openwrt/staging/blogic.git
14 years agoMIPS: Support 36-bit iomem on 32-bit Au1x00
pascal@pabr.org [Sun, 3 Jan 2010 12:39:12 +0000 (13:39 +0100)]
MIPS: Support 36-bit iomem on 32-bit Au1x00

I believe these changes are needed on Alchemy SoCs in order to
use iomem above 4G with the usual platform_device machinery:

- Set CONFIG_ARCH_PHYS_ADDR_T_64BIT to make resource_size_t 64-bit.
- Increase IOMEM_RESOURCE_END so that platforms can register resources.

To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/814/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Cleanup the halt and poweroff action
Wu Zhangjin [Wed, 27 Jan 2010 14:39:46 +0000 (22:39 +0800)]
MIPS: Loongson: Cleanup the halt and poweroff action

In the old source code, I have let halt and poweroff do the same action,
but in reality, they have different meanings.

As the manpage of shutdown shows:

-r     Reboot after shutdown.
-H     Halt action is to halt or drop into boot monitor on systems that support it.
-P     Halt action is to turn off the power.

and in the real world, some machines(e.g. NAS) did not provide a power
button and the shutdown works as reset, so, we need to provide a
mechanism to let the users turn off the power safely without breaking
the system, such a mechanism is "halt", which only put the system into a
dead loop or a power-save mode and print some information to the screen
to tell the users to turn off the power safely.

$ shutdown -hH now /* loongson_halt, not turn off the power */
$ shutdown -hP now /* loongson_poweroff, work as poweroff */

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Tested-by: Liu Shiwei <liushiwei@gmail.com>
Cc: Liu Shiwei <liushiwei@gmail.com>
Cc: linux-mips <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/883/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Simplify the weak annotation with __weak
Wu Zhangjin [Tue, 26 Jan 2010 15:02:34 +0000 (23:02 +0800)]
MIPS: Simplify the weak annotation with __weak

Found by

$ find arch/mips/ -name "*.c" | xargs -i grep -H weak {} | grep -v __weak

[Ralf: Made this bulletproof by including <linux/compiler.h>]

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/874/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: debug output for compressed kernels
Manuel Lauss [Wed, 13 Jan 2010 17:46:58 +0000 (18:46 +0100)]
MIPS: Alchemy: debug output for compressed kernels

Hook up the compressed debug output for all Alchemy systems supported
by current kernel codebase.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
To: Linux-MIPS <linux-mips@linux-mips.org>
Cc: Wu Zhangjin <wuzhangjin@gmail.com>
Patchwork: http://patchwork.linux-mips.org/patch/879/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: AR7: Implement clock API
Florian Fainelli [Wed, 27 Jan 2010 08:10:06 +0000 (09:10 +0100)]
MIPS: AR7: Implement clock API

This patch makes the ar7 clock code implement the Linux clk API. Drivers
using the various clocks available in the SoC are updated accordingly.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Acked-by: Wim Van Sebroeck <wim@iguana.be>
To: linux-mips@linux-mips.org
Cc: Wim Van Sebroeck <wim@iguana.be>
Cc: netdev@vger.kernel.org
Cc: David Miller <davem@davemloft.net>
Patchwork: http://patchwork.linux-mips.org/patch/881/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: AR7: Implement gpiolib
Florian Fainelli [Sun, 3 Jan 2010 20:16:51 +0000 (21:16 +0100)]
MIPS: AR7: Implement gpiolib

This patch implements gpiolib for the AR7 SoC.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/816/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: msp71xx: remove unused prom_getcmdline()
Yoichi Yuasa [Tue, 26 Jan 2010 09:02:58 +0000 (18:02 +0900)]
MIPS: msp71xx: remove unused prom_getcmdline()

Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
Cc: linux-mips <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/868/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Cleanup the Makefile of compressed kernel support
Wu Zhangjin [Tue, 26 Jan 2010 09:04:02 +0000 (17:04 +0800)]
MIPS: Cleanup the Makefile of compressed kernel support

This patch removes a useless "\" (line break) and tunes the format of a
long line.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/869/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: AR7: replace prom_getcmdline() to arcs_cmdline[]
Yoichi Yuasa [Tue, 26 Jan 2010 09:08:34 +0000 (18:08 +0900)]
MIPS: AR7: replace prom_getcmdline() to arcs_cmdline[]

Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
Cc: linux-mips <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/872/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: AR7: use strlcat() for the command line arguments
Yoichi Yuasa [Tue, 26 Jan 2010 09:07:02 +0000 (18:07 +0900)]
MIPS: AR7: use strlcat() for the command line arguments

Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
Cc: linux-mips <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/871/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Nuke trailing blank lines
Ralf Baechle [Sat, 27 Feb 2010 11:53:14 +0000 (12:53 +0100)]
MIPS: Nuke trailing blank lines

Recent git versions now warn about those and they've always been a bit of
an annoyance.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Cleanup switches with cases that can be merged
Roel Kluin [Tue, 19 Jan 2010 23:59:27 +0000 (00:59 +0100)]
MIPS: Cleanup switches with cases that can be merged

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
To: linux-mips@linux-mips.org
To: Andrew Morton <akpm@linux-foundation.org>
To: LKML <linux-kernel@vger.kernel.org>
Patchwork: http://patchwork.linux-mips.org/patch/860/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Decode c0_config4 for large TLBs.
David Daney [Fri, 22 Jan 2010 22:41:15 +0000 (14:41 -0800)]
MIPS: Decode c0_config4 for large TLBs.

For processors that have more than 64 TLBs, we need to decode both
config1 and config4 to determine the total number TLBs.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/866/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Remove probe_tlb().
David Daney [Fri, 22 Jan 2010 22:41:14 +0000 (14:41 -0800)]
MIPS: Remove probe_tlb().

The function probe_tlb() only does anything for processors that are
not PRID_COMP_LEGACY.  This is precisely the set of processors for
which decode_configs() is called to do identical tlbsize probing
calculations.  Therefore probe_tlb() is completely redundant and may
be removed.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/865/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: add readl/write_be accessors
Florian Fainelli [Wed, 16 Dec 2009 10:29:06 +0000 (11:29 +0100)]
MIPS: add readl/write_be accessors

MIPS currently lacks the readl_be and writel_be accessors
which are required by BCM63xx for OHCI and EHCI support.
Let's define them globally for MIPS. This also fixes the
compilation of the bcm63xx defconfig against USB.

Signed-off-by: Florian Fainelli <ffainelli@freebox.fr>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: linux-mips@linux-mips.org
Cc: Maxime Bizon <mbizon@freebox.fr>
Patchwork: http://patchwork.linux-mips.org/patch/793/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Lemote-2F: update defconfig
Wu Zhangjin [Mon, 4 Jan 2010 09:16:52 +0000 (17:16 +0800)]
MIPS: Loongson: Lemote-2F: update defconfig

Changes:

  o Serial port related configuration
    Disable EARLY_PRINTK, CONFIG_SYS_SUPPORTS_ZBOOT_UART16550
    Enable the serial port support as module.
  o PM related support
    Enable CPUFreq as module, use the external timer(MFGPT) instead of
    r4k timer.
    Enable Suspend support
    Enable Run Time PM support
  o Enable SM7XX Video Driver
    Disable the buggy 2d acceleration
  o Enable CONFIG_OPROFILE as module
  o Use GZIP instead of LZMA, which need less decompression time
  o Enable more USB devices support
  o Enable initrd support(needed by gNewsense)
  o Enable more crypto support

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/830/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Change the Email address of Wu Zhangjin
Wu Zhangjin [Mon, 4 Jan 2010 09:16:51 +0000 (17:16 +0800)]
MIPS: Loongson: Change the Email address of Wu Zhangjin

Currently wuzj@lemote.com is not usable; change it to wuzhangjin@gmail.com.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/829/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Fixup mem.c indentation
Wu Zhangjin [Mon, 4 Jan 2010 09:16:50 +0000 (17:16 +0800)]
MIPS: Loongson: Fixup mem.c indentation

Replace whitespace by tabs.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/828/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: arch/mips/Makefile: Add missing whitespace
Wu Zhangjin [Mon, 4 Jan 2010 09:16:49 +0000 (17:16 +0800)]
MIPS: Loongson: arch/mips/Makefile: Add missing whitespace

This patch add missing whitespace after every "+=" in the loongson
related part of arch/mips/Makefile.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/827/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Cleanup of the environment variables
Wu Zhangjin [Mon, 4 Jan 2010 09:16:48 +0000 (17:16 +0800)]
MIPS: Loongson: Cleanup of the environment variables

Changes:

o Move bus_clock into prom_init_env()
o Initialize the cpu_clock_freq to the default values for the
correspoding processor revisions if no such environment variable
passed by BIOS/Bootloader.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/826/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Move prom_argc and prom_argv into prom_init_cmdline()
Wu Zhangjin [Mon, 4 Jan 2010 09:16:47 +0000 (17:16 +0800)]
MIPS: Loongson: Move prom_argc and prom_argv into prom_init_cmdline()

prom_argc and prom_argv are only used by prom_init_cmdline(), move them
into the function.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/825/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Remove the serial port output of compressed kernel support
Wu Zhangjin [Mon, 4 Jan 2010 09:16:46 +0000 (17:16 +0800)]
MIPS: Loongson: Remove the serial port output of compressed kernel support

The compressed kernel support on loongson family machines is stable now,
so, remove the debug information via using SYS_SUPPORTS_ZBOOT instead of
SYS_SUPPORTS_ZBOOT_UART16550. This may reduce the image size and speedup
the booting.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/824/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Convert loongson_halt() to use unreachable()
Wu Zhangjin [Mon, 4 Jan 2010 09:16:45 +0000 (17:16 +0800)]
MIPS: Loongson: Convert loongson_halt() to use unreachable()

Use the new unreachable() macro instead of while(1);

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/823/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Lemote-2F: USB: Not Emulate Non-Posted Writes
Wu Zhangjin [Mon, 4 Jan 2010 09:16:44 +0000 (17:16 +0800)]
MIPS: Loongson: Lemote-2F: USB: Not Emulate Non-Posted Writes

Without this patch, when copying large amounts of data between the USB
storage devices and the hard disk, the USB device will disconnect
regularly.

Signed-off-by: Hu Hongbing <huhb@lemote.com>
Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Patchwork: http://patchwork.linux-mips.org/patch/822/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Loongson: Lemote-2F: Get the machine type from PMON_VER
Wu Zhangjin [Mon, 4 Jan 2010 09:16:43 +0000 (17:16 +0800)]
MIPS: Loongson: Lemote-2F: Get the machine type from PMON_VER

Lemote have used the PMON_VER strings to indicate the loongson-2f
machine series:

  PMON_VER=LM8089 Lemote 8.9'' netbook
           LM8101 Lemote 10.1'' netbook
  (The above two netbooks have the same kernel support)
         LM6XXX Lemote FuLoong(2F) box series
         LM9XXX Lemote LynLoong PC series

Before the machtype is supported by the PMON, we can get the machine
type from the PMON_VER for these machines, this will help the users a
lot.

Signed-off-by: Wu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: yanh@lemote.com
Cc: huhb@lemote.com
Cc: zhangfx@lemote.com
Cc: Wu Zhangjin <wuzhangjin@gmail.com>
Patchwork: http://patchwork.linux-mips.org/patch/821/
Patchwork: http://patchwork.linux-mips.org/patch/908/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Use constants from in.h
David Daney [Thu, 7 Jan 2010 19:05:06 +0000 (11:05 -0800)]
Staging: Octeon Ethernet: Use constants from in.h

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/837/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Enable scatter-gather.
David Daney [Thu, 7 Jan 2010 19:05:05 +0000 (11:05 -0800)]
Staging: Octeon Ethernet: Enable scatter-gather.

Octeon ethernet hardware can handle NETIF_F_SG, so we enable it.

A gather list of up to six fragments will fit in the SKB's CB
structure, so no extra memory is required.  If a SKB has more than six
fragments, we must linearize it.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/838/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Convert to NAPI.
David Daney [Thu, 7 Jan 2010 19:05:04 +0000 (11:05 -0800)]
Staging: Octeon Ethernet: Convert to NAPI.

Convert the driver to be a reasonably well behaved NAPI citizen.

There is one NAPI instance per CPU shared between all input ports.  As
receive backlog increases, NAPI is scheduled on additional CPUs.

Receive buffer refill code factored out so it can also be called from
the periodic timer.  This is needed to recover from temporary buffer
starvation conditions.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/839/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Rewrite transmit code.
David Daney [Thu, 7 Jan 2010 19:05:03 +0000 (11:05 -0800)]
Staging: Octeon Ethernet: Rewrite transmit code.

Stop the queue if too many packets are queued.  Restart it from a high
resolution timer.

Rearrange and simplify locking and SKB freeing code

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/843/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Fix memory allocation.
David Daney [Wed, 27 Jan 2010 21:22:53 +0000 (13:22 -0800)]
Staging: Octeon Ethernet: Fix memory allocation.

After aligning the blocks returned by kmalloc, we need to save the original
pointer so they can be correctly freed.

There are no guarantees about the alignment of SKB data, so we need to
handle worst case alignment.

Since right shifts over subtraction have no distributive property, we need
to fix the back pointer calculation.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/884/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoStaging: Octeon Ethernet: Remove unused code.
David Daney [Thu, 7 Jan 2010 19:05:01 +0000 (11:05 -0800)]
Staging: Octeon Ethernet: Remove unused code.

Remove unused code, reindent, and join some spilt strings.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/842/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Octeon: Fix EOI handling.
David Daney [Thu, 7 Jan 2010 19:05:00 +0000 (11:05 -0800)]
MIPS: Octeon: Fix EOI handling.

If an interrupt handler disables interrupts, the EOI function will
just reenable them.  This will put us in an endless loop when the
upcoming Ethernet driver patches are applied.

Only reenable the interrupt on EOI if it is not IRQ_DISABLED.  This
requires that the EOI function be separate from the ENABLE function.
We also rename the ACK functions to correspond with their function.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: gregkh@suse.de
Patchwork: http://patchwork.linux-mips.org/patch/840/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Octeon: Use optimized memory barrier primitives.
David Daney [Sat, 9 Jan 2010 01:17:44 +0000 (17:17 -0800)]
MIPS: Octeon: Use optimized memory barrier primitives.

In order to achieve correct synchronization semantics, the Octeon port
had defined CONFIG_WEAK_REORDERING_BEYOND_LLSC.  This resulted in code
that looks like:

   sync
   ll ...
   .
   .
   .
   sc ...
   .
   .
   sync

The second SYNC was redundant, but harmless.

Octeon has a SYNCW instruction that acts as a write-memory-barrier
(due to an erratum in some parts two SYNCW are used).  It is much
faster than SYNC because it imposes ordering on the writes, but
doesn't otherwise stall the execution pipeline.  On Octeon, SYNC
stalls execution until all preceeding writes are committed to the
coherent memory system.

Using:

    syncw;syncw
    ll
    .
    .
    .
    sc
    .
    .

Has identical semantics to the first sequence, but is much faster.
The SYNCW orders the writes, and the SC will not complete successfully
until the write is committed to the coherent memory system.  So at the
end all preceeding writes have been committed.  Since Octeon does not
do speculative reads, this functions as a full barrier.

The patch removes CONFIG_WEAK_REORDERING_BEYOND_LLSC, and substitutes
SYNCW for SYNC in write-memory-barriers.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/850/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: New macro smp_mb__before_llsc.
David Daney [Sat, 9 Jan 2010 01:17:43 +0000 (17:17 -0800)]
MIPS: New macro smp_mb__before_llsc.

Replace some instances of smp_llsc_mb() with a new macro
smp_mb__before_llsc().  It is used before ll/sc sequences that are
documented as needing write barrier semantics.

The default implementation of smp_mb__before_llsc() is just smp_llsc_mb(),
so there are no changes in semantics.

Also simplify definition of smp_mb(), smp_rmb(), and smp_wmb() to be just
barrier() in the non-SMP case.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/851/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Remove unused macros from barrier.h
David Daney [Thu, 7 Jan 2010 22:33:30 +0000 (14:33 -0800)]
MIPS: Remove unused macros from barrier.h

The smp_llsc_rmb() and smp_llsc_wmb() macros are not used in the tree,
remove them.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/848/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Octeon: Register some devices on the I2C bus.
David Daney [Thu, 7 Jan 2010 19:54:21 +0000 (11:54 -0800)]
MIPS: Octeon: Register some devices on the I2C bus.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: linux-i2c@vger.kernel.org
To: ben-linux@fluff.org
To: khali@linux-fr.org
Cc: Rade Bozic <rade.bozic.ext@nsn.com>
Patchwork: http://patchwork.linux-mips.org/patch/845/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: I2C: Add driver for Cavium OCTEON I2C ports.
Rade Bozic [Thu, 28 Jan 2010 20:47:07 +0000 (12:47 -0800)]
MIPS: I2C: Add driver for Cavium OCTEON I2C ports.

Signed-off-by: Rade Bozic <rade.bozic.ext@nsn.com>
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Cc: Michael Lawnick <michael.lawnick.ext@nsn.com>
To: linux-mips@linux-mips.org
To: linux-i2c@vger.kernel.org
To: ben-linux@fluff.org
To: khali@linux-fr.org
Cc: rade.bozic.ext@nsn.com
Cc: Michael Lawnick <michael.lawnick.ext@nsn.com>
Patchwork: http://patchwork.linux-mips.org/patch/890/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Octeon: Add I2C platform device.
David Daney [Thu, 7 Jan 2010 21:23:41 +0000 (13:23 -0800)]
MIPS: Octeon: Add I2C platform device.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
To: linux-i2c@vger.kernel.org
To: ben-linux@fluff.org
To: khali@linux-fr.org
Cc: Rade Bozic <rade.bozic.ext@nsn.com>
Patchwork: http://patchwork.linux-mips.org/patch/847/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Simplify param.h by using <asm-generic/param.h>
Robert P. J. Day [Thu, 31 Dec 2009 20:39:00 +0000 (15:39 -0500)]
MIPS: Simplify param.h by using <asm-generic/param.h>

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/810/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: get rid of common/reset.c
Manuel Lauss [Tue, 8 Dec 2009 18:18:13 +0000 (19:18 +0100)]
MIPS: Alchemy: get rid of common/reset.c

Implement reset / poweroff in the board code instead.  The peripheral reset
code is gone too since YAMON which all in-tree boards use does the same
work when it boots.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Yoichi Yuasa <yuasa@linux-mips.org>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/783/
Patchwork: http://patchwork.linux-mips.org/patch/882/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Two-level pagetables for 64-bit kernels with 64KB pages.
David Daney [Fri, 4 Dec 2009 21:52:36 +0000 (13:52 -0800)]
MIPS: Two-level pagetables for 64-bit kernels with 64KB pages.

For 64-bit kernels with 64KB pages and two level page tables, there are
42 bits worth of virtual address space This is larger than the 40 bits of
virtual address space obtained with the default 4KB Page size and three
levels, so there are no draw backs for using two level tables with this
configuration.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/761/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: irq: use runtime CPU type detection
Manuel Lauss [Mon, 23 Nov 2009 19:40:02 +0000 (20:40 +0100)]
MIPS: Alchemy: irq: use runtime CPU type detection

Use runtime CPU detection instead of relying on preprocessor symbols.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/701/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Only build AU1000 INTC code for compatible cpus
Manuel Lauss [Mon, 23 Nov 2009 19:40:01 +0000 (20:40 +0100)]
MIPS: Alchemy: Only build AU1000 INTC code for compatible cpus

Use the GPIO config symbol to only build Au1000 interrupt code on chips with
compatible hw.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/670/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: use runtime cpu detection in GPIO code.
Manuel Lauss [Mon, 23 Nov 2009 19:40:00 +0000 (20:40 +0100)]
MIPS: Alchemy: use runtime cpu detection in GPIO code.

Remove the cpu subtype cpp macros in favor of runtime detection,
to improve compile coverage of the alchemy common code.
(Increases kernel size by 700 bytes).

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/699/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoNET: au1000-eth: Convert to platform_driver model
Florian Fainelli [Tue, 10 Nov 2009 00:13:38 +0000 (01:13 +0100)]
NET: au1000-eth: Convert to platform_driver model

This patch converts the au1000-eth driver to become a full platform-driver
as it ought to be. We now pass PHY-speficic configurations through
platform_data but for compatibility the driver still assumes the default
settings (search for PHY1 on MAC0) when no platform_data is passed. Tested
on my MTX-1 board.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Cc: linux-mips@linux-mips.org
Cc: netdev@vger.kernel.org
Acked-by: David S. Miller <davem@davemloft.net>
Patchwork: http://patchwork.linux-mips.org/patch/619/
Patchwork: http://patchwork.linux-mips.org/patch/963/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Add au1000-eth platform device
Florian Fainelli [Tue, 10 Nov 2009 00:13:30 +0000 (01:13 +0100)]
MIPS: Alchemy: Add au1000-eth platform device

This patch makes the board code register the au1000-eth platform device. The
au1000-eth platform data can be overriden with the au1xxx_override_eth_cfg
function like it has to be done for the Bosporus board which uses a
different MAC/PHY setup.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Cc: David Miller <davem@davemloft.net>
Cc: linux-mips@linux-mips.org
Cc: netdev@vger.kernel.org
Patchwork: http://patchwork.linux-mips.org/patch/618/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: DB1200 defconfig update
Manuel Lauss [Mon, 2 Nov 2009 20:21:45 +0000 (21:21 +0100)]
MIPS: Alchemy: DB1200 defconfig update

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS/SOUND: Alchemy: DB1200 AC97+I2S audio support.
Manuel Lauss [Mon, 2 Nov 2009 20:21:44 +0000 (21:21 +0100)]
MIPS/SOUND: Alchemy: DB1200 AC97+I2S audio support.

Machine driver for DB1200 AC97 and I2S audio systems, intended as a proper
reference asoc machine for Alchemy-based systems.  AC97/I2S can be selected
at boot time by setting switch S6.7.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Cc: alsa-devel@alsa-project.org
Cc: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Extended DB1200 board support.
Manuel Lauss [Mon, 2 Nov 2009 20:21:43 +0000 (21:21 +0100)]
MIPS: Alchemy: Extended DB1200 board support.

Create own directory for DB1200 code and update it with new features.

- SPI support:
  - tmp121 temperature sensor
  - SPI flash on DB1200
- I2C support
  - NE1619 sensor
  - AT24 eeprom
- I2C/SPI can be selected at boot time via switch S6.8
- Carddetect IRQs for SD cards.
- gen_nand based NAND support.
- hexleds count sleep/wake transitions.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: physmap-flash for all devboards
Manuel Lauss [Mon, 19 Oct 2009 10:53:37 +0000 (12:53 +0200)]
MIPS: Alchemy: physmap-flash for all devboards

Replace the devboard NOR MTD mapping driver with physmap-flash support.
Also honor the "swapboot" switch settings wrt. to the layout of the
NOR partitions.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Acked-By: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Don't include <linux/smp_lock.h> unnecessarily.
Ralf Baechle [Sat, 27 Feb 2010 11:52:57 +0000 (12:52 +0100)]
MIPS: Don't include <linux/smp_lock.h> unnecessarily.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoSERIAL 8250: Fixes for Alchemy UARTs.
Manuel Lauss [Wed, 28 Oct 2009 20:37:28 +0000 (21:37 +0100)]
SERIAL 8250: Fixes for Alchemy UARTs.

Limit the amount of address space claimed for Alchemy serial ports to
0x1000.  On the Au1300, ports are only 0x1000 apart, and the registers
only extend to 0x110 at most on all supported alchemy models.

On the Au1300 the autodetect logic no longer works and this makes it
necessary to specify the port type through platform data.  Because of
this the MSR quirk needs to be moved outside the autoconfig() function
which will no longer be called when UPF_FIXED_TYPE is specified.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Cc: linux-serial@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-mips@linux-mips.org
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: UARTs are of type 16550A
Manuel Lauss [Wed, 28 Oct 2009 20:49:46 +0000 (21:49 +0100)]
MIPS: Alchemy: UARTs are of type 16550A

UART autodetection breaks on the Au1300 but the IP blocks are identical,
at least according to the datasheets.  Help the 8250 driver by passing
on uart type information via platform data.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Turn on -Werror for devboards and xss1500
Florian Fainelli [Sun, 18 Oct 2009 14:04:41 +0000 (16:04 +0200)]
MIPS: Alchemy: Turn on -Werror for devboards and xss1500

Warnings being suppressed, we can now turn on -Werror for boards which did
not have it already (devboards and xss1500).

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Fix warnings in DB1x00 / PB1000 / PB1550 board setup code
Florian Fainelli [Sun, 18 Oct 2009 14:04:09 +0000 (16:04 +0200)]
MIPS: Alchemy: Fix warnings in DB1x00 / PB1000 / PB1550 board setup code

This patch fixes warnings due to potentially unused
variables in board setup code or mixed variables
declaration and code (forbidden by ISO C90).

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: remove unused SYS area structure
Manuel Lauss [Thu, 15 Oct 2009 17:32:01 +0000 (19:32 +0200)]
MIPS: Alchemy: remove unused SYS area structure

Nothing in-tree uses it, so get rid of it.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: get rid of superfluous UART definitions
Manuel Lauss [Thu, 15 Oct 2009 17:07:34 +0000 (19:07 +0200)]
MIPS: Alchemy: get rid of superfluous UART definitions

Remove unused uart bit definitions and base macros.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: prom_putchar is board dependent
Manuel Lauss [Thu, 15 Oct 2009 16:49:27 +0000 (18:49 +0200)]
MIPS: Alchemy: prom_putchar is board dependent

This patch replaces the general alchemy prom_putchar() implementation
in favor of board-specific versions:  The UART where the output of
prom_putchar is directed to really depends on the board, the current
implementation hardcodes this on a per-SoC basis which is just wrong.

So a generic uart tx function is provided in the alchemy headers,
and the boards can provide their own prom_putchar with custom
destination uart, and all in-kernel alchemy boards support
early printk.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: change dbdma to accept physical memory addresses
Manuel Lauss [Tue, 13 Oct 2009 18:22:35 +0000 (20:22 +0200)]
MIPS: Alchemy: change dbdma to accept physical memory addresses

DMA can only be done from physical addresses; move the "virt_to_phys"
source/destination buffer address translation from the dbdma queueing
functions (since the hardware can only DMA to/from physical addresses)
to their respective users.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: remove dbdma compat macros
Manuel Lauss [Tue, 13 Oct 2009 18:22:34 +0000 (20:22 +0200)]
MIPS: Alchemy: remove dbdma compat macros

Remove dbdma compat macros, move remaining users over to default
queueing functions and -flags.

(Queueing function signature has changed in order to give
 a build failure instead of silent functional changes due
 to the no longer implicitly specified DDMA_FLAGS_IE flag)

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: reduce size of irq dispatcher
Manuel Lauss [Tue, 13 Oct 2009 18:26:31 +0000 (20:26 +0200)]
MIPS: Alchemy: reduce size of irq dispatcher

By replacing an extra do_IRQ with a goto, the assembly shrinks
from 260 to 212 bytes (gcc-4.3.4).

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Stop IRQ name sharing
Manuel Lauss [Wed, 7 Oct 2009 18:15:15 +0000 (20:15 +0200)]
MIPS: Alchemy: Stop IRQ name sharing

Eliminate the sharing of IRQ names among the differenct Alchemy
variants.  IRQ numbers need no longer be hidden behind a
CONFIG_SOC_AU1XXX symbol: step 1 in my quest to make the Alchemy
code less reliant on a hardcoded subtype.

This patch also renames the GPIO irq number constants. It's really
an interrupt line, NOT a GPIO number!

Code which relied on certain irq numbers to have the same name
across all supported cpu subtypes is changed to determine current
cpu subtype at runtime; in some places this isn't possible so
a "compat" symbol is used.

Run-tested on DB1200.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Simple cpu subtype detector
Manuel Lauss [Wed, 7 Oct 2009 18:15:14 +0000 (20:15 +0200)]
MIPS: Alchemy: Simple cpu subtype detector

Extract the alchemy chip variant from c0_prid register.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/14/
Patchwork: http://patchwork.linux-mips.org/patch/707/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: higher priority for system timer.
Manuel Lauss [Wed, 7 Oct 2009 18:15:13 +0000 (20:15 +0200)]
MIPS: Alchemy: higher priority for system timer.

Raise RTCMATCH2 interrupt priority in case it is used as the system
timer tick.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: Remove USB_DEV_REQ_INT prioritization hack
Manuel Lauss [Wed, 7 Oct 2009 18:15:12 +0000 (20:15 +0200)]
MIPS: Alchemy: Remove USB_DEV_REQ_INT prioritization hack

The Alchemy hardware provides a method to prioritize interrupts
on a controller by assigning them to a differenct core request line.
Assign usb device request interrupt to IC0 Request 0 (which has
highest priority in the core and the dispatcher) and others to
Request 1.  The explicit check for usb device request occurrence
should be obsolete now.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: XXS1500 PCMCIA driver rewrite
Manuel Lauss [Sun, 4 Oct 2009 12:55:29 +0000 (14:55 +0200)]
MIPS: Alchemy: XXS1500 PCMCIA driver rewrite

Rewritten XXS1500 PCMCIA socket driver, standalone (doesn't depend on
au1000_generic.c) and added carddetect IRQ support.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Florian Fainelli <florian@openwrt.org>
Cc: Linux-PCMCIA <linux-pcmcia@lists.infradead.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: devboards: wire up new PCMCIA driver.
Manuel Lauss [Sun, 4 Oct 2009 12:55:28 +0000 (14:55 +0200)]
MIPS: Alchemy: devboards: wire up new PCMCIA driver.

Register the PCMCIA driver on all boards supported by it,
get rid of now-unused pcmcia macros in the board headers
(and subsequently empty pb1100/pb1500 ones).

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: PCMCIA: new socket driver for Au1000 demoboards.
Manuel Lauss [Sun, 4 Oct 2009 12:55:27 +0000 (14:55 +0200)]
MIPS: PCMCIA: new socket driver for Au1000 demoboards.

New PCMCIA socket driver for all Db/Pb1xxx boards (except Pb1000),
which replaces au1000_db1x00.c and (most of) au1000_pb1x00.c.
Notable improvements:
        - supports Db1000, DB/PB1100/1500/1550/1200.
        - support for carddetect and statuschange IRQs.
        - pcmcia socket mem/io/attr areas and irqs passed through
          platform resource information.
        - doesn't freeze system during card insertion/ejection like
          the one it replaces.
        - boardtype is automatically detected using BCSR ID register.

Run-tested on the DB1200.

Cc: Linux-PCMCIA <linux-pcmcia@lists.infradead.org>
Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: remove board_init_irq() function.
Manuel Lauss [Sun, 4 Oct 2009 12:55:26 +0000 (14:55 +0200)]
MIPS: Alchemy: remove board_init_irq() function.

remove board_init_irq():  On all in-kernel boards it is sufficient to
initialize board interrupts in an arch_initcall by using the default
linux irq functions.

Some small irqmap.c files have been folded into board_setup files.

Run-tested on DB1200; compile-tested on all other affected boards.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: devboards: factor out PB1200 IRQ cascade code.
Manuel Lauss [Sun, 4 Oct 2009 12:55:25 +0000 (14:55 +0200)]
MIPS: Alchemy: devboards: factor out PB1200 IRQ cascade code.

Move the PB1200 IRQ cascade code out to the BCSR support code:
upcoming DB1300 support can use it too.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Alchemy: devboard register abstraction
Manuel Lauss [Sun, 4 Oct 2009 12:55:24 +0000 (14:55 +0200)]
MIPS: Alchemy: devboard register abstraction

All Alchemy development boards have external CPLDs with a few registers
in them.  They all share an identical register layout with only a few
minor differences (except the PB1000) in bit functions and base
addresses.

This patch
- adds a primitive facility to initialize and use these external
  registers,
- replaces all occurrences of bcsr->xxx accesses with calls to the new
  functions (the pb1200 cascade irq handling code is special).
- collects BCSR register information scattered throughout the board
  headers in a central place.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMIPS: Fix build error for uncompressed non-plain vmlinux kernels
Ralf Baechle [Fri, 26 Feb 2010 12:57:12 +0000 (13:57 +0100)]
MIPS: Fix build error for uncompressed non-plain vmlinux kernels

Seen on rm200_defconfig for example:

  CC      arch/mips/boot/compressed/decompress.o
/home/ralf/src/linux/upstream-linus/arch/mips/boot/compressed/decompress.c: In function ‘decompress_kernel’:
/home/ralf/src/linux/upstream-linus/arch/mips/boot/compressed/decompress.c:116: error: implicit declaration of function ‘decompress’
make[3]: *** [arch/mips/boot/compressed/decompress.o] Error 1
make[2]: *** [vmlinuz.ecoff] Error 2
make[1]: *** [sub-make] Error 2
make: *** [all] Error 2

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 27 Feb 2010 01:22:53 +0000 (17:22 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/rafael/suspend-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/suspend-2.6:
  PM / Hibernate: Fix preallocating of memory
  PM / Hibernate: Remove swsusp.c finally
  PM / Hibernate: Remove trailing space in message
  PM: Allow SCSI devices to suspend/resume asynchronously
  PM: Allow USB devices to suspend/resume asynchronously
  USB: implement non-tree resume ordering constraints for PCI host controllers
  PM: Allow PCI devices to suspend/resume asynchronously
  PM / Hibernate: Swap, remove useless check from swsusp_read()
  PM / Hibernate: Really deprecate deprecated user ioctls
  PM: Allow device drivers to use dpm_wait()
  PM: Start asynchronous resume threads upfront
  PM: Add facility for advanced testing of async suspend/resume
  PM: Add a switch for disabling/enabling asynchronous suspend/resume
  PM: Asynchronous suspend and resume of devices
  PM: Add parent information to timing messages
  PM: Document device power attributes in sysfs
  PM / Runtime: Add sysfs switch for disabling device run-time PM

14 years agoRemove EXPERIMENTAL from NFS_FSCACHE
Christian Kujau [Fri, 26 Feb 2010 17:25:14 +0000 (17:25 +0000)]
Remove EXPERIMENTAL from NFS_FSCACHE

There's currently an open Ubuntu bug[0], with the intent to compile NFS_FSCACHE
(and possibly AFS_FSCACHE, 9P_FSCACHE) into the standard Ubuntu kernel.
However, since *_FSCACHE still depends on EXPERIMENTAL, this won't happen.

As Arjan van de Ven pointed out[1], the EXPERIMENTAL flag doesn't mean that
much any more, I propose the following patch to fs/nfs/Kconfig.  I'd do the
same for fs/9p/Kconfig and fs/afs/Kconfig, but as I did not test 9p or AFS, I
feel it would not be appropriate for me to remove the flag.

[0] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/440522/comments/5
[1] http://lkml.org/lkml/2010/1/23/145

Signed-off-by: Christian Kujau <lists@nerdbynature.de>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platf...
Linus Torvalds [Sat, 27 Feb 2010 01:20:18 +0000 (17:20 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/mjg59/platform-drivers-x86

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86:
  toshiba_acpi: Add full hotkey support
  hp-wmi: Add support for tablet rotation key
  dell-laptop: Add another Dell laptop to the DMI whitelist
  classmate-laptop: use a single MODULE_DEVICE_TABLE to get correct aliases
  dell-laptop: Pay attention to which devices the hardware switch controls
  dell-laptop: Use buffer with 32-bit physical address
  dell-laptop: Blacklist machines not supporting dell-laptop
  dell-laptop: Block software state changes when rfkill hard blocked
  dell-laptop: Fix small memory leak
  dell-laptop: Fix platform device unregistration
  dell-laptop: Update rfkill state on kill switch
  compal-laptop: Replace sysfs support with rfkill support
  compal-laptop: Add support for known Compal made Dell laptops
  MAINTAINERS: update drivers/platform/x86 information

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm
Linus Torvalds [Sat, 27 Feb 2010 01:19:30 +0000 (17:19 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: use bastmode in debugfs output
  dlm: Send lockspace name with uevents
  dlm: send reply before bast
  dlm: fix ordering of bast and cast

14 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Sat, 27 Feb 2010 01:18:52 +0000 (17:18 -0800)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs: (52 commits)
  fs/xfs: Correct NULL test
  xfs: optimize log flushing in xfs_fsync
  xfs: only clear the suid bit once in xfs_write
  xfs: kill xfs_bawrite
  xfs: log changed inodes instead of writing them synchronously
  xfs: remove invalid barrier optimization from xfs_fsync
  xfs: kill the unused XFS_QMOPT_* flush flags V2
  xfs: Use delay write promotion for dquot flushing
  xfs: Sort delayed write buffers before dispatch
  xfs: Don't issue buffer IO direct from AIL push V2
  xfs: Use delayed write for inodes rather than async V2
  xfs: Make inode reclaim states explicit
  xfs: more reserved blocks fixups
  xfs: turn off sign warnings
  xfs: don't hold onto reserved blocks on remount,ro
  xfs: quota limit statvfs available blocks
  xfs: replace KM_LARGE with explicit vmalloc use
  xfs: cleanup up xfs_log_force calling conventions
  xfs: kill XLOG_VEC_SET_TYPE
  xfs: remove duplicate buffer flags
  ...

14 years agoMerge branch 'ibft-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad...
Linus Torvalds [Sat, 27 Feb 2010 01:18:11 +0000 (17:18 -0800)]
Merge branch 'ibft-fixes' of git://git./linux/kernel/git/konrad/ibft-2.6

* 'ibft-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/ibft-2.6:
  ibft: Update MAINTAINERS file.
  drivers/firmware/iscsi_ibft.c: remove NIPQUAD_FMT, use %pI4

14 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Sat, 27 Feb 2010 01:16:20 +0000 (17:16 -0800)]
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-2.6

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (362 commits)
  V4L-DVB: cx88-dvb: remove extra attribution for core
  V4L/DVB: v4l: soc_camera: fix bound checking of mbus_fmt[] index
  V4L/DVB: Add support for SMT7020 to cx88
  V4L/DVB: radio-si470x: Use UTF-8 encoding on a comment
  V4L/DVB: MAINTAINERS: Telegent tlg2300 section fix
  V4L/DVB: gspca_stv06xx: Add support for camera button
  V4L/DVB: gspca_ov519: add support for the button on ov511 based cams
  V4L/DVB: gspca_ov519: Add support for the button on ov518 based cams
  V4L/DVB: gspca_ov519: add support for the button on ov519 based cams
  V4L/DVB: gspca_main: Fix a compile error when CONFIG_INPUT is not set
  V4L/DVB: gspca_main: some input error handling fixes
  V4L/DVB: gspca_main: Allow use of input device creation code for non int. inputs
  V4L/DVB: gspca_pac7302: much improved exposure control
  V4L/DVB: gspca_sonixb: Make sonixb driver handle pas106 and pas202 cameras
  V4L/DVB: gspca_sonixb: pas106: fixup bright ctrl and add gain and exposure ctrls
  V4L/DVB: Documentation: gspca.txt: update known mr97310a cams
  V4L/DVB: gspca_mr97310a: add support for the Sakar 1638x CyberPix
  V4L/DVB: gscpa_sonixb: limit ov7630 max framerate at 640x480
  V4L/DVB: gspca_sonixb: pas202: fixup brightness ctrl and add gain and exposure ctrls
  V4L/DVB: gscpa_sonixb: Differentiate between sensors with a coarse and fine expo ctrl
  ...

14 years agodvb-core: Fix DoS bug in ULE decapsulation code that can be triggered by an invalid...
Ang Way Chuang [Thu, 25 Feb 2010 01:45:03 +0000 (09:45 +0800)]
dvb-core: Fix DoS bug in ULE decapsulation code that can be triggered by an invalid Payload Pointer

ULE (Unidirectional Lightweight Encapsulation RFC 4326) decapsulation
has a bug that causes endless loop when Payload Pointer of MPEG2-TS
frame is 182 or 183.  Anyone who sends malicious MPEG2-TS frame will
cause the receiver of ULE SNDU to go into endless loop.

This patch was generated and tested against linux-2.6.32.9 and should
apply cleanly to linux-2.6.33 as well because there was only one typo
fix to dvb_net.c since v2.6.32.

This bug was brought to you by modern day Santa Claus who decided to
shower the satellite dish at Keio University with heavy snow causing
huge burst of errors.  We, receiver end, received Santa Claus's gift in
the form of kernel bug.

Care has been taken not to introduce more bug by fixing this bug, but
please scrutinize the code for I always produces buggy code.

Signed-off-by: Ang Way Chuang <wcang79@gmail.com>
Acked-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'kmemcheck-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 27 Feb 2010 01:11:11 +0000 (17:11 -0800)]
Merge branch 'kmemcheck-for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'kmemcheck-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  kmemcheck: Test the full object in kmemcheck_is_obj_initialized()

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/xfs-vipt
Linus Torvalds [Sat, 27 Feb 2010 01:05:10 +0000 (17:05 -0800)]
Merge git://git./linux/kernel/git/jejb/xfs-vipt

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/xfs-vipt:
  xfs: fix xfs to work with Virtually Indexed architectures
  sh: add mm API for DMA to vmalloc/vmap areas
  arm: add mm API for DMA to vmalloc/vmap areas
  parisc: add mm API for DMA to vmalloc/vmap areas
  mm: add coherence API for DMA to vmalloc/vmap areas

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
Linus Torvalds [Sat, 27 Feb 2010 00:55:27 +0000 (16:55 -0800)]
Merge git://git./linux/kernel/git/jejb/scsi-misc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (158 commits)
  [SCSI] Fix printing of failed 32-byte commands
  [SCSI] Fix printing of variable length commands
  [SCSI] libsrp: fix bug in ADDITIONAL CDB LENGTH interpretation
  [SCSI] scsi_dh_alua: Add IBM Power Virtual SCSI ALUA device to dev list
  [SCSI] scsi_dh_alua: add netapp to dev list
  [SCSI] qla2xxx: Update version number to 8.03.02-k1.
  [SCSI] qla2xxx: EEH: Restore PCI saved state during pci slot reset.
  [SCSI] qla2xxx: Add firmware ETS burst support.
  [SCSI] qla2xxx: Correct loop-resync issues during SNS scans.
  [SCSI] qla2xxx: Correct use-after-free issue in terminate_rport_io callback.
  [SCSI] qla2xxx: Correct EH bus-reset handling.
  [SCSI] qla2xxx: Proper clean-up of BSG requests when request times out.
  [SCSI] qla2xxx: Initialize payload receive length in failure path of vendor commands
  [SCSI] fix duplicate removal on error path in scsi_sysfs_add_sdev
  [SCSI] fix refcounting bug in scsi_get_host_dev
  [SCSI] fix memory leak in scsi_report_lun_scan
  [SCSI] lpfc: correct PPC build failure
  [SCSI] raid_class: add raid1e
  [SCSI] mpt2sas: Do not call sas_is_tlr_enabled for RAID volumes.
  [SCSI] zfcp: Introduce header file for qdio structs and inline functions
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6
Linus Torvalds [Sat, 27 Feb 2010 00:54:27 +0000 (16:54 -0800)]
Merge git://git./linux/kernel/git/lethal/sh-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6: (187 commits)
  sh: remove dead LED code for migo-r and ms7724se
  sh: ecovec build fix for CONFIG_I2C=n
  sh: ecovec r-standby support
  sh: ms7724se r-standby support
  sh: SH-Mobile R-standby register save/restore
  clocksource: Fix up a registration/IRQ race in the sh drivers.
  sh: ms7724: modify scan_timing for KEYSC
  sh: ms7724: Add sh_sir support
  sh: mach-ecovec24: Add sh_sir support
  sh: wire up SET/GET_UNALIGN_CTL.
  sh: allow alignment fault mode to be configured at kernel boot.
  sh: sh7724: Update FSI/SPU2 clock
  sh: always enable sh7724 vpu_clk and set to 166MHz on Ecovec
  sh: add sh7724 kick callback to clk_div4_table
  sh: introduce struct clk_div4_table
  sh: clock-cpg div4 set_rate() shift fix
  sh: Turn on speculative return for SH7785 and SH7786
  sh: Merge legacy and dynamic PMB modes.
  sh: Use uncached I/O helpers in PMB setup.
  sh: Provide uncached I/O helpers.
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Sat, 27 Feb 2010 00:50:02 +0000 (16:50 -0800)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (31 commits)
  crypto: aes_generic - Fix checkpatch errors
  crypto: fcrypt - Fix checkpatch errors
  crypto: ecb - Fix checkpatch errors
  crypto: des_generic - Fix checkpatch errors
  crypto: deflate - Fix checkpatch errors
  crypto: crypto_null - Fix checkpatch errors
  crypto: cipher - Fix checkpatch errors
  crypto: crc32 - Fix checkpatch errors
  crypto: compress - Fix checkpatch errors
  crypto: cast6 - Fix checkpatch errors
  crypto: cast5 - Fix checkpatch errors
  crypto: camellia - Fix checkpatch errors
  crypto: authenc - Fix checkpatch errors
  crypto: api - Fix checkpatch errors
  crypto: anubis - Fix checkpatch errors
  crypto: algapi - Fix checkpatch errors
  crypto: blowfish - Fix checkpatch errors
  crypto: aead - Fix checkpatch errors
  crypto: ablkcipher - Fix checkpatch errors
  crypto: pcrypt - call the complete function on error
  ...

14 years agoibft: Update MAINTAINERS file.
Konrad Rzeszutek Wilk [Fri, 26 Feb 2010 03:53:00 +0000 (03:53 +0000)]
ibft: Update MAINTAINERS file.

Provide the right e-mail and names for me and Peter.

Signed-off-by: Konrad Rzeszutek Wilk <konrad@kernel.org>
14 years agodrivers/firmware/iscsi_ibft.c: remove NIPQUAD_FMT, use %pI4
Joe Perches [Thu, 14 Jan 2010 16:34:19 +0000 (11:34 -0500)]
drivers/firmware/iscsi_ibft.c: remove NIPQUAD_FMT, use %pI4

Convert netmask to __be32 and format it with %pI4

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Konrad Rzeszutek <ketuzsezr@darnok.org>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Peter Jones <pjones@redhat.com>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
14 years agoMerge branch 'linux-2.6.33'
Alex Elder [Fri, 26 Feb 2010 20:34:02 +0000 (14:34 -0600)]
Merge branch 'linux-2.6.33'

14 years agoPM / Hibernate: Fix preallocating of memory
Rafael J. Wysocki [Thu, 25 Feb 2010 21:32:37 +0000 (22:32 +0100)]
PM / Hibernate: Fix preallocating of memory

The hibernate memory preallocation code allocates memory to push some
user space data out of physical RAM, so that the hibernation image is
not too large.  It allocates more memory than necessary for creating
the image, so it has to release some pages to make room for
allocations made while suspending devices and disabling nonboot CPUs,
or the system will hang due to the lack of free pages to allocate
from.  Unfortunately, the function used for freeing these pages,
free_unnecessary_pages(), contains a bug that prevents it from doing
the job on all systems without highmem.

Fix this problem, which is a regression from the 2.6.30 kernel, by
using the right condition for the termination of the loop in
free_unnecessary_pages().

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Reported-and-tested-by: Alan Jenkins <sourcejedi.lkml@googlemail.com>
Cc: stable@kernel.org
14 years agoPM / Hibernate: Remove swsusp.c finally
Jiri Slaby [Sun, 21 Feb 2010 21:14:44 +0000 (22:14 +0100)]
PM / Hibernate: Remove swsusp.c finally

Its contents and entry in Makefile were already removed in
8e60c6a1348e17e68ad73589a52a03876e7059be
(Shift remaining code from swsusp.c to hibernate.c)
but somehow it remained in-place (rjw: which most likely was my
mistake).

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Nigel Cunningham <nigel@tuxonice.net>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM / Hibernate: Remove trailing space in message
Frans Pop [Thu, 11 Feb 2010 22:09:08 +0000 (23:09 +0100)]
PM / Hibernate: Remove trailing space in message

Remove a trailing space from a message in swsusp_save().

Signed-off-by: Frans Pop <elendil@planet.nl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Allow SCSI devices to suspend/resume asynchronously
Rafael J. Wysocki [Mon, 8 Feb 2010 18:18:26 +0000 (19:18 +0100)]
PM: Allow SCSI devices to suspend/resume asynchronously

Set power.async_suspend for all SCSI devices, targets and hosts, so
that they can be suspended and resumed in parallel with the main
suspend/resume thread and possibly with other devices they don't
depend on in a known way (i.e. devices which are not their parents or
children).

The power.async_suspend flag is also set for devices that don't have
suspend or resume callbacks, because otherwise they would make the
main suspend/resume thread wait for their "asynchronous" children
(during suspend) or parents (during resume), effectively negating the
possible gains from executing these devices' suspend and resume
callbacks asynchronously.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Allow USB devices to suspend/resume asynchronously
Rafael J. Wysocki [Mon, 8 Feb 2010 18:18:16 +0000 (19:18 +0100)]
PM: Allow USB devices to suspend/resume asynchronously

Set power.async_suspend for USB devices, endpoints and interfaces,
allowing them to be suspended and resumed asynchronously during
system sleep transitions.

The power.async_suspend flag is also set for devices that don't have
suspend or resume callbacks, because otherwise they would make the
main suspend/resume thread wait for their "asynchronous" children
(during suspend) or parents (during resume), effectively negating the
possible gains from executing these devices' suspend and resume
callbacks asynchronously.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoUSB: implement non-tree resume ordering constraints for PCI host controllers
Alan Stern [Fri, 12 Feb 2010 11:21:11 +0000 (12:21 +0100)]
USB: implement non-tree resume ordering constraints for PCI host controllers

This patch (as1331) adds non-tree ordering constraints needed for
proper resume of PCI USB host controllers from hibernation.  The main
issue is that non-high-speed devices must not be resumed before the
high-speed root hub, because it is the ehci_bus_resume() routine which
takes care of handing the device connection over to the companion
controller.  If the device resume is attempted before the handover
then the device won't be found and it will be treated as though it had
disconnected.

The patch adds a new field to the usb_bus structure; for each
full/low-speed bus this field will contain a pointer to the companion
high-speed bus (if one exists).  It is used during normal device
resume; if the hs_companion pointer isn't NULL then we wait for the
root-hub device on the hs_companion bus.

A secondary issue is that an EHCI controlller shouldn't be resumed
before any of its companions.  On some machines I have observed
handovers failing if the companion controller is reinitialized after
the handover.  Thus, the EHCI resume routine must wait for the
companion controllers to be resumed.

The patch also fixes a small bug in usb_hcd_pci_probe(); an error path
jumps to the wrong label, causing a memory leak.

[rjw: Fixed compilation for CONFIG_PM_SLEEP unset.]

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Allow PCI devices to suspend/resume asynchronously
Rafael J. Wysocki [Mon, 8 Feb 2010 18:16:33 +0000 (19:16 +0100)]
PM: Allow PCI devices to suspend/resume asynchronously

Set power.async_suspend for all PCI devices and PCIe port services,
so that they can be suspended and resumed in parallel with other
devices they don't depend on in a known way (i.e. devices which are
not their parents or children).

This only affects the "regular" suspend and resume stages, which
means in particular that the restoration of the PCI devices' standard
configuration registers during resume will still be carried out
synchronously (at the "early" resume stage).

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM / Hibernate: Swap, remove useless check from swsusp_read()
Jiri Slaby [Wed, 27 Jan 2010 22:47:56 +0000 (23:47 +0100)]
PM / Hibernate: Swap, remove useless check from swsusp_read()

It will never reach here if the sws_resume_bdev is erratic.
swsusp_read() is called only from software_resume(), but after
swsusp_check() which would catch the error state.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM / Hibernate: Really deprecate deprecated user ioctls
Jiri Slaby [Wed, 27 Jan 2010 22:47:50 +0000 (23:47 +0100)]
PM / Hibernate: Really deprecate deprecated user ioctls

They were deprecated and removed from exported headers more than 2
years ago. Inform users about their removal in the future now.

(Switch cases needed to be reorderded for an easy fall through.)

And add an entry to feature-removal-schedule.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Allow device drivers to use dpm_wait()
Rafael J. Wysocki [Wed, 27 Jan 2010 22:47:38 +0000 (23:47 +0100)]
PM: Allow device drivers to use dpm_wait()

There are some dependencies between devices (in particular, between
EHCI USB controllers and their OHCI/UHCI siblings) which are not
reflected by the structure of the device tree.  With synchronous
suspend and resume these dependencies are taken into accout
automatically, because the devices in question are always registered
in the right order, but to meet these constraints with asynchronous
suspend and resume the drivers of these devices will need to use
dpm_wait() in their suspend/resume routines, so introduce a helper
function allowing them to do that.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Start asynchronous resume threads upfront
Rafael J. Wysocki [Sat, 23 Jan 2010 21:25:31 +0000 (22:25 +0100)]
PM: Start asynchronous resume threads upfront

It has been shown by testing that total device resume time can be
reduced significantly (by as much as 50% or more) if the async
threads executing some devices' resume routines are all started
before the main resume thread starts to handle the "synchronous"
devices.

This is a consequence of the fact that the slowest devices tend to be
located at the end of dpm_list, so their resume routines are started
very late.  Consequently, they have to wait for all the preceding
"synchronous" devices before their resume routines can be started
by the main resume thread, even if they are "asynchronous".  By
starting their async threads upfront we effectively move those
devices towards the beginning of dpm_list, without breaking their
ordering with respect to their parents and children.  As a result,
their resume routines are started much earlier and we are able to
save much more device resume time this way.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
14 years agoPM: Add facility for advanced testing of async suspend/resume
Rafael J. Wysocki [Sat, 23 Jan 2010 21:25:23 +0000 (22:25 +0100)]
PM: Add facility for advanced testing of async suspend/resume

Add configuration switch CONFIG_PM_ADVANCED_DEBUG for compiling in
extra PM debugging/testing code allowing one to access some
PM-related attributes of devices from the user space via sysfs.

If CONFIG_PM_ADVANCED_DEBUG is set, add sysfs attribute power/async
for every device allowing the user space to access the device's
power.async_suspend flag and modify it, if desired.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>