14 years agopowerpc: Remove common stuff from setup_64.c
Paul Mackerras [Thu, 27 Oct 2005 01:08:31 +0000]
powerpc: Remove common stuff from setup_64.c

This should have been in commit 03501dab035ab7da5e1373f5e130cfd6346d3f21
but got missed by accident.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: 32-bit powermac needs the mpc106 code
Paul Mackerras [Wed, 26 Oct 2005 12:54:17 +0000]
powerpc: 32-bit powermac needs the mpc106 code

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fix incorrect timer register addresses in mpic.c
Paul Mackerras [Wed, 26 Oct 2005 11:55:33 +0000]
powerpc: Fix incorrect timer register addresses in mpic.c

We were computing the wrong address for the MPIC timer registers,
so when we went to initialize them we would have been hitting some
unrelated ioremap... oops.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc32: Limit memory to lowmem if !CONFIG_HIGHMEM.
Paul Mackerras [Wed, 26 Oct 2005 11:54:21 +0000]
powerpc32: Limit memory to lowmem if !CONFIG_HIGHMEM.

This trims off the extra unusable memory from the lmb structure,
so we don't try to use it.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopower: Update the multiple inclusion protection symbol on machdep.h
Paul Mackerras [Wed, 26 Oct 2005 11:53:19 +0000]
power: Update the multiple inclusion protection symbol on machdep.h

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fixes to get the Longtrail CHRP a bit further
Paul Mackerras [Wed, 26 Oct 2005 11:52:53 +0000]
powerpc: Fixes to get the Longtrail CHRP a bit further

Talk about buggy firmware...  the OF on the Longtrail returns 0
from the claim client service rather than -1 when the claim fails.
It also has no device_type on the /memory node and blows up if
the output buffer for package-to-path is too big.

This also fixes a bug with calling alloc_up with align == 0, where
we did _ALIGN_UP(alloc_bottom, 0) which will end up as 0.

Lastly, we now check the return value (in r3) from calling the
prom, and return -1 from call_prom if we get a negative value back.
That is supposed to indicate that the requested client service
doesn't exist.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Merge 32-bit CHRP support.
Paul Mackerras [Wed, 26 Oct 2005 11:45:56 +0000]
powerpc: Merge 32-bit CHRP support.

SMP still needs more work but UP gets as far as starting userspace
at least.  This uses the 64-bit-style code for spinning up the cpus.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Don't limit pmac_get_rtc_time to return only positive values
Paul Mackerras [Wed, 26 Oct 2005 07:22:33 +0000]
powerpc: Don't limit pmac_get_rtc_time to return only positive values

If the machine's clock is set to a bogus value, this check resulted
in userland waiting effectively forever for the RTC value to change,
so remove the check.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fix interrupt-tree parsing
Paul Mackerras [Wed, 26 Oct 2005 07:19:06 +0000]
powerpc: Fix interrupt-tree parsing

The interrupt-tree parsing code wasn't offsetting interrupt numbers
by 16 on 32-bit platforms with an i8259 interrupt controller, and
it was confused about the encoding of interrupt sense and level
(which is different for i8259 and openpic interrupt controllers,
just to make things interesting).

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Pull out MPC106 (grackle) initialization code into its own file
Paul Mackerras [Wed, 26 Oct 2005 07:16:38 +0000]
powerpc: Pull out MPC106 (grackle) initialization code into its own file

This is so that the 32-bit CHRP code can use it.  The MPC106
initialization code is now in arch/powerpc/sysdev/grackle.c and
is controlled by CONFIG_PPC_MPC106.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Pull common bits of setup_{32,64}.c into setup-common.c
Paul Mackerras [Wed, 26 Oct 2005 07:11:18 +0000]
powerpc: Pull common bits of setup_{32,64}.c into setup-common.c

This creates a new arch/powerpc/kernel/setup-common.c with various
bits that setup_32.c and setup_64.c had in common - functions like
machine_shutdown/restart/power_off, show_cpuinfo, set_preferred_console
etc.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Merge rtas.c into arch/powerpc/kernel
Paul Mackerras [Wed, 26 Oct 2005 07:05:24 +0000]
powerpc: Merge rtas.c into arch/powerpc/kernel

This splits arch/ppc64/kernel/rtas.c into arch/powerpc/kernel/rtas.c,
which contains generic RTAS functions useful on any CHRP platform,
and arch/powerpc/platforms/pseries/rtas-fw.[ch], which contain
some pSeries-specific firmware flashing bits.  The parts of rtas.c
that are to do with pSeries-specific error logging are protected
by a new CONFIG_RTAS_ERROR_LOGGING symbol.  The inclusion of rtas.o
is controlled by the CONFIG_PPC_RTAS symbol, and the relevant
platforms select that.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Merge i8259.c into arch/powerpc/sysdev
Paul Mackerras [Wed, 26 Oct 2005 06:47:42 +0000]
powerpc: Merge i8259.c into arch/powerpc/sysdev

This changes the parameters for i8259_init so that it takes two
parameters: a physical address for generating an interrupt
acknowledge cycle, and an interrupt number offset.  i8259_init
now sets the irq_desc[] for its interrupts; all the callers
were doing this, and that code is gone now.  This also defines
a CONFIG_PPC_I8259 symbol to select i8259.o for inclusion, and
makes the platforms that need it select that symbol.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc: Use the indirect_pci.c from arch/powerpc/sysdev
Paul Mackerras [Wed, 26 Oct 2005 06:36:55 +0000]
ppc: Use the indirect_pci.c from arch/powerpc/sysdev

This defines a CONFIG_INDIRECT_PCI symbol to control whether it
gets used or not, and fixes the Kconfig to select that symbol for
platforms that need it.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Move ras.c into arch/powerpc/platforms/pseries
Michael Ellerman [Mon, 24 Oct 2005 05:07:30 +0000]
[PATCH] powerpc: Move ras.c into arch/powerpc/platforms/pseries

ras.o is only built for CONFIG_PPC_PSERIES, so move it into
arch/powerpc/platforms/pseries. Update Makefiles to suit.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Move firmware.h into include/asm-powerpc
Michael Ellerman [Mon, 24 Oct 2005 05:07:29 +0000]
[PATCH] powerpc: Move firmware.h into include/asm-powerpc

Move firmware.h into include/asm-powerpc.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Remove trailing \n" in HMT macros
Michael Ellerman [Mon, 24 Oct 2005 05:07:27 +0000]
[PATCH] powerpc: Remove trailing \n" in HMT macros

GCC 3.3.3 barfs on the trailing \n" in the HMT macros.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Remove duplicate definition of set_tb()
Michael Ellerman [Mon, 24 Oct 2005 05:07:26 +0000]
[PATCH] powerpc: Remove duplicate definition of set_tb()

Somewhere along the line we got two definitions of set_tb(). They look to
be identical although they're not textually identical. So remove the #ifdef
CONFIG_PPC64 version, leaving the common version in time.h.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Some minor cleanups to setup_32.c
Kumar Gala [Wed, 26 Oct 2005 04:57:33 +0000]
[PATCH] powerpc: Some minor cleanups to setup_32.c

* Removed of_show_percpuinfo and just report CPU frequency in generic
  show_cpuinfo code.
* Killed OCP and PPC_SYS related code which doesn't belong in the
  merge tree

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Add support for Book-E timer config to generic_calibrate_decr
Kumar Gala [Wed, 26 Oct 2005 04:02:59 +0000]
[PATCH] powerpc: Add support for Book-E timer config to generic_calibrate_decr

We need to initialize some control SPRS for timers on Book-E before
we start taking decrementer interrupts.

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] ppc: Fix m82xx_pci build
Becky Bruce [Mon, 24 Oct 2005 19:50:46 +0000]
[PATCH] ppc: Fix m82xx_pci build

A recent patch updated the name of pci_assign_all_busses to
pci_assign_all_buses.  This instance of its use wasn't corrected
by the original patch to use the new name.

Builds cleanly on ads8272.

Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Fix types.h
Becky Bruce [Fri, 21 Oct 2005 21:16:58 +0000]
[PATCH] powerpc: Fix types.h

Powerpc: Fix types.h

I noticed that Paul had already pulled the version of types.h that
is missing the config.h include into the merge tree - this patch adds
it back in.

Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: merge include/asm-ppc*/checksum.h into include/asm-powerpc/checksum.h
Kumar Gala [Thu, 20 Oct 2005 16:44:46 +0000]
[PATCH] powerpc: merge include/asm-ppc*/checksum.h into include/asm-powerpc/checksum.h

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: some prom.c cleanups
Kumar Gala [Thu, 20 Oct 2005 16:44:03 +0000]
[PATCH] powerpc: some prom.c cleanups

On !CONFIG_PPC_MULTIPLATFORM _machine is defined as 0.  This is ok, but
we can't assign a value to _machine then.

We may not have CONFIG_PCI available, so only build in support for
find_parent_pci_resource(), request_OF_resource(), release_OF_resource()
if PCI is enabled.  This is probably not the long term fix but works out
for now.

Make reg_property64 contain 64-bit elements on a 32-bit machine.

Mark the deprecated prom.c functions as __deprecated.

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Make sure we have an RTC before trying to adjust it
Kumar Gala [Thu, 20 Oct 2005 16:43:33 +0000]
[PATCH] powerpc: Make sure we have an RTC before trying to adjust it

Its valid for ppc_md.set_rtc_time to be NULL.  We need to check
that its non-NULL before trying to update the RTC.

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Add a shutdown member to vio_driver
Stephen Rothwell [Mon, 24 Oct 2005 07:40:23 +0000]
powerpc: Add a shutdown member to vio_driver

Add a shutdown member to struct vio_driver. We also need vio_bus_shutdown()
which converts from struct device to struct vio_dev and knows how to extract
the struct vio_driver.

Original patch adjusted for different location of vio.c.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: set the driver.owner field for all vio drivers
Stephen Rothwell [Mon, 24 Oct 2005 05:12:22 +0000]
powerpc: set the driver.owner field for all vio drivers

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: iseries: Fix a bogus comment
Stephen Rothwell [Mon, 24 Oct 2005 05:21:52 +0000]
powerpc: iseries: Fix a bogus comment

A comment in lpevents.c refers to code that's actually in HvCallEvent.h.
The code in HvCallEvent.h is pretty obvious, so just remove the comment
altogether.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Acked-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: don't duplicate name between vio_driver and device_driver
Stephen Rothwell [Mon, 24 Oct 2005 04:53:21 +0000]
powerpc: don't duplicate name between vio_driver and device_driver

Just set the name field directly in the device_driver structure
contained in the vio_driver struct.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: Move include/asm-ppc64/vio.h to include/asm-powerpc/vio.h
Stephen Rothwell [Mon, 24 Oct 2005 04:32:35 +0000]
powerpc: Move include/asm-ppc64/vio.h to include/asm-powerpc/vio.h

Move include/asm-ppc64/vio.h to include/asm-powerpc/vio.h, that's it.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: Merge arch/ppc64/kernel/vio.c into arch/powerpc/kernel/vio.c
Stephen Rothwell [Mon, 24 Oct 2005 04:22:37 +0000]
powerpc: Merge arch/ppc64/kernel/vio.c into arch/powerpc/kernel/vio.c

Merge arch/ppc64/kernel/vio.c into arch/powerpc/kernel/vio.c, update
the Makefiles to make it work, and make ARCH=ppc64 still work.

Michael's version put vio.c in arch/powerpc/sysedv but after consolting
Paulus, this one puts it in arch/powerpc/kernel.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

14 years agopowerpc: Make coff boot wrapper load the kernel at 8M
Paul Mackerras [Sun, 23 Oct 2005 07:26:32 +0000]
powerpc: Make coff boot wrapper load the kernel at 8M

Previously it claimed 7MB starting at the 9M point and loaded the
kernel there.  That meant that prom_init put the flattened device
tree above 16M.  On the 601 that caused the early device tree scan
to fail, since only 16MB are mapped with BATs on the 601.  Moving
this down to 8MB allows prom_init to put the flattened device tree
between 15M and 16M, so it works on the 601.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Run on old powermacs.
Paul Mackerras [Sun, 23 Oct 2005 07:23:21 +0000]
powerpc: Run on old powermacs.

Old powermacs have a number of differences from current machines:
- there is no interrupt tree in the device tree, just interrupt
  or AAPL,interrupt properties
- the chosen node in the device tree is called /chosen@0
- the OF claim method doesn't map the memory, so we have to do
  an explicit map call as well
- there is no /chosen/cpu property on SMP machines
- the NVRAM isn't structured as a set of partitions.

This adapts the merged powermac support code to cope with these
issues.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fix time code for 601 processors
Paul Mackerras [Sun, 23 Oct 2005 07:14:56 +0000]
powerpc: Fix time code for 601 processors

The 601 doesn't have the timebase register; instead it has an RTCL
register that counts nanoseconds and wraps at 1000000000, and an
RTCU register that counts seconds.  This makes the necessary changes
for the merged time code to use the RTCL/U registers when the kernel
is running on a 601.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoMerge changes from linux-2.6 by hand
Paul Mackerras [Sat, 22 Oct 2005 06:51:34 +0000]
Merge changes from linux-2.6 by hand

14 years agoppc64: Use arch/powerpc/platforms/powermac for powermac build.
Paul Mackerras [Sat, 22 Oct 2005 06:06:27 +0000]
ppc64: Use arch/powerpc/platforms/powermac for powermac build.

This switches the ARCH=ppc64 build to use arch/powerpc/platforms/powermac
instead of arch/ppc64/kernel/pmac*, and deletes the latter set of files.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Merge in 64-bit powermac support.
Paul Mackerras [Sat, 22 Oct 2005 06:02:39 +0000]
powerpc: Merge in 64-bit powermac support.

This brings in a lot of changes from arch/ppc64/kernel/pmac_*.c to
arch/powerpc/platforms/powermac/*.c and makes various minor tweaks
elsewhere.  On the powermac we now initialize ppc_md by copying
the whole pmac_md structure into it, which required some changes in
the ordering of initializations of individual fields of it.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Make set_rtc_time() return error code from lower-level function
Paul Mackerras [Sat, 22 Oct 2005 05:57:55 +0000]
powerpc: Make set_rtc_time() return error code from lower-level function

Previously it ignored the return value from ppc_md.set_rtc_time,
but in fact the functions that that can point to do return a
useful error code, so return it from set_rtc_time().

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Simplify secondary CPU startup on powermacs
Paul Mackerras [Sat, 22 Oct 2005 05:55:58 +0000]
ppc64: Simplify secondary CPU startup on powermacs

... for consistency with ppc32, and because this way is neater.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Rearrange btext initialization for consistency with ppc32
Paul Mackerras [Sat, 22 Oct 2005 05:36:52 +0000]
ppc64: Rearrange btext initialization for consistency with ppc32

Moved init_boot_display from arch/ppc64/kernel/pmac_setup.c to
arch/ppc64/kernel/btext.c and declared it in asm-ppc64/btext.h.
Call it from init_early rather than pmac_init_early.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Add a `primary' argument to pci_process_bridge_OF_ranges
Paul Mackerras [Sat, 22 Oct 2005 05:03:21 +0000]
ppc64: Add a `primary' argument to pci_process_bridge_OF_ranges

... for consistency with ppc32 and to make the powermac merge easier.
Also make it use just a single resource in the host bridge for multiple
consecutive elements of the ranges property.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64/powerpc: Fix time initialization on SMP systems
Paul Mackerras [Sat, 22 Oct 2005 04:55:23 +0000]
ppc64/powerpc: Fix time initialization on SMP systems

This moves smp_space_timers from arch/ppc64/kernel/smp.c to
arch/powerpc/kernel/time.c and makes it initialize last_jiffy[]
instead of paca[].next_jiffy_update_tb, since last_jiffy[] is
now what the time code uses.  It also declares smp_space_timers
in include/asm-powerpc/time.h and gets rid of an ifdef in
div128_by_32.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Fix delivery of RT signals to 32-bit processes.
Paul Mackerras [Sat, 22 Oct 2005 04:46:33 +0000]
ppc64: Fix delivery of RT signals to 32-bit processes.

An error in merging led to 32-bit processes getting the wrong link
register value on entry to RT signal handlers, and the wrong stack
chain as well.  This fixes it.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Move agp_special_page export to where it is defined
Paul Mackerras [Sat, 22 Oct 2005 04:42:51 +0000]
powerpc: Move agp_special_page export to where it is defined

... instead of exporting it in arch/*/kernel/ppc_ksyms.c.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] typo fix in last cpufreq powernow patch
Chris Wright [Fri, 21 Oct 2005 23:56:08 +0000]
[PATCH] typo fix in last cpufreq powernow patch

Not sure how it slipped by, but here's a trivial typo fix for powernow.

Signed-off-by: Chris Wright <chrisw@osdl.org>
[ It's "nurter" backwards.. Maybe we have a hillbilly The Shining fan? ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Call exit_itimers from do_exit, not __exit_signal
Roland McGrath [Fri, 21 Oct 2005 22:03:29 +0000]
[PATCH] Call exit_itimers from do_exit, not __exit_signal

When I originally moved exit_itimers into __exit_signal, that was the only
place where we could reliably know it was the last thread in the group
dying, without races.  Since then we've gotten the signal_struct.live
counter, and do_exit can reliably do group-wide cleanup work.

This patch moves the call to do_exit, where it's made without locks.  This
avoids the deadlock issues that the old __exit_signal code's comment talks
about, and the one that Oleg found recently with process CPU timers.

[ This replaces e03d13e985d48ac4885382c9e3b1510c78bd047f, which is why
  it was just reverted. ]

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agoRevert "Fix cpu timers exit deadlock and races"
Linus Torvalds [Fri, 21 Oct 2005 22:36:00 +0000]
Revert "Fix cpu timers exit deadlock and races"

Revert commit e03d13e985d48ac4885382c9e3b1510c78bd047f, to be replaced
by a much nicer fix from Roland.

14 years ago[PATCH] cpufreq: fix pending powernow timer stuck condition
Dave Jones [Fri, 21 Oct 2005 21:21:03 +0000]
[PATCH] cpufreq: fix pending powernow timer stuck condition

AMD recently discovered that on some hardware, there is a race condition
possible when a C-state change request goes onto the bus at the same
time as a P-state change request.

Both requests happen, but the southbridge hardware only acknowledges the
C-state change.  The PowerNow! driver is then stuck in a loop, waiting
for the P-state change acknowledgement.  The driver eventually times
out, but can no longer perform P-state changes.

It turns out the solution is to resend the P-state change, which the
southbridge will acknowledge normally.

Thanks to Johannes Winkelmann for reporting this and testing the fix.

Signed-off-by: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc64: Fix typo bug in iSeries hash code
David Gibson [Fri, 21 Oct 2005 03:41:19 +0000]
[PATCH] ppc64: Fix typo bug in iSeries hash code

This fixes a stupid typo bug in the iSeries hash table code.

When we place a hash PTE in the secondary bucket, instead of setting the
SECONDARY flag bit, as we should, we (redundantly) set the VALID flag.

This was introduced with the patch abolishing bitfields from the hash
table code.  Mea culpa, oops.  It hasn't been noticed until now because
in practice we don't hit the secondary bucket terribly often.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Fri, 21 Oct 2005 19:23:07 +0000]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

14 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/acme/net-2.6
Linus Torvalds [Fri, 21 Oct 2005 19:22:33 +0000]
Merge /pub/scm/linux/kernel/git/acme/net-2.6

14 years ago[PATCH] drm: another mga bug
Dave Airlie [Thu, 20 Oct 2005 22:49:00 +0000]
[PATCH] drm: another mga bug

The wrong state emission routines were being called for G550, and
consistent maps weren't correctly mapped...

Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc64: Fix pages marked dirty abusively
Benjamin Herrenschmidt [Fri, 21 Oct 2005 04:12:51 +0000]
[PATCH] ppc64: Fix pages marked dirty abusively

While working on 64K pages, I found this little buglet in our
update_mmu_cache() implementation.

The code calls __hash_page() passing it an "access" parameter (the type
of access that triggers the hash) containing the bits _PAGE_RW and
_PAGE_USER of the linux PTE.  The latter is useless in this case and the
former is wrong.  In fact, if we have a writeable PTE and we pass
_PAGE_RW to hash_page(), it will set _PAGE_DIRTY (since we track dirty
that way, by hash faulting !dirty) which is not what we want.

In fact, the correct fix is to always pass 0. That means that only
read-only or already dirty read write PTEs will be preloaded. The
(hopefully rare) case of a non dirty read write PTE can't be preloaded
this way, it will have to fault in hash_page on the actual access.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc64: Fix typo in time calculations
Paul Mackerras [Fri, 21 Oct 2005 12:39:36 +0000]
[PATCH] ppc64: Fix typo in time calculations

This fixes a typo in the div128_by_32 function used in the timekeeping
calculations on ppc64.  If you look at the code it's quite obvious
that we need (rb + c) rather than (rb + b).  The "b" is clearly just a
typo.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] mptsas: fix phy identifiers
Eric Moore [Fri, 21 Oct 2005 18:56:36 +0000]
[PATCH] mptsas: fix phy identifiers

This fixes handling of the phy identifiers in mptsas.

Signed-off-by: Eric Moore <Eric.Moore@lsil.com>
[ split it a pre-2.6.14 portion from Eric's bigger patch ]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] powerpc: Fix mmap returning 64 bit addresses
Michael Ellerman [Fri, 21 Oct 2005 06:01:34 +0000]
[PATCH] powerpc: Fix mmap returning 64 bit addresses

The merge of syscalls.c & sys_ppc32.c (30286ef6e044bc3d9019c3d8b900572e3fa05e65)
broke mmap, if the mmap returned a 64 bit address.

do_mmap2 was taking the return value from do_mmap_pgoff (an unsigned long), and
storing it in an int, before returning it to sys_mmap as an unsigned long. So
we were losing the high bits of the address.

You would have thought the compiler could catch this for us ...

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Don't blow away load_addr in start_thread
Michael Ellerman [Fri, 21 Oct 2005 06:01:33 +0000]
[PATCH] powerpc: Don't blow away load_addr in start_thread

The patch to make process.c work for 32-bit and 64-bit
(06d67d54741a5bfefa31945ef195dfa748c29025) broke some 64-bit binaries.
We were blowing away load_addr in gpr[2], so we weren't properly relocating
the entry point.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Merge thread_info.h
David Gibson [Fri, 21 Oct 2005 05:45:50 +0000]
[PATCH] powerpc: Merge thread_info.h

Merge ppc32 and ppc64 versions of thread_info.h.  They were pretty
similar already, the chief changes are:

- Instead of inline asm to implement current_thread_info(),
which needs to be different for ppc32 and ppc64, we use C with an
asm("r1") register variable.  gcc turns it into the same asm as we
used to have for both platforms.
- We replace ppc32's 'local_flags' with the ppc64
'syscall_noerror' field.  The noerror flag was in fact the only thing
in the local_flags field anyway, so the ppc64 approach is simpler, and
means we only need a load-immediate/store instead of load/mask/store
when clearing the flag.
- In readiness for 64k pages, when THREAD_SIZE will be less
than a page, ppc64 used kmalloc() rather than get_free_pages() to
allocate the kernel stack.  With this patch we do the same for ppc32,
since there's no strong reason not to.
- For ppc64, we no longer export THREAD_SHIFT and THREAD_SIZE
via asm-offsets, thread_info.h can now be safely included in asm, as
on ppc32.

Built and booted on G4 Powerbook (ARCH=ppc and ARCH=powerpc) and
Power5 (ARCH=ppc64 and ARCH=powerpc).

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] Fix broken initialization of conswitchp for ARCH=ppc64
David Gibson [Fri, 21 Oct 2005 05:43:42 +0000]
[PATCH] Fix broken initialization of conswitchp for ARCH=ppc64

In the merge tree, commit 0458060c1c59c5378d8fb5daabe18cf4681c35cd
broke boot on some machines because the initialization of conswitchp
was moved to arch/powerpc/kernel/setup_64.c, but a corresponding copy
was not added to arch/ppc64/kernel/setup.c.  This patch fixes it.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[ARM] Fix Integrator IM/PD-1 support
Russell King [Fri, 21 Oct 2005 09:17:37 +0000]
[ARM] Fix Integrator IM/PD-1 support

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[ARM] 3028/1: S3C2410 - add DCLK mask definitions
Ben Dooks [Thu, 20 Oct 2005 22:21:20 +0000]
[ARM] 3028/1: S3C2410 - add DCLK mask definitions

Patch from Ben Dooks

From: Guillaume Gourat <guillaume.gourat@nexvision.fr>

Add MASK definitions for DCLK0 and DCLK1

Signed-off-by: Guillaume Gourat <guillaume.gourat@nexvision.fr>
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[ARM] 3027/1: BAST - reduce NAND timings slightly
Ben Dooks [Thu, 20 Oct 2005 22:21:19 +0000]
[ARM] 3027/1: BAST - reduce NAND timings slightly

Patch from Ben Dooks

The current Simtec BAST nand area timings are a little
too slow to be obtained by a 2410 running at 266MHz,
so reduce the timings slightly to bring them into the
acceptable range.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[ARM] 3026/1: S3C2410 - avoid possible overflow in pll calculations
Ben Dooks [Thu, 20 Oct 2005 22:21:18 +0000]
[ARM] 3026/1: S3C2410 - avoid possible overflow in pll calculations

Patch from Ben Dooks

Avoid the possiblity that if the board is using
a 16.9334 or higher crystal with a high PLL
multiplier, then the pll value could overflow
the capability of an int.

Also fix the value types of the intermediate
variables to unsigned int.

Rewrite of patch from Guillaume Gourat

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[ARM] 3025/1: Add I2S platform device for PXA
Matt Reimer [Thu, 20 Oct 2005 22:21:18 +0000]
[ARM] 3025/1: Add I2S platform device for PXA

Patch from Matt Reimer

Adds an I2S platform_device for PXA. I2S is used to interface
with sound chips on systems like iPAQ h1910/h2200/hx4700 and
Asus 716.

Signed-off-by: mreimer@vpop.net
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

14 years ago[TCP] Allow len == skb->len in tcp_fragment
Herbert Xu [Thu, 20 Oct 2005 19:13:13 +0000]
[TCP] Allow len == skb->len in tcp_fragment

It is legitimate to call tcp_fragment with len == skb->len since
that is done for FIN packets and the FIN flag counts as one byte.
So we should only check for the len > skb->len case.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>

14 years ago[DCCP]: Clear the IPCB area
Herbert Xu [Tue, 18 Oct 2005 02:03:28 +0000]
[DCCP]: Clear the IPCB area

Turns out the problem has nothing to do with use-after-free or double-free.
It's just that we're not clearing the CB area and DCCP unlike TCP uses a CB
format that's incompatible with IP.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ian McDonald <imcdnzl@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>

14 years ago[DCCP]: Make dccp_write_xmit always free the packet
Herbert Xu [Sun, 16 Oct 2005 11:08:46 +0000]
[DCCP]: Make dccp_write_xmit always free the packet

icmp_send doesn't use skb->sk at all so even if skb->sk has already
been freed it can't cause crash there (it would've crashed somewhere
else first, e.g., ip_queue_xmit).

I found a double-free on an skb that could explain this though.
dccp_sendmsg and dccp_write_xmit are a little confused as to what
should free the packet when something goes wrong.  Sometimes they
both go for the ball and end up in each other's way.

This patch makes dccp_write_xmit always free the packet no matter
what.  This makes sense since dccp_transmit_skb which in turn comes
from the fact that ip_queue_xmit always frees the packet.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>

14 years ago[DCCP]: Use skb_set_owner_w in dccp_transmit_skb when skb->sk is NULL
Herbert Xu [Fri, 14 Oct 2005 06:38:49 +0000]
[DCCP]: Use skb_set_owner_w in dccp_transmit_skb when skb->sk is NULL

David S. Miller <davem@davemloft.net> wrote:
> One thing you can probably do for this bug is to mark data packets
> explicitly somehow, perhaps in the SKB control block DCCP already
> uses for other data.  Put some boolean in there, set it true for
> data packets.  Then change the test in dccp_transmit_skb() as
> appropriate to test the boolean flag instead of "skb_cloned(skb)".

I agree.  In fact we already have that flag, it's called skb->sk.
So here is patch to test that instead of skb_cloned().

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Ian McDonald <imcdnzl@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>

14 years ago[PATCH] Fix handling spurious page fault for hugetlb region
Hugh Dickins [Thu, 20 Oct 2005 15:24:28 +0000]
[PATCH] Fix handling spurious page fault for hugetlb region

This reverts commit 3359b54c8c07338f3a863d1109b42eebccdcf379 and
replaces it with a cleaner version that is purely based on page table
operations, so that the synchronization between inode size and hugetlb
mappings becomes moot.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agopowerpc: Fix some bugs in the new merged time code
Paul Mackerras [Thu, 20 Oct 2005 12:33:06 +0000]
powerpc: Fix some bugs in the new merged time code

I had the sense of the test for when to use the old 601-style RTC
registers inverted.  pmac_calibrate_decr and via_calibrate_decr
weren't setting ppc_tb_freq, on which all the further calculations
depended.  Lastly, update_gtod was losing the top 32 bits of
the new tb_to_xs value.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc/ppc/ppc64: Various compile fixes.
Paul Mackerras [Thu, 20 Oct 2005 11:10:09 +0000]
powerpc/ppc/ppc64: Various compile fixes.

This declares powersave_nap in system.h and makes it an int everywhere,
fixes typos for the maple platform, fixes a couple of places where
I missed removing the last two arguments from a message_pass function,
and makes ppc64 consistent with ppc32 in the type of the
pci_bridge.cfg_data field.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Use the merged mpic.c
Paul Mackerras [Thu, 20 Oct 2005 11:06:44 +0000]
ppc64: Use the merged mpic.c

This means we now compile in arch/powerpc/sysdev for ARCH=ppc64.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Move some calculations from xxx_calibrate_decr to time_init
Paul Mackerras [Thu, 20 Oct 2005 11:04:51 +0000]
powerpc: Move some calculations from xxx_calibrate_decr to time_init

Previously the individual xxx_calibrate_decr functions would each
print the timebase and cpu frequency and calculate several values
such as tb_to_us and tb_to_xs.  This moves those printks and
calculations into time_init just after the call to the platform's
calibrate_decr function.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Move init_boot_text call and conswitchp init into setup_arch
Paul Mackerras [Thu, 20 Oct 2005 11:00:20 +0000]
ppc64: Move init_boot_text call and conswitchp init into setup_arch

This way they get done in one place for all platforms, and it is
more consistent with what ppc32 does.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc: rename pci_assign_all_busses to pci_assign_all_buses
Paul Mackerras [Thu, 20 Oct 2005 10:57:05 +0000]
ppc: rename pci_assign_all_busses to pci_assign_all_buses

... for consistency with ppc64 and to make merging easier.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Merge various powermac-related header files.
Paul Mackerras [Thu, 20 Oct 2005 10:53:39 +0000]
powerpc: Merge various powermac-related header files.

Except for smu.h, which moved from asm-ppc64 to asm-powerpc, all
of these moved from asm-ppc to asm-powerpc.  In each case the
asm-ppc64 version (if there was one) was just a single line
including the asm-ppc version.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fix places where ppc_md.show_[per]cpuinfo was treated as int
Paul Mackerras [Thu, 20 Oct 2005 10:48:19 +0000]
powerpc: Fix places where ppc_md.show_[per]cpuinfo was treated as int

I missed a few places where ppc code was still assuming that the
ppc_md.show_[per]cpuinfo functions returned int.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Move smp_mpic_message_pass into mpic.c
Paul Mackerras [Thu, 20 Oct 2005 07:09:51 +0000]
powerpc: Move smp_mpic_message_pass into mpic.c

Having it here rather than in arch/ppc64/kernel/smp.c means that
we can use it on 32-bit SMP systems easily with ARCH=powerpc.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoppc64: Change ppc_md.get_cpuinfo to ppc_md.show_cpuinfo
Paul Mackerras [Thu, 20 Oct 2005 07:02:01 +0000]
ppc64: Change ppc_md.get_cpuinfo to ppc_md.show_cpuinfo

... for consistency with ppc32; also add in ppc32's show_percpuinfo
function.

Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agoLinux v2.6.14-rc5
Linus Torvalds [Thu, 20 Oct 2005 06:23:05 +0000]
Linux v2.6.14-rc5

The -rc4 release was supposed to be the last -rc, but here goes.  The
RCU fixes and the swiotlb changes need an -rc for final testing.

14 years ago[PATCH] powerpc: Merge types.h
Becky Bruce [Wed, 19 Oct 2005 23:45:03 +0000]
[PATCH] powerpc: Merge types.h

This patch merges types.h into include/asm-powerpc. The only real change is
the removal of the include of linux/config.h from the 32-bit version - it
doesn't appear to be necessary.

This patch has been built on several different 32 and 64-bit platforms,
and booted on mpc8540_ads.

Signed-off-by: Becky Bruce <becky.bruce@freescale.com>
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Merge ppc64 pmc.[ch] with ppc32 perfmon.[ch]
David Gibson [Wed, 19 Oct 2005 04:53:32 +0000]
[PATCH] powerpc: Merge ppc64 pmc.[ch] with ppc32 perfmon.[ch]

This patches the ppc32 and ppc64 versions of the headers and .c files
with helper functions for manipulating the performance counting
hardware.  As a side effect, it removes use of the term "perfmon" from
ppc32, thus avoiding confusion with the unrelated performance counter
interface from HP Labs also called "perfmon".

Built, but not booted, for g5, pSeries, iSeries, and 32-bit Powermac
with both ARCH=powerpc and ARCH=ppc{,64} as appropriate.

Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] build fix for uml/amd64
Al Viro [Tue, 18 Oct 2005 21:45:17 +0000]
[PATCH] build fix for uml/amd64

Missing half of the [PATCH] uml: Fix sysrq-r support for skas mode
We need to remove these (UPT_[DEFG]S) from the read side as well as the
write one - otherwise it simply won't build.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] scsi_error thread exits in TASK_INTERRUPTIBLE state.
Steven Rostedt [Wed, 19 Oct 2005 12:22:13 +0000]
[PATCH] scsi_error thread exits in TASK_INTERRUPTIBLE state.

Found in the -rt patch set.  The scsi_error thread likely will be in the
TASK_INTERRUPTIBLE state upon exit.  This patch fixes this bug.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc64: update defconfigs
Paul Mackerras [Wed, 19 Oct 2005 22:34:56 +0000]
[PATCH] ppc64: update defconfigs

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Thu, 20 Oct 2005 06:12:03 +0000]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

14 years ago[PATCH] swiotlb: make sure initial DMA allocations really are in DMA memory
Yasunori Goto [Wed, 19 Oct 2005 22:52:18 +0000]
[PATCH] swiotlb: make sure initial DMA allocations really are in DMA memory

This introduces a limit parameter to the core bootmem allocator; The new
parameter indicates that physical memory allocated by the bootmem
allocator should be within the requested limit.

We also introduce alloc_bootmem_low_pages_limit, alloc_bootmem_node_limit,
alloc_bootmem_low_pages_node_limit apis, but alloc_bootmem_low_pages_limit
is the only api used for swiotlb.

The existing alloc_bootmem_low_pages() api could instead have been
changed and made to pass right limit to the core allocator.  But that
would make the patch more intrusive for 2.6.14, as other arches use
alloc_bootmem_low_pages().  We may be done that post 2.6.14 as a
cleanup.

With this, swiotlb gets memory within 4G for both x86_64 and ia64
arches.

Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ravikiran G Thirumalai <kiran@scalex86.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] `unaligned access' in acpi get_root_bridge_busnr()
Peter Chubb [Thu, 20 Oct 2005 05:45:14 +0000]
[PATCH] `unaligned access' in acpi get_root_bridge_busnr()

In drivers/acpi/glue.c the address of an integer is cast to the address of
an unsigned long.  This breaks on systems where a long is larger than an
int --- for a start the int can be misaligned; for a second the assignment
through the pointer will overwrite part of the next variable.

Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au>
Acked-by: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] fix MGA DRM regression before 2.6.14
Dave Airlie [Thu, 20 Oct 2005 04:23:51 +0000]
[PATCH] fix MGA DRM regression before 2.6.14

I've gotten a report on lkml, of a possible regression in the MGA DRM in
2.6.14-rc4 (since -rc1), I haven't been able to reproduce it here, but I've
figured out some possible issues in the mga code that were definitely
wrong, some of these are from DRM CVS, the main fix is the agp enable bit
on the old code path still used by everyone.....

Signed-off-by: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Threads shouldn't inherit PF_NOFREEZE
Alan Stern [Thu, 20 Oct 2005 04:23:51 +0000]
[PATCH] Threads shouldn't inherit PF_NOFREEZE

The PF_NOFREEZE process flag should not be inherited when a thread is
forked.  This patch (as585) removes the flag from the child.

This problem is starting to show up more and more as drivers turn to the
kthread API instead of using kernel_thread().  As a result, their kernel
threads are now children of the kthread worker instead of modprobe, and
they inherit the PF_NOFREEZE flag.  This can cause problems during system
suspend; the kernel threads are not getting frozen as they ought to be.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Export RCS_TAR_IGNORE for rpm targets
Tom Rini [Thu, 20 Oct 2005 04:23:49 +0000]
[PATCH] Export RCS_TAR_IGNORE for rpm targets

The variable RCS_TAR_IGNORE is used in scripts/packaging/Makefile, but not
exported from the main Makefile, so it's never used.

This results in the rpm targets being very unhappy in quilted trees.

Signed-off-by: Tom Rini <trini@kernel.crashing.org>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc64: Fix error in vDSO 32 bits date
Benjamin Herrenschmidt [Thu, 20 Oct 2005 04:23:49 +0000]
[PATCH] ppc64: Fix error in vDSO 32 bits date

The implementation of __kernel_gettimeofday() in the 32 bits vDSO has a
small bug (a typo actually) that will cause it to lose 1 bit of precision.
Not terribly bad but worth fixing.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Three one-liners in md.c
NeilBrown [Thu, 20 Oct 2005 04:23:47 +0000]
[PATCH] Three one-liners in md.c

The main problem fixes is that in certain situations stopping md arrays may
take longer than you expect, or may require multiple attempts.  This would
only happen when resync/recovery is happening.

This patch fixes three vaguely related bugs.

1/ The recent change to use kthreads got the setting of the
   process name wrong.  This fixes it.
2/ The recent change to use kthreads lost the ability for
   md threads to be signalled with SIG_KILL.  This restores that.
3/ There is a long standing bug in that if:
    - An array needs recovery (onto a hot-spare) and
    - The recovery is being blocked because some other array being
       recovered shares a physical device and
    - The recovery thread is killed with SIG_KILL
   Then the recovery will appear to have completed with no IO being
   done, which can cause data corruption.
   This patch makes sure that incomplete recovery will be treated as
   incomplete.

Note that any kernel affected by bug 2 will not suffer the problem of bug
3, as the signal can never be delivered.  Thus the current 2.6.14-rc
kernels are not susceptible to data corruption.  Note also that if arrays
are shutdown (with "mdadm -S" or "raidstop") then the problem doesn't
occur.  It only happens if a SIGKILL is independently delivered as done by
'init' when shutting down.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] raw1394: fix locking in the presence of SMP and interrupts
Andy Wingo [Thu, 20 Oct 2005 04:23:46 +0000]
[PATCH] raw1394: fix locking in the presence of SMP and interrupts

Changes all spinlocks that can be held during an irq handler to disable
interrupts while the lock is held.  Changes spin_[un]lock_irq to use the
irqsave/irqrestore variants for robustness and readability.

In raw1394.c:handle_iso_listen(), don't grab host_info_lock at all -- we're
not accessing host_info_list or host_count, and holding this lock while
trying to tasklet_kill the iso tasklet this can cause an ABBA deadlock if
ohci:dma_rcv_tasklet is running and tries to grab host_info_lock in
raw1394.c:receive_iso.  Test program attached reliably deadlocks all SMP
machines I have been able to test without this patch.

Signed-off-by: Andy Wingo <wingo@pobox.com>
Acked-by: Ben Collins <bcollins@ubuntu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] orinoco: limit message rate
Andrew Morton [Thu, 20 Oct 2005 04:23:44 +0000]
[PATCH] orinoco: limit message rate

Brice Goglin <Brice.Goglin@ens-lyon.org> reports a printk storm from this
driver.  Fix.

Acked-by: David Gibson <hermes@gibson.dropbear.id.au>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] mm: hugetlb truncation fixes
Hugh Dickins [Thu, 20 Oct 2005 04:23:43 +0000]
[PATCH] mm: hugetlb truncation fixes

hugetlbfs allows truncation of its files (should it?), but hugetlb.c often
forgets that: crashes and misaccounting ensue.

copy_hugetlb_page_range better grab the src page_table_lock since we don't
want to guess what happens if concurrently truncated.  unmap_hugepage_range
rss accounting must not assume the full range was mapped.  follow_hugetlb_page
must guard with page_table_lock and be prepared to exit early.

Restyle copy_hugetlb_page_range with a for loop like the others there.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] Fix cpu timers exit deadlock and races
Roland McGrath [Thu, 20 Oct 2005 05:21:23 +0000]
[PATCH] Fix cpu timers exit deadlock and races

Oleg Nesterov reported an SMP deadlock.  If there is a running timer
tracking a different process's CPU time clock when the process owning
the timer exits, we deadlock on tasklist_lock in posix_cpu_timer_del via
exit_itimers.

That code was using tasklist_lock to check for a race with __exit_signal
being called on the timer-target task and clearing its ->signal.
However, there is actually no such race.  __exit_signal will have called
posix_cpu_timers_exit and posix_cpu_timers_exit_group before it does
that.  Those will clear those k_itimer's association with the dying
task, so posix_cpu_timer_del will return early and never reach the code
in question.

In addition, posix_cpu_timer_del called from exit_itimers during execve
or directly from timer_delete in the process owning the timer can race
with an exiting timer-target task to cause a double put on timer-target
task struct.  Make sure we always access cpu_timers lists with sighand
lock held.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Chris Wright <chrisw@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

14 years ago[PATCH] ppc32: replace use of _GLOBAL with .globl for ppc32
Kumar Gala [Tue, 18 Oct 2005 22:42:41 +0000]
[PATCH] ppc32: replace use of _GLOBAL with .globl for ppc32

The _GLOBAL() macro is for text symbols only. Changed to using
.globl for .data symbols.  This is also needed in ppc32 land
to allow FSL Book-E, 40x, and 44x to work.

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years ago[PATCH] powerpc: Some more fixes to allow building for a Book-E processor
Kumar Gala [Tue, 18 Oct 2005 22:42:09 +0000]
[PATCH] powerpc: Some more fixes to allow building for a Book-E processor

Some minor fixes that are needed if we are building for a book-e
processor.

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

14 years agopowerpc: Fix a corner case in __div64_32
Paul Mackerras [Wed, 19 Oct 2005 23:37:02 +0000]
powerpc: Fix a corner case in __div64_32

The code was incorrectly doing a division by 0 in the case where
the denominator was 0x100000000 and the divisor was 0xffffffff.
Thanks to Fred Liu of Motorola for pointing this out.

Signed-off-by: Paul Mackerras <paulus@samba.org>