11 Jan, 2006
7 commits
-
The following patch (against 2.6.15-rc5-mm3) fixes a kprobes build break
due to changes introduced in the kprobe locking in 2.6.15-rc5-mm3. In
addition, the patch reverts back the open-coding of kprobe_mutex.Signed-off-by: Ananth N Mavinakayanahalli
Acked-by: Anil S Keshavamurthy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Currently arch_remove_kprobes() is only implemented/required for x86_64 and
powerpc. All other architecture like IA64, i386 and sparc64 implementes a
dummy function which is being called from arch independent kprobes.c file.This patch removes the dummy functions and replaces it with
#define arch_remove_kprobe(p, s) do { } while(0)Signed-off-by: Anil S Keshavamurthy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Based on some feedback from Oleg Nesterov, I have made few changes to
previously posted patch.Signed-off-by: Anil S Keshavamurthy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Since Kprobes runtime exception handlers is now lock free as this code path is
now using RCU to walk through the list, there is no need for the
register/unregister{_kprobe} to use spin_{lock/unlock}_isr{save/restore}. The
serialization during registration/unregistration is now possible using just a
mutex.In the above process, this patch also fixes a minor memory leak for x86_64 and
powerpc.Signed-off-by: Anil S Keshavamurthy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Now that all these entries in the arch ioctl32.c files are gone [1], we can
build fs/compat_ioctl.c as a normal object and kill tons of cruft. We need a
special do_ioctl32_pointer handler for s390 so the compat_ptr call is done.
This is not needed but harmless on all other architectures. Also remove some
superflous includes in fs/compat_ioctl.cTested on ppc64.
[1] parisc still had it's PPP handler left, which is not fully correct
for ppp and besides that ppp uses the generic SIOCPRIV ioctl so it'd
kick in for all netdevice users. We can introduce a proper handler
in one of the next patch series by adding a compat_ioctl method to
struct net_device but for now let's just kill it - parisc doesn't
compile in mainline anyway and I don't want this to block this
patchset.Signed-off-by: Christoph Hellwig
Cc: Matthew Wilcox
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The comment in compat.c is wrong, every architecture provides a
get_compat_sigevent() for the IPC compat code already.This basically moves the x86_64 version to common code and removes all the
others.Signed-off-by: Christoph Hellwig
Acked-by: Paul Mackerras
Cc: Arnd Bergmann
Cc: Benjamin Herrenschmidt
Cc: Martin Schwidefsky
Cc: "David S. Miller"
Acked-by: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Cc: Eric Dumazet
Cc: Paul Mackerras
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
10 Jan, 2006
4 commits
-
This patch converts the inode semaphore to a mutex. I have tested it on
XFS and compiled as much as one can consider on an ia64. Anyway your
luck with it might be different.Modified-by: Ingo Molnar
(finished the conversion)
Signed-off-by: Jes Sorensen
Signed-off-by: Ingo Molnar -
Signed-off-by: Adrian Bunk
09 Jan, 2006
29 commits
-
We don't set CONFIG_PPC_MULTIPLATFORM on iSeries (yet). Avoid
compiling in the prom_init stuff on iSeries.Signed-off-by: Paul Mackerras
-
In add_memory() we should be using __va() to get a virtual address.
Spotted by Mike Kravetz.Signed-off-by: Michael Ellerman
Signed-off-by: Paul Mackerras -
The CONFIG_PPC_OF symbol is used to mean that the firmware device tree
access functions are available. Since we always have a device tree
with ARCH=powerpc, make CONFIG_PPC_OF always Y for ARCH=powerpc.This fixes some compile errors reported by Kumar Gala, but in a
different way to his patch. This also makes prom_parse.o be compiled
only if CONFIG_PPC_OF so that non-OF ARCH=ppc platforms will compile.Signed-off-by: Paul Mackerras
-
When taking a DABR exception we were reporting the PC. It makes more
sense to report the address that caused the exception, and the gdb guys
would like it that way.Signed-off-by: Anton Blanchard
Signed-off-by: Paul Mackerras -
POWER5+ adds new PMU groups and as such needs to be treated differently
by oprofile userspace. Change it to report itself as power5+.Signed-off-by: Anton Blanchard
Signed-off-by: Paul Mackerras -
My recent changes to oprofile broke it when built as a module. Fix it by
using an enum instead of a function pointer. This way we still retain
the oprofile configuration in the cputable.Signed-off-by: Anton Blanchard
Signed-off-by: Paul Mackerras -
This patch adds cpufreq support for all desktop "tower" G5 models. The
only G5 models still lacking cpufreq support at this point are the
Xserve and possibly the new iMac iSight (not tested). I'll have those
added soon. That patch uses the new platform functions interpreter to
implement frequency and voltage switching on most models.Note that in order to find the low frequency value, I had to hack
something that might now work properly on all models, so if the
frequency value reported when running low speed looks bogus to you,
please report it to me. (Appart from a bogus reported value, things
should work fine).Signed-off-by: Benjamin Herrenschmidt
Signed-off-by: Paul Mackerras -
This is the platform function interpreter itself along with the backends
for UniN/U3/U4, mac-io, GPIOs and i2c. It adds the ability to execute
those do-platform-* scripts in the device-tree (at least for most
devices for which a backend is provided). This should replace the clock
spreading hacks properly. It might also have an impact on all sort of
machines since some of the scripts marked "at init" will now be executed
on boot (or some other on sleep/wakeup), those will possibly do things
that the kernel didn't do at all, like setting some values into some i2c
devices (changing thermal sensor calibration or conversion rate) etc...
Thus regression testing is MUCH welcome. Also loook for errors in dmesg.
That's also why I've left rather verbose debugging enabled in this
version of the patch.(I do expect some Windtunnel G4s to show some errors as they have an i2c
clock chip on the PMU bus that uses some primitives that the i2c backend
doesn't implement yet. I really need users that have one of those
machine to come back to me so we can get that done right, though the
errors themselves should be harmless, I suspect the machine might not
run at full speed).Signed-off-by: Benjamin Herrenschmidt
Signed-off-by: Paul Mackerras -
This is the continuation of the previous patch. This one removes the old
PowerMac i2c drivers (i2c-keywest and i2c-pmac-smu) and replaces them
both with a single stub driver that uses the new PowerMac low i2c layer.Now that i2c-keywest is gone, the low-i2c code is extended to support
interrupt driver transfers. All i2c busses now appear as platform
devices. Compatibility with existing drivers should be maintained as the
i2c bus names have been kept identical, except for the SMU bus but in
that later case, all users has been fixed.With that patch added, matching a device node to an i2c_adapter becomes
trivial.Signed-off-by: Benjamin Herrenschmidt
Signed-off-by: Paul Mackerras -
This is the first part of a rework of the PowerMac i2c code. It
completely reworks the "low_i2c" layer. It is now more flexible,
supports KeyWest, SMU and PMU i2c busses, and provides functions to
match device nodes to i2c busses and adapters.This patch also extends & fix some bugs in the SMU driver related to i2c
support and removes the clock spreading hacks from the pmac feature code
rather than adapting them to the new API since they'll be replaced by
the platform function code completely in patch 3/5Signed-off-by: Benjamin Herrenschmidt
Signed-off-by: Paul Mackerras -
ppc64 has its own version of sys_time. It looks pretty scary, touching
a whole bunch of variables without any locking or memory ordering.In fact, a recent bugreport has shown it can actually go backwards. Time
to remove it and just use the generic sys_time, which is implemented on
top of do_gettimeofday.Signed-off-by: Anton Blanchard
Signed-off-by: Paul Mackerras -
For far, all SPU triggered interrupts always end up on
the first SMT thread, which is a bad solution.This patch implements setting the affinity to the
CPU that was running last when entering execution on
an SPU. This should result in a significant reduction
in IPI calls and better cache locality for SPE thread
specific data.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
One local variable is missing an __iomem modifier,
in another place, we pass a completely unused argument
with a missing __user modifier.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
In a hypervisor based setup, direct access to the first
priviledged register space can typically not be allowed
to the kernel and has to be implemented through hypervisor
calls.As suggested by Masato Noguchi, let's abstract the register
access trough a number of function calls. Since there is
currently no public specification of actual hypervisor
calls to implement this, I only provide a place that
makes it easier to hook into.Cc: Masato Noguchi
Cc: Geoff Levand
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
The logic for sys_spu_run keeps growing and it does
not really belong into file.c any more since we
moved away from using regular file operations to our
own syscall.No functional change in here.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
checking bits manually might not be synchonized with
the use of set_bit/clear_bit. Make sure we always use
the correct bitops by removing the unnecessary
identifiers.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
If creating one entry failed in spufs_fill_dir,
we never cleaned up the freshly created entries.
Fix this by calling the cleanup function on error.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
If get_unused_fd failed in sys_spu_create, we never cleaned
up the created directory. Fix that by restructuring the
error path.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
When spu_activate fails in spu_acquire_runnable, the
state must still be SPU_STATE_SAVED, we were
incorrectly setting it to SPU_STATE_RUNNABLE.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
During an earlier cleanup, we lost the serialization
of multiple spu_run calls performed on the same
spu_context. In order to get this back, introduce a
mutex in the spu_context that is held inside of spu_run.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
Only checking for SPUFS_MAGIC is not reliable, because
it might not be unique in theory. Worse than that,
we accidentally allow spu_run to be performed on
any file in spufs, not just those returned from
spu_create as intended.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
spu_forget will do mmput on the DMA address space,
which can lead to lots of other stuff getting triggered.
We better not hold a semaphore here that we might
need in the process.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
We need to check for validity of owner under down_write,
down_read is not enough.Noticed by Al Viro.
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
This patch enables support for pause(0) power management state
for the Cell Broadband Processor, which is import for power efficient
operation. The pervasive infrastructure will in the future enable
us to introduce more functionality specific to the Cell's
pervasive unit.From: Maximino Aguilar
Signed-off-by: Arnd Bergmann
Signed-off-by: Paul Mackerras -
Only build in support for ISA and PCI cases if we have enabled CONFIG_ISA
and CONFIG_PCI. Additionally, isa_bridge is a global so we shouldn't use
it a parameter name since it gets redefined to NULL when !CONFIG_PCI.Signed-off-by: Kumar Gala
Signed-off-by: Paul Mackerras -
At present, we are not looking at all interrupt controller nodes in the
device tree even though the proper node was not found. This is causing
the system panic. The attached patch will scan all nodes until it finds
the proper interrupt controller type.Signed-off-by: Haren Myneni
Signed-off-by: Paul Mackerras -
In setup_arch and setup_system call find_legacy_serial_ports() if we
build in support for 8250 serial ports instead of basing it on PPC_MULTIPLATFORM.Signed-off-by: Kumar Gala
Signed-off-by: Paul Mackerras -
Added a common udbg_progress for use by ppc_md.progress()
Signed-off-by: Kumar Gala
Signed-off-by: Paul Mackerras -
Add the ability to configure and initialize legacy 8250 serials
ports on an SOC bus. Also, fixed an issue that we would not
configure any serial ports if "linux,stdout-path" was not found.Signed-off-by: Kumar Gala
Signed-off-by: Paul Mackerras
