10 Jul, 2007
2 commits
-
track TSC-unstable events and propagate it to the scheduler code.
Also allow sched_clock() to be used when the TSC is unstable,
the rq_clock() wrapper creates a reliable clock out of it.Signed-off-by: Ingo Molnar
-
the SMP load-balancer uses the boot-time migration-cost estimation
code to attempt to improve the quality of balancing. The reason for
this code is that the discrete priority queues do not preserve
the order of scheduling accurately, so the load-balancer skips
tasks that were running on a CPU 'recently'.this code is fundamental fragile: the boot-time migration cost detector
doesnt really work on systems that had large L3 caches, it caused boot
delays on large systems and the whole cache-hot concept made the
balancing code pretty undeterministic as well.(and hey, i wrote most of it, so i can say it out loud that it sucks ;-)
under CFS the same purpose of cache affinity can be achieved without
any special cache-hot special-case: tasks are sorted in the 'timeline'
tree and the SMP balancer picks tasks from the left side of the
tree, thus the most cache-cold task is balanced automatically.Signed-off-by: Ingo Molnar
08 Jul, 2007
1 commit
-
The printk level in this printk is bogus, as the previous printk
didn't have a terminating \n resulting in ..Intel E7520/7320/7525 detected.Disabling irq balancing and affinity
It also never printed a \n at all in the case where we didn't do
the quirk.Change it to only make noise if it actually does something useful.
Signed-off-by: Dave Jones
Signed-off-by: Linus Torvalds
07 Jul, 2007
5 commits
-
Writing to MSR 0x51400017 forces a hard reset on CS5536-based machines,
this has the reboot fixup do just that if such a board is detected.Acked-by: Jordan Crouse
Signed-off-by: Andres Salomon
Cc: Alan Cox
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
[MIPS] Fix scheduling latency issue on 24K, 34K and 74K cores
[MIPS] Add macros to encode processor revisions.
[MIPS] RM7000: Enable ICACHE_REFILLS_WORKAROUND_WAR.
[MIPS] SMTC: Fix cut'n'paste bug in Kconfig.debug
[MIPS] Change libgcc-style functions from lib-y to obj-y
[MIPS] Fix timer/performance interrupt detection
[MIPS] AP/SP: Avoid triggering the 34K E125 performance issue
[MIPS] 64-bit TO_PHYS_MASK macro for RM9000 processors -
o Commit 1833d6bc72893265f22addd79cf52e6987496e0f broke the build if
compiled with CONFIG_ES7000=y and CONFIG_X86_GENERICARCH=narch/i386/kernel/built-in.o(.init.text+0x4fa9): In function `acpi_parse_madt':
: undefined reference to `acpi_madt_oem_check'
arch/i386/kernel/built-in.o(.init.text+0x7406): In function `smp_read_mpc':
: undefined reference to `mps_oem_check'
arch/i386/kernel/built-in.o(.init.text+0x8990): In function
`connect_bsp_APIC':
: undefined reference to `enable_apic_mode'
make: *** [.tmp_vmlinux1] Error 1o Fix the build issue. Provided the definitions of missing functions.
o Don't have ES7000 machine. Only compile tested.
Cc: Len Brown
Cc: Natalie Protasevich
Cc: Roland Dreier
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Processors synchronization in set_mtrr requires the .gate field to be set
after .count field is properly initialized. Without an explicit barrier,
the compiler was reordering those memory stores. That was sometimes
causing a processor (in ipi_handler) to see the .gate change and decrement
.count before the latter is set by set_mtrr() (which then hangs in a
infinite loop with irqs disabled).Signed-off-by: Loic Prylli
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The commit 635cf99a80f4ebee59d70eb64bb85ce829e4591f introduced a
regression. Executing a ptrace single step after certain int80
accesses will infinitely loop and never advance the PC.The TIF_SINGLESTEP check should be done on the return from the syscall
and not before it.I loops on each single step on the pop right after the int80 which writes out
to the console. At that point you can issue as many single steps as you want
and it will not advance any further.The test case is below:
/* Test whether singlestep through an int80 syscall works.
*/
#define _GNU_SOURCE
#include
#include
#include
#include
#include
#include
#include
#includestatic int child, status;
static struct user_regs_struct regs;static void do_child()
{
char str[80] = "child: int80 test\n";ptrace(PTRACE_TRACEME, 0, 0, 0);
kill(getpid(), SIGUSR1);
write(fileno(stdout),str,strlen(str));
asm ("int $0x80" : : "a" (20)); /* getpid */
}static void do_parent()
{
unsigned long eip, expected = 0;
again:
waitpid(child, &status, 0);
if (WIFEXITED(status) || WIFSIGNALED(status))
return;if (WIFSTOPPED(status)) {
ptrace(PTRACE_GETREGS, child, 0, ®s);
eip = regs.eip;
if (expected)
fprintf(stderr, "child stop @ %08lx, expected %08lx %s\n",
eip, expected,
eip == expected ? "" : "
Cc: Jeremy Fitzhardinge
Cc:
Cc: Chuck Ebbert
Acked-by: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
06 Jul, 2007
8 commits
-
The idle loop goes to sleep using the WAIT instruction if !need_resched().
This has is suffering from from a race condition that if if just after
need_resched has returned 0 an interrupt might set TIF_NEED_RESCHED but
we've just completed the test so go to sleep anyway. This would be
trivial to fix by just disabling interrupts during that sequence as in:local_irq_disable();
if (!need_resched())
__asm__("wait");
local_irq_enable();but the processor architecture leaves it undefined if a processor calling
WAIT with interrupts disabled will ever restart its pipeline and indeed
some processors have made use of the freedom provided by the architecture
definition. This has been resolved and the Config7.WII bit indicates that
the use of WAIT is safe on 24K, 24KE and 34K cores. It also is safe on
74K starting revision 2.1.0 so enable the use of WAIT with interrupts
disabled for 74K based on a c0_prid of at least that.Signed-off-by: Ralf Baechle
-
This effectivly turned the SMTC_IDLE_HOOK_DEBUG debug option into a no-op.
Signed-off-by: Ralf Baechle
-
Reported by Eugene Surovegin .
If only modules were users of these functions they did not get linked into
the kernel proper, so later module loads would fail as well.Signed-off-by: Ralf Baechle
-
Signed-off-by: Chris Dearman
Signed-off-by: Ralf Baechle -
C0_status doesn't need to be initialized at this point anyway; the register
will be initialized later.Signed-off-by: Ralf Baechle
-
* master.kernel.org:/home/rmk/linux-2.6-arm:
[ARM] always allow dump_stack() to produce a backtrace
[ARM] Fix non-page aligned boot time mappings
[ARM] 4458/1: pxa: Fix CKEN usage and hence fix pxa suspend/resume
[ARM] 4454/1: Use word accesses in Versatile PCI config reads -
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
[POWERPC] Update defconfigs
[POWERPC] Uninline and export virq_to_hw() for the pasemi_mac driver
[POWERPC] Fix PMI breakage in cbe_cbufreq driver
[POWERPC] Disable old EMAC driver in arch/powerpc -
Don't make this dependent on CONFIG_DEBUG_KERNEL - if we hit a WARN_ON
we need the stack trace to work out how we got to that point.Signed-off-by: Russell King
05 Jul, 2007
1 commit
-
AT91SAM9260 stopped booting with the recent changes to MM
initialisation - it was asking for a non-aligned virtual address
which caused loops to be non-terminal. Fix this by rounding
virtual addresses down, but remember to include the offset in
the length, and round the length up to the following page.This means that asking for a mapping of 4K starting at 2K into
a page maps two pages as one would expect.Signed-off-by: Russell King
04 Jul, 2007
4 commits
-
Signed-off-by: Ralf Baechle
-
When Andi reverted the HPET resource reservation (in commit
0f8dc2f06560e2ca126d1670a24126ba08357d38), he didn't remove the now
unused variables, which just causes gcc to be noisy.Signed-off-by: Linus Torvalds
-
With this change it works again when the nmi watchdog is disabled.
Signed-off-by: Andi Kleen
Cc: Björn Steinbrink
Cc: Stephane Eranian
Cc: Andrew Morton
Signed-off-by: Linus Torvalds -
Matthias Lenk reports that the PCI subsystem would move the HPET on
SB400/SB600-based systems, where the HPET is in BAR1 of the SMbus
controller.The reason? The ACPI layer registered the PCI MMIO range as being busy
too early, before PCI enumeration had happened, causing the PCI layer to
decide that it should relocate the resources somewhere else.Firmware resources should be marked busy _after_ the PCI enumeration and
probing has happened, not before.Remove the too-early reservation, we'll fix it up to do it properly
later. In the meantime, this solves the regression.Tested-by: Matthias Lenk
Cc: Aaron Durbin
Cc: Andrew Morton
Signed-off-by: Linus Torvalds
02 Jul, 2007
9 commits
-
The PXA CKEN changes broken syspend/resume on the pxa27x. This patch
corrects the problem and fixes another couple of bad references.Signed-off-by: Richard Purdie
Signed-off-by: Russell King -
ARM Versatile PCI config reads of one byte width have the lowest two
bits of the address cleared and result in reading from a wrong place
in the config space. This change is to use word size accesses like it is done for halfword reads.Byte reads are used for retrieving the IRQ number of a PCI device and the problem was not exposed until 2.6.20 because the value read was discarded in drivers/pci/setup-irq.c (recently fixed).
Signed-off-by: Andrzej Zaborowski
Acked-by: Paul Brook
Signed-off-by: Russell King -
Signed-off-by: Mike Frysinger
Signed-off-by: Bryan Wu -
Signed-off-by: Kumar Gala
-
Uninline virq_to_hw and export it so modules can use it. The alternative
would be to export the irq_map array instead, but it's an infrequently
called function, and keeping the array unexported seems considerably
cleaner.This is needed so that the pasemi_mac driver can be compiled as a module.
Signed-off-by: Olof Johansson
Signed-off-by: Paul Mackerras -
The recent change to cell_defconfig to enable cpufreq on Cell exposed
the fact that the cbe_cpufreq driver currently needs the PMI interface
code to compile, but Kconfig doesn't make sure that the PMI interface
code gets built if cbe_cpufreq is enabled.In fact cbe_cpufreq can work without PMI, so this ifdefs out the code
that deals with PMI. This is a minimal solution for 2.6.22; a more
comprehensive solution will be merged for 2.6.23.Signed-off-by: Christian Krafft
Signed-off-by: Paul Mackerras -
Commit 52ade9b3b97fd3bea42842a056fe0786c28d0555 changed the suspend code
ordering to execute pm_ops->prepare() after the device model per-device
.suspend() calls in order to fix some ACPI-related issues. Unfortunately, it
broke the at91 platform which assumed that pm_ops->prepare() would be called
before suspending devices.at91 used pm_ops->prepare() to get notified of the target system sleep state,
so that it could use this information while suspending devices. However, with
the current suspend code ordering pm_ops->prepare() is called too late for
this purpose. Thus, at91 needs an additional method in 'struct pm_ops' that
will be used for notifying the platform of the target system sleep state.
Moreover, in the future such a method will also be needed by ACPI.This patch adds the .set_target() method to 'struct pm_ops' and makes the
suspend code call it, if implemented, before executing the device model
per-device .suspend() calls. It also modifies the at91 code to use
pm_ops->set_target() instead of pm_ops->prepare().Signed-off-by: Rafael J. Wysocki
Acked-by: David Brownell
Cc: Pavel Machek
Cc: Johannes Berg
Cc: Len Brown
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Commit 3ebad5905609476a4ff1151a66b21d9794009961 ("[PATCH] x86: Save and
restore the fixed-range MTRRs of the BSP when suspending") added mtrr
operations without verifying that the CPU has MTRRs. Crashes transmeta
CPUs.Cc: "H. Peter Anvin"
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Commit 9215da33209b861b01c51382254b178a3fe92a30 "fixed" the MTRR range
check to not allow any MTRR's under the 1MB mark (since that's where the
fixed MTRR's are active).However, that was totally bogus, since it's normal (and almost required)
to have a large variable MTRR that starts at 0, and covers some large
percentage of the whole RAM, and then using the fixed MTRR's to override
that large MTRR to handle the special ISA hole in the 640k-1M region.The old check was bogus too (checking that no variable MTRR is used that
is entirely under the 1MB range), but at least it wasn't actively
detrimental, because no sane situation would ever trigger such MTRR
usage in the first place.That said, the whole notion of not allowing variable MTRR's in the low
1MB is just stupid, so rather than revert the commit, this just removes
the whole sad and unnecessary check entirely.Cc: Jan Beulich
Cc: Andi Kleen
Cc: Andrew Morton
Tested-by: Luca Palermo
Signed-off-by: Linus Torvalds
30 Jun, 2007
2 commits
-
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
[SPARC64]: Add linux/pagemap.h to asm/tlb.h
[SPARC64]: Need to set state to IDLE during sun4v IRQ enable.
[SPARC64]: Fix VIRQ enabling.
[SPARC64]: Add irqs to mdesc_node. -
The vdso64 portion of patch 74609f4536f2b8fd6a48381bbbe3cd37da20a527 for
fixing problems with NULL gettimeofday input mistakenly checks for a
null tz field twice, when it should be checking for null tz once, and
null tv once; by way of a r10/r11 typo.Any application calling gettimeofday(&tv,NULL) will "fail".
This corrects that typo, and makes my G5 happy.
Tested on G5.
Signed-off-by: Will Schmidt
Cc: Tony Breeds
Forwarded-by: Ben Herrenschmidt[ Ben says: "I checked the 32 bits part of the change is correct. You
can probably blame me for originally writing the 2 versions with
inversed usage of r10 and r11, thus confusing Tony :-)"Ben duly blamed. - Linus ]
Signed-off-by: Linus Torvalds
29 Jun, 2007
3 commits
-
Not all the world is an i386. Many architectures need 64-bit arguments to be
aligned in suitable pairs of registers, and the original
sys_sync_file_range(int, loff_t, loff_t, int) was therefore wasting an
argument register for padding after the first integer. Since we don't
normally have more than 6 arguments for system calls, that left no room for
the final argument on some architectures.Fix this by introducing sys_sync_file_range2(int, int, loff_t, loff_t) which
all fits nicely. In fact, ARM already had that, but called it
sys_arm_sync_file_range. Move it to fs/sync.c and rename it, then implement
the needed compatibility routine. And stop the missing syscall check from
bitching about the absence of sys_sync_file_range() if we've implemented
sys_sync_file_range2() instead.Tested on PPC32 and with 32-bit and 64-bit userspace on PPC64.
Signed-off-by: David Woodhouse
Acked-by: Russell King
Cc: Arnd Bergmann
Cc: Paul Mackerras
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
arch/mips/jazz/setup.c:55:4: error: Initializer entry defined twice
Signed-off-by: Alexey Dobriyan
Acked-by: Ralf Baechle
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
main.c::mtrr_add() or mtrr_del() [exported]
calls main.c::mtrr_add_page() or mtrr_del_page() or mtrr_restore() [resume]
calls main.c::set_mtrr()
calls main.c::ipi_handler()
calls main.c::mtrr_if->set_all() == which can be cyrix_set_allWARNING: arch/i386/kernel/built-in.o(.text+0x8657): Section mismatch: reference to .init.data: (between 'cyrix_set_all' and 'centaur_get_free_region')
WARNING: arch/i386/kernel/built-in.o(.text+0x866b): Section mismatch: reference to .init.data: (between 'cyrix_set_all' and 'centaur_get_free_region')
WARNING: arch/i386/kernel/built-in.o(.text+0x867e): Section mismatch: reference to .init.data: (between 'cyrix_set_all' and 'centaur_get_free_region')
WARNING: arch/i386/kernel/built-in.o(.text+0x8684): Section mismatch: reference to .init.data: (between 'cyrix_set_all' and 'centaur_get_free_region')
WARNING: arch/i386/kernel/built-in.o(.text+0x868a): Section mismatch: reference to .init.data: (between 'cyrix_set_all' and 'centaur_get_free_region')Signed-off-by: Randy Dunlap
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
28 Jun, 2007
1 commit
-
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
[IA64] Make SN2 PCI code use ioremap rather than manually mangle the address
[IA64] Force error to surface in nofault code
[IA64] change sh_change_coherence oemcall to use nolock
[IA64] remove duplicate header include line
[IA64] Correct unwind validation code
[IA64] is_power_of_2-ia64/mm/hugetlbpage.c
27 Jun, 2007
4 commits
-
Force irq migration path during cpu offline, is not using proper locks and
irq_chip mask/unmask routines. This will result in some races(especially
the device generating the interrupt can see some inconsistent state,
resulting in issues like stuck irq,..).Appended patch fixes the issue by taking proper lock and encapsulating
irq_chip set_affinity() with a mask() before and an unmask() after.This fixes a MSI irq stuck issue reported by Darrick Wong.
There are several more general bugs in this area(irq migration in the
process context). For example,1. Possibility of missing edge triggered irq.
2. Reliable method of migrating level triggered irq in the process context.We plan to look and close these in the near future.
Eric says:
In addition even with the fix from Suresh there is still at least one
nasty hardware race in fixup_irqs(). However we exercise that code
path rarely enough that we are unlikely to hit it in the real world,
and that race seems to have existed since the code was merged. And a
fix for that is not coming soon as it is an open investigation area
if we can fix irq migration to work outside of irq context or if
we have to rework the requirements imposed by the generic cpu hotplug
and layer on fixup_irqs(). So this may come up again.Signed-off-by: Suresh Siddha
Cc: Eric W. Biederman
Reported-and-tested-by: Darrick Wong
Cc: Andi Kleen
Acked-by: Rafael J. Wysocki
Signed-off-by: Linus Torvalds -
set the irq_chip name for lapic.
Signed-off-by: Suresh Siddha
Signed-off-by: Linus Torvalds -
* master.kernel.org:/home/rmk/linux-2.6-arm:
[ARM] 4449/1: more entries in arch/arm/boot/.gitignore
[ARM] 4452/1: Force the literal pool dump before reloc_end
[ARM] Update show_regs/oops register format
[ARM] Add support for pause_on_oops and display preempt/smp options -
* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
[MIPS] Count timer interrupts correctly.
[MIPS] SMTC and non-SMTC kernel and modules are incompatible
[MIPS] EMMA2RH: Disable GEN_RTC, it can't possibly work.
[MIPS] Remove a duplicated local variable in test_and_clear_bit()
[MIPS] use compat_siginfo in rt_sigframe_n32
[MIPS] 20K: Handle WAIT related bugs according to errata information
[MIPS] AP/SP requires shadow registers, auto enable support.
[MIPS] Fix pb1500 reg B access
[MIPS] Alchemy: Fix wrong cast
[MIPS] remove "support for" from system type entry
[MIPS] add io_map_base to pci_controller on Cobalt
[MIPS] __ucmpdi2 arguments are unsigned long long.
