?
spi_mpc83xx.c:149:1: warning: symbol 'mpc83xx_spi_tx_buf_u16' was not declared. Should it be static?
spi_mpc83xx.c:150:1: warning: symbol 'mpc83xx_spi_tx_buf_u32' was not declared. Should it be static?
spi_mpc83xx.c:175:32: warning: incorrect type in initializer (different address spaces)
spi_mpc83xx.c:175:32: expected void *tmp_ptr
spi_mpc83xx.c:175:32: got unsigned int [noderef] <asn:2>*<noident>
spi_mpc83xx.c:183:26: warning: incorrect type in argument 1 (different address spaces)
spi_mpc83xx.c:183:26: expected unsigned int [noderef] [usertype] <asn:2>*reg
spi_mpc83xx.c:183:26: got void *tmp_ptr
spi_mpc83xx.c:184:26: warning: incorrect type in argument 1 (different address spaces)
spi_mpc83xx.c:184:26: expected unsigned int [noderef] [usertype] <asn:2>*reg
spi_mpc83xx.c:184:26: got void *tmp_ptr
spi_mpc83xx.c:287:31: warning: incorrect type in initializer (different address spaces)
spi_mpc83xx.c:287:31: expected void *tmp_ptr
spi_mpc83xx.c:287:31: got unsigned int [noderef] <asn:2>*<noident>
spi_mpc83xx.c:295:25: warning: incorrect type in argument 1 (different address spaces)
spi_mpc83xx.c:295:25: expected unsigned int [noderef] [usertype] <asn:2>*reg
spi_mpc83xx.c:295:25: got void *tmp_ptr
spi_mpc83xx.c:296:25: warning: incorrect type in argument 1 (different address spaces)
spi_mpc83xx.c:296:25: expected unsigned int [noderef] [usertype] <asn:2>*reg
spi_mpc83xx.c:296:25: got void *tmp_ptr
spi_mpc83xx.c:486:13: warning: symbol 'mpc83xx_spi_irq' was not declared. Should it be static?
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Cc: David Brownell <david-b@pacbell.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Kumar Gala <galak@gate.crashing.org> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
"I use ramfs instead of tmpfs for /tmp because I don't use swap on my
laptop. Some apps need 1777 mode for /tmp directory, but ramfs does not
support 'mode=' mount option."
Daniel Mack [Tue, 31 Mar 2009 22:24:33 +0000 (15:24 -0700)]
lis3: SPI transport layer
Make use of the new abstraction layer and add a new transport layer for
spi. Works fine on a PXA based board.
Signed-off-by: Daniel Mack <daniel@caiaq.de> Acked-by: Pavel Machek <pavel@ucw.cz> Acked-by: Eric Piel <eric.piel@tremplin-utc.net> Cc: David Brownell <david-b@pacbell.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Daniel Mack [Tue, 31 Mar 2009 22:24:32 +0000 (15:24 -0700)]
lis3: solve dependency between core and ACPI
This solves the dependency between lis3lv02d.[ch] and ACPI specific
methods. It introduces a ->bus_priv pointer to the device struct which is
casted to 'struct acpi_device' in the ACIP layer. Changed hp_accel.c
accordingly.
Signed-off-by: Daniel Mack <daniel@caiaq.de> Acked-by: Pavel Machek <pavel@ucw.cz> Acked-by: Eric Piel <eric.piel@tremplin-utc.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Daniel Mack [Tue, 31 Mar 2009 22:24:31 +0000 (15:24 -0700)]
lis3: reorder functions to make forward decl obsolete
Move lis3lv02d_init_device() down so that the forward declaration of
lis3lv02d_add_fs() becomes unnecessary.
Signed-off-by: Daniel Mack <daniel@caiaq.de> Acked-by: Pavel Machek <pavel@ucw.cz> Acked-by: Eric Piel <eric.piel@tremplin-utc.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Luca Cappa [Tue, 31 Mar 2009 22:24:31 +0000 (15:24 -0700)]
hp_accel: axis conversion for hp compaq 8710w
I have a laptop HP Compaq 8710W, I compiled into my kernel the LIS3LV02DL
and HP_ACCEL module drivers. While loading it cannot recognize the laptop
model, so i am sending the necessary information to update the database of
axis orientations.
>When the laptop is horizontal the position reported is about 0 for X and Y
>and a positive value for Z
Yes, it is about 0,0,1000, the actual reading says: (-17,-26,1018);
> If the left side is elevated, X increases (becomes positive)
Yes, X goes toward to positive 1000.
>If the front side (where the touchpad is) is elevated, Y decreases (becomes negative)
No, Y goes toward to positive 1000.
>If the laptop is put upside-down, Z becomes negative
Yes, the laptop on a table Z gives 1000, and if upsidedown the Z reads
-1000.
So in few words the Y axis is inverted.
Cc: Eric Piel <eric.piel@tremplin-utc.net> Signed-off-by: Pavel Machek <pavel@ucw.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pavel Machek [Tue, 31 Mar 2009 22:24:29 +0000 (15:24 -0700)]
hp_accel: add two more axis information
Add two more laptops to whitelist.
Signed-off-by: Michal Marek <mmarek@suse.cz> Signed-off-by: Pavel Machek <pavel@ucw.cz> Cc: Daniel Mack <daniel@caiaq.de> Cc: Eric Piel <eric.piel@tremplin-utc.net> Cc: Vladimir Botka <vbotka@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ira Snyder [Tue, 31 Mar 2009 22:24:29 +0000 (15:24 -0700)]
hwmon: Add LTC4215 driver
Add Linux support for the Linear Technology LTC4215 Hot Swap controller
I2C monitoring interface.
I have tested the driver with my board, and it appears to work fine. With
the power supplies disabled, it reads 11.93V input, 1.93V output, no
current and no power. With the supplies enabled, it reads 11.93V input,
11.98V output, no current, no power. I'm not drawing any current at the
moment, so this is reasonable. The value in the sense register never
reads anything except 0, so I expect to get zero from the current and
power calculations.
I didn't attempt to support changing any of the chip's settings or
enabling the FET. I'm not sure even how to do that and still fit within
the hwmon framework. :)
Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu> Cc: Jean Delvare <khali@linux-fr.org> Cc: "Mark M. Hoffman" <mhoffman@lightlink.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:24 +0000 (15:24 -0700)]
epoll keyed wakeups: make tty use keyed wakeups
Introduce keyed event wakeups inside the TTY code.
Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:23 +0000 (15:24 -0700)]
epoll keyed wakeups: make eventfd use keyed wakeups
Introduce keyed event wakeups inside the eventfd code.
Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:22 +0000 (15:24 -0700)]
epoll keyed wakeups: teach epoll about hints coming with the wakeup key
Use the events hint now sent by some devices, to avoid unnecessary wakeups
for events that are of no interest for the caller. This code handles both
devices that are sending keyed events, and the ones that are not (and
event the ones that sometimes send events, and sometimes don't).
[akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:21 +0000 (15:24 -0700)]
epoll keyed wakeups: make sockets use keyed wakeups
Add support for event-aware wakeups to the sockets code. Events are
delivered to the wakeup target, so that epoll can avoid spurious wakeups
for non-interesting events.
Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:20 +0000 (15:24 -0700)]
epoll keyed wakeups: introduce new *_poll() wakeup macros
Introduce new wakeup macros that allow passing an event mask to the wakeup
targets. They exactly mimic their non-_poll() counterpart, with the added
event mask passing capability. I did add only the ones currently
requested, avoiding the _nr() and _all() for the moment.
Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:20 +0000 (15:24 -0700)]
epoll keyed wakeups: add __wake_up_locked_key() and __wake_up_sync_key()
This patchset introduces wakeup hints for some of the most popular (from
epoll POV) devices, so that epoll code can avoid spurious wakeups on its
waiters.
The problem with epoll is that the callback-based wakeups do not, ATM,
carry any information about the events the wakeup is related to. So the
only choice epoll has (not being able to call f_op->poll() from inside the
callback), is to add the file* to a ready-list and resolve the real events
later on, at epoll_wait() (or its own f_op->poll()) time. This can cause
spurious wakeups, since the wake_up() itself might be for an event the
caller is not interested into.
The rate of these spurious wakeup can be pretty high in case of many
network sockets being monitored.
By allowing devices to report the events the wakeups refer to (at least
the two major classes - POLLIN/POLLOUT), we are able to spare useless
wakeups by proper handling inside the epoll's poll callback.
Epoll will have in any case to call f_op->poll() on the file* later on,
since the change to be done in order to have the full event set sent via
wakeup, is too invasive for the way our f_op->poll() system works (the
full event set is calculated inside the poll function - there are too many
of them to even start thinking the change - also poll/select would need
change too).
Epoll is changed in a way that both devices which send event hints, and
the ones that don't, are correctly handled. The former will gain some
efficiency though.
As a general rule for devices, would be to add an event mask by using
key-aware wakeup macros, when making up poll wait queues. I tested it
(together with the epoll's poll fix patch Andrew has in -mm) and wakeups
for the supported devices are correctly filtered.
Test program available here:
http://www.xmailserver.org/epoll_test.c
This patch:
Nothing revolutionary here. Just using the available "key" that our
wakeup core already support. The __wake_up_locked_key() was no brainer,
since both __wake_up_locked() and __wake_up_locked_key() are thin wrappers
around __wake_up_common().
The __wake_up_sync() function had a body, so the choice was between
borrowing the body for __wake_up_sync_key() and calling it from
__wake_up_sync(), or make an inline and calling it from both. I chose the
former since in most archs it all resolves to "mov $0, REG; jmp ADDR".
Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: William Lee Irwin III <wli@movementarian.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:18 +0000 (15:24 -0700)]
eventfd: improve support for semaphore-like behavior
People started using eventfd in a semaphore-like way where before they
were using pipes.
That is, counter-based resource access. Where a "wait()" returns
immediately by decrementing the counter by one, if counter is greater than
zero. Otherwise will wait. And where a "post(count)" will add count to
the counter releasing the appropriate amount of waiters. If eventfd the
"post" (write) part is fine, while the "wait" (read) does not dequeue 1,
but the whole counter value.
The problem with eventfd is that a read() on the fd returns and wipes the
whole counter, making the use of it as semaphore a little bit more
cumbersome. You can do a read() followed by a write() of COUNTER-1, but
IMO it's pretty easy and cheap to make this work w/out extra steps. This
patch introduces a new eventfd flag that tells eventfd to only dequeue 1
from the counter, allowing simple read/write to make it behave like a
semaphore. Simple test here:
http://www.xmailserver.org/eventfd-sem.c
To be back-compatible with earlier kernels, userspace applications should
probe for the availability of this feature via
Tony Battersby [Tue, 31 Mar 2009 22:24:15 +0000 (15:24 -0700)]
epoll: clean up ep_modify
ep_modify() doesn't need to set event.data from within the ep->lock
spinlock as the comment suggests. The only place event.data is used is
ep_send_events_proc(), and this is protected by ep->mtx instead of
ep->lock. Also update the comment for mutex_lock() at the top of
ep_scan_ready_list(), which mentions epoll_ctl(EPOLL_CTL_DEL) but not
epoll_ctl(EPOLL_CTL_MOD).
ep_modify() can also use spin_lock_irq() instead of spin_lock_irqsave().
Signed-off-by: Tony Battersby <tonyb@cybernetics.com> Acked-by: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tony Battersby [Tue, 31 Mar 2009 22:24:15 +0000 (15:24 -0700)]
epoll: remove unnecessary xchg
xchg in ep_unregister_pollwait() is unnecessary because it is protected by
either epmutex or ep->mtx (the same protection as ep_remove()).
If xchg was necessary, it would be insufficient to protect against
problems: if multiple concurrent calls to ep_unregister_pollwait() were
possible then a second caller that returns without doing anything because
nwait == 0 could return before the waitqueues are removed by the first
caller, which looks like it could lead to problematic races with
ep_poll_callback().
So remove xchg and add comments about the locking.
Signed-off-by: Tony Battersby <tonyb@cybernetics.com> Acked-by: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tony Battersby [Tue, 31 Mar 2009 22:24:14 +0000 (15:24 -0700)]
epoll: remember the event if epoll_wait returns -EFAULT
If epoll_wait returns -EFAULT, the event that was being returned when the
fault was encountered will be forgotten. This is not a big deal since
EFAULT will happen only if a buggy userspace program passes in a bad
address, in which case what happens later usually doesn't matter.
However, it is easy to remember the event for later, and this patch makes
a simple change to do that.
Signed-off-by: Tony Battersby <tonyb@cybernetics.com> Acked-by: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tony Battersby [Tue, 31 Mar 2009 22:24:13 +0000 (15:24 -0700)]
epoll: don't use current in irq context
ep_call_nested() (formerly ep_poll_safewake()) uses "current" (without
dereferencing it) to detect callback recursion, but it may be called from
irq context where the use of current is generally discouraged. It would
be better to use get_cpu() and put_cpu() to detect the callback recursion.
Signed-off-by: Tony Battersby <tonyb@cybernetics.com> Acked-by: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Davide Libenzi [Tue, 31 Mar 2009 22:24:10 +0000 (15:24 -0700)]
epoll: fix epoll's own poll
Fix a bug inside the epoll's f_op->poll() code, that returns POLLIN even
though there are no actual ready monitored fds. The bug shows up if you
add an epoll fd inside another fd container (poll, select, epoll).
The problem is that callback-based wake ups used by epoll does not carry
(patches will follow, to fix this) any information about the events that
actually happened. So the callback code, since it can't call the file*
->poll() inside the callback, chains the file* into a ready-list.
So, suppose you added an fd with EPOLLOUT only, and some data shows up on
the fd, the file* mapped by the fd will be added into the ready-list (via
wakeup callback). During normal epoll_wait() use, this condition is
sorted out at the time we're actually able to call the file*'s
f_op->poll().
Inside the old epoll's f_op->poll() though, only a quick check
!list_empty(ready-list) was performed, and this could have led to
reporting POLLIN even though no ready fds would show up at a following
epoll_wait(). In order to correctly report the ready status for an epoll
fd, the ready-list must be checked to see if any really available fd+event
would be ready in a following epoll_wait().
Operation (calling f_op->poll() from inside f_op->poll()) that, like wake
ups, must be handled with care because of the fact that epoll fds can be
added to other epoll fds.
Test code:
/*
* epoll_test by Davide Libenzi (Simple code to test epoll internals)
* Copyright (C) 2008 Davide Libenzi
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* Davide Libenzi <davidel@xmailserver.org>
*
*/
for (i = 0; i < tcfg->size; i++) {
nfd = xepoll_create(1);
xepoll_ctl(bfd, EPOLL_CTL_ADD, nfd, &evt);
bfd = nfd;
}
xpipe(pfds);
if (tcfg->flags & EPOLL_TF_LOOP)
{
xepoll_ctl(bfd, EPOLL_CTL_ADD, epfd, &evt);
/*
* If we're testing for loop, we want that the wakeup
* triggered by the write to the pipe done in the child
* process, triggers a fake event. So we add the pipe
* read size with EPOLLOUT events. This will trigger
* an addition to the ready-list, but no real events
* will be there. The the epoll kernel code will proceed
* in calling f_op->poll() of the epfd, triggering the
* loop we want to test.
*/
evt.events = EPOLLOUT;
}
xepoll_ctl(bfd, EPOLL_CTL_ADD, pfds[0], &evt);
/*
* The pipe write must come after the poll(2) call inside
* check_events(). This tests the nested wakeup code in
* fs/eventpoll.c:ep_poll_safewake()
* By having the check_events() (hence poll(2)) happens first,
* we have poll wait queue filled up, and the write(2) in the
* child will trigger the wakeup chain.
*/
if ((pid = xfork()) == 0) {
sleep(1);
write(pfds[1], "w", 1);
exit(0);
}
for (i = 0; i < tcfg->size; i++) {
nfd = xepoll_create(1);
xepoll_ctl(bfd, EPOLL_CTL_ADD, nfd, &evt);
bfd = nfd;
}
xpipe(pfds);
if (tcfg->flags & EPOLL_TF_LOOP)
{
xepoll_ctl(bfd, EPOLL_CTL_ADD, epfd, &evt);
/*
* If we're testing for loop, we want that the wakeup
* triggered by the write to the pipe done in the child
* process, triggers a fake event. So we add the pipe
* read size with EPOLLOUT events. This will trigger
* an addition to the ready-list, but no real events
* will be there. The the epoll kernel code will proceed
* in calling f_op->poll() of the epfd, triggering the
* loop we want to test.
*/
evt.events = EPOLLOUT;
}
xepoll_ctl(bfd, EPOLL_CTL_ADD, pfds[0], &evt);
/*
* The pipe write mush come before the poll(2) call inside
* check_events(). This tests the nested f_op->poll calls code in
* fs/eventpoll.c:ep_eventpoll_poll()
* By having the pipe write(2) happen first, we make the kernel
* epoll code to load the ready lists, and the following poll(2)
* done inside check_events() will test nested poll code in
* ep_eventpoll_poll().
*/
if ((pid = xfork()) == 0) {
write(pfds[1], "w", 1);
exit(0);
}
sleep(1);
res = check_events(epfd, 1000) & POLLIN;
Daniel Mack [Tue, 31 Mar 2009 22:23:53 +0000 (15:23 -0700)]
drivers/misc/isl29003.c: driver for the ISL29003 ambient light sensor
Add a driver for Intersil's ISL29003 ambient light sensor device plus some
documentation. Inspired by tsl2550.c, a driver for a similar device.
It is put in drivers/misc for now until the industrial I/O framework gets
merged.
Signed-off-by: Daniel Mack <daniel@caiaq.de> Acked-by: Jonathan Cameron <jic23@cam.ac.uk> Cc: Jean Delvare <khali@linux-fr.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David Altobelli [Tue, 31 Mar 2009 22:23:53 +0000 (15:23 -0700)]
hpilo: reduce frequency of IO operations
Change hpilo open and close logic to spin for 10usec between checking device,
rather than every usec.
Because the loop is coded to take up to 10ms, it seemed prudent to
increase the interval between polling the device, to reduce the load on
the system and allow more other work to happen.
Signed-off-by: David Altobelli <david.altobelli@hp.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Michael Buesch [Tue, 31 Mar 2009 22:23:49 +0000 (15:23 -0700)]
bcm47xx: fix GPIO API return codes
The GPIO API is supposed to return 0 or a negative error code,
but the SSB GPIO functions return the bitmask of the GPIO register.
Fix this by ignoring the bitmask and always returning 0. The SSB GPIO functions can't fail.
Signed-off-by: Michael Buesch <mb@bu3sch.de> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: David Brownell <david-b@pacbell.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Eric Sandeen [Tue, 31 Mar 2009 22:23:46 +0000 (15:23 -0700)]
filesystem freeze: allow SysRq emergency thaw to thaw frozen filesystems
Now that the filesystem freeze operation has been elevated to the VFS, and
is just an ioctl away, some sort of safety net for unintentionally frozen
root filesystems may be in order.
The timeout thaw originally proposed did not get merged, but perhaps
something like this would be useful in emergencies.
For example, freeze /path/to/mountpoint may freeze your root filesystem if
you forgot that you had that unmounted.
I chose 'j' as the last remaining character other than 'h' which is sort
of reserved for help (because help is generated on any unknown character).
I've tested this on a non-root fs with multiple (nested) freezers, as well
as on a system rendered unresponsive due to a frozen root fs.
[randy.dunlap@oracle.com: emergency thaw only if CONFIG_BLOCK enabled] Signed-off-by: Eric Sandeen <sandeen@redhat.com> Cc: Takashi Sato <t-sato@yk.jp.nec.com> Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Wolfram Strepp [Tue, 31 Mar 2009 22:23:45 +0000 (15:23 -0700)]
lib/rbtree.c: optimize rb_erase()
Tfour 4 redundant if-conditions in function __rb_erase_color() in
lib/rbtree.c are removed.
In pseudo-source-code, the structure of the code is as follows:
if ((!A || B) && (!C || D)) {
.
.
.
} else {
if (!C || D) {//if this is true, it implies: (A == true) && (B == false)
if (A) {//hence this always evaluates to 'true'...
.
}
.
//at this point, C always becomes true, because of:
__rb_rotate_right/left();
//and:
other = parent->rb_right/left;
}
.
.
if (C) {//...and this too !
.
}
}
Signed-off-by: Wolfram Strepp <wstrepp@gmx.de> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <andrea@qumranet.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
J. R. Okajima [Tue, 31 Mar 2009 22:23:43 +0000 (15:23 -0700)]
loop: add ioctl to resize a loop device
Add the ability to 'resize' the loop device on the fly.
One practical application is a loop file with XFS filesystem, already
mounted: You can easily enlarge the file (append some bytes) and then call
ioctl(fd, LOOP_SET_CAPACITY, new); The loop driver will learn about the
new size and you can use xfs_growfs later on, which will allow you to use
full capacity of the loop file without the need to unmount.
void usage(FILE *f)
{
fprintf(f, "%s [options] loop_dev [backend_file]\n"
"-s, --set new_size_in_bytes\n"
"\twhen backend_file is given, "
"it will be expanded too while keeping the original contents\n",
me);
}
WANG Cong [Tue, 31 Mar 2009 22:23:41 +0000 (15:23 -0700)]
uml: remove useless comments
These comments are useless now, remove them.
Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
WANG Cong [Tue, 31 Mar 2009 22:23:41 +0000 (15:23 -0700)]
uml: improve error messages
These error messages are from check_sysemu(), not check_ptrace().
Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
WANG Cong [Tue, 31 Mar 2009 22:23:40 +0000 (15:23 -0700)]
uml: don't use a too long string literal
uml uses a concatenated string literal to store the contents of .config,
but .config file content is varaible, it can be very long.
Use an array of string literals instead.
Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
MAJOR_NR isn't needed anymore since very early 2.5 kernels.
[akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Christoph Hellwig <hch@lst.de> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Magnus Damm [Tue, 31 Mar 2009 22:23:37 +0000 (15:23 -0700)]
pm: rework includes, remove arch ifdefs
Make the following header file changes:
- remove arch ifdefs and asm/suspend.h from linux/suspend.h
- add asm/suspend.h to disk.c (for arch_prepare_suspend())
- add linux/io.h to swsusp.c (for ioremap())
- x86 32/64 bit compile fixes
Signed-off-by: Magnus Damm <damm@igel.co.jp> Cc: Paul Mundt <lethal@linux-sh.org> Acked-by: "Rafael J. Wysocki" <rjw@sisk.pl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Randy Dunlap [Tue, 31 Mar 2009 22:23:36 +0000 (15:23 -0700)]
alpha: convert u64 to unsigned long long
Convert alpha architecture to use u64 as unsigned long long. This is
being done so that (a) all arches use u64 as unsigned long long and (b)
printk of a u64 as %ll[ux] will not generate format warnings by gcc.
The only gcc cross-compiler that I have is 4.0.2, which generates errors
about miscompiling __weak references, so I have commented out that line in
compiler-gcc4.h so that most of these compile, but more builds and real
machine testing would be Real Good.
[akpm@linux-foundation.org: fix warning]
[akpm@linux-foundation.org: fix build]
[akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
From: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ivan Kokshaysky [Tue, 31 Mar 2009 22:23:35 +0000 (15:23 -0700)]
alpha: xchg/cmpxchg cleanup and fixes
- "_local" versions of xchg/cmpxchg functions duplicate code
of non-local ones (quite a few pages of assembler), except
memory barriers. We can generate these two variants from a
single header file using simple macros;
- convert xchg macro back to inline function using always_inline
attribute;
- use proper argument types for cmpxchg_u8/u16 functions
to fix a problem with negative arguments.
Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Richard Henderson <rth@twiddle.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Hugh Dickins [Tue, 31 Mar 2009 22:23:33 +0000 (15:23 -0700)]
shmem: writepage directly to swap
Synopsis: if shmem_writepage calls swap_writepage directly, most shmem
swap loads benefit, and a catastrophic interaction between SLUB and some
flash storage is avoided.
shmem_writepage() has always been peculiar in making no attempt to write:
it has just transferred a shmem page from file cache to swap cache, then
let that page make its way around the LRU again before being written and
freed.
The idea was that people use tmpfs because they want those pages to stay
in RAM; so although we give it an overflow to swap, we should resist
writing too soon, giving those pages a second chance before they can be
reclaimed.
That was always questionable, and I've toyed with this patch for years;
but never had a clear justification to depart from the original design.
It became more questionable in 2.6.28, when the split LRU patches classed
shmem and tmpfs pages as SwapBacked rather than as file_cache: that in
itself gives them more resistance to reclaim than normal file pages. I
prepared this patch for 2.6.29, but the merge window arrived before I'd
completed gathering statistics to justify sending it in.
Then while comparing SLQB against SLUB, running SLUB on a laptop I'd
habitually used with SLAB, I found SLUB to run my tmpfs kbuild swapping
tests five times slower than SLAB or SLQB - other machines slower too, but
nowhere near so bad. Simpler "cp -a" swapping tests showed the same.
slub_max_order=0 brings sanity to all, but heavy swapping is too far from
normal to justify such a tuning. The crucial factor on that laptop turns
out to be that I'm using an SD card for swap. What happens is this:
By default, SLUB uses order-2 pages for shmem_inode_cache (and many other
fs inodes), so creating tmpfs files under memory pressure brings lumpy
reclaim into play. One subpage of the order is chosen from the bottom of
the LRU as usual, then the other three picked out from their random
positions on the LRUs.
In a tmpfs load, many of these pages will be ones which already passed
through shmem_writepage, so already have swap allocated. And though their
offsets on swap were probably allocated sequentially, now that the pages
are picked off at random, their swap offsets are scattered.
But the flash storage on the SD card is very sensitive to having its
writes merged: once swap is written at scattered offsets, performance
falls apart. Rotating disk seeks increase too, but less disastrously.
So: stop giving shmem/tmpfs pages a second pass around the LRU, write them
out to swap as soon as their swap has been allocated.
It's surely possible to devise an artificial load which runs faster the
old way, one whose sizing is such that the tmpfs pages on their second
pass are the ones that are wanted again, and other pages not.
But I've not yet found such a load: on all machines, under the loads I've
tried, immediate swap_writepage speeds up shmem swapping: especially when
using the SLUB allocator (and more effectively than slub_max_order=0), but
also with the others; and it also reduces the variance between runs. How
much faster varies widely: a factor of five is rare, 5% is common.
One load which might have suffered: imagine a swapping shmem load in a
limited mem_cgroup on a machine with plenty of memory. Before 2.6.29 the
swapcache was not charged, and such a load would have run quickest with
the shmem swapcache never written to swap. But now swapcache is charged,
so even this load benefits from shmem_writepage directly to swap.
Apologies for the #ifndef CONFIG_SWAP swap_writepage() stub in swap.h:
it's silly because that will never get called; but refactoring shmem.c
sensibly according to CONFIG_SWAP will be a separate task.
Signed-off-by: Hugh Dickins <hugh@veritas.com> Acked-by: Pekka Enberg <penberg@cs.helsinki.fi> Acked-by: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
try_to_free_pages() is used for the direct reclaim of up to
SWAP_CLUSTER_MAX pages when watermarks are low. The caller to
alloc_pages_nodemask() can specify a nodemask of nodes that are allowed to
be used but this is not passed to try_to_free_pages(). This can lead to
unnecessary reclaim of pages that are unusable by the caller and int the
worst case lead to allocation failure as progress was not been make where
it is needed.
This patch passes the nodemask used for alloc_pages_nodemask() to
try_to_free_pages().
Johannes Weiner [Tue, 31 Mar 2009 22:23:29 +0000 (15:23 -0700)]
ramfs-nommu: use generic lru cache
Instead of open-coding the lru-list-add pagevec batching when expanding a
file mapping from zero, defer to the appropriate page cache function that
also takes care of adding the page to the lru list.
This is cleaner, saves code and reduces the stack footprint by 16 words
worth of pagevec.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: David Howells <dhowells@redhat.com> Cc: Nick Piggin <npiggin@suse.de> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Rik van Riel <riel@redhat.com> Cc: Peter Zijlstra <peterz@infradead.com> Cc: MinChan Kim <minchan.kim@gmail.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com> Cc: Greg Ungerer <gerg@snapgear.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
David Rientjes [Tue, 31 Mar 2009 22:23:29 +0000 (15:23 -0700)]
vmscan: print shrink_slab symbol name on negative shrinker objects
When a shrinker has a negative number of objects to delete, the symbol
name of the shrinker should be printed, not shrink_slab. This also makes
the error message slightly more informative.
Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
projects / mv-sheeva.git / log