summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'master' of ↵Paul Mundt2010-04-265-6/+24
|\ | | | | | | | | | | | | | | | | | | | | master.kernel.org:/pub/scm/linux/kernel/git/mfleming/sh-2.6 * 'master' of master.kernel.org:/pub/scm/linux/kernel/git/mfleming/sh-2.6: sh: Use correct mask when comparing PMB DATA array values sh: Do not try merging two 128MB PMB mappings sh: Fix zImage load address when CONFIG_32BIT=y sh: Fix address to decompress at when CONFIG_32BIT=y sh: Assembly friendly __pa and __va definitions
| * sh: Use correct mask when comparing PMB DATA array valuesMatt Fleming2010-04-251-1/+4
| | | | | | | | | | | | | | Previously we were masking the PMB DATA array values with the value of __MEMORY_START | PMB_V, which misses some PFN bits off the mask. Signed-off-by: Matt Fleming <matt@console-pimps.org>
| * sh: Do not try merging two 128MB PMB mappingsMatt Fleming2010-04-251-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | There is a logic error in pmb_merge() that means we will incorrectly try to merge two 128MB PMB mappings into one mapping. However, 256MB isn't a valid PMB map size and pmb_merge() will actually drop the second 128MB mapping. This patch allows my SDK7786 board to boot when configured with CONFIG_MEMORY_SIZE=0x10000000. Signed-off-by: Matt Fleming <matt@console-pimps.org>
| * sh: Fix zImage load address when CONFIG_32BIT=yMatt Fleming2010-04-251-0/+6
| | | | | | | | | | | | | | We can't necessarily use the P1SEG region to access RAM when running in 32BIT mode, so use CONFIG_MEMORY_START as the base address. Signed-off-by: Matt Fleming <matt@console-pimps.org>
| * sh: Fix address to decompress at when CONFIG_32BIT=yMatt Fleming2010-04-251-0/+4
| | | | | | | | | | | | | | | | | | When running in 32BIT mode the P1SEG region doesn't necessarily provide a window onto RAM (it depends how the bootloader setup the PMB). The correct location to place the decompressed kernel is the physical address of _text. Signed-off-by: Matt Fleming <matt@console-pimps.org>
| * sh: Assembly friendly __pa and __va definitionsMatt Fleming2010-04-251-4/+9
| | | | | | | | | | | | | | | | | | This patch defines ___pa and ___va which return the physical and virtual address of an address, respectively. These macros are suitable for calling from assembly because they don't include the C casts required by __pa and __va. Signed-off-by: Matt Fleming <matt@console-pimps.org>
* | SH: constify multiple DMA related objects and references to themGuennadi Liakhovetski2010-04-267-67/+30
| | | | | | | | | | | | | | | | | | Lists of DMA channels and slaves are not changed, make them constant. Besides, SH7724 channel and slave configuration of both DMA controllers is identical, remove the extra copy of the configuration data. Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* | sh: invoke oom-killer from page faultNick Piggin2010-04-262-24/+8
|/ | | | | | | | | | | | As explained in commit 1c0fe6e3bd, we want to call the architecture independent oom killer when getting an unexplained OOM from handle_mm_fault, rather than simply killing current. Cc: linux-sh@vger.kernel.org Cc: linux-arch@vger.kernel.org Signed-off-by: Nick Piggin <npiggin@suse.de> Acked-by: David Rientjes <rientjes@google.com> Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: __cpuinit annotate the CPU init path.Paul Mundt2010-04-215-14/+13
| | | | | | | All of the regular CPU init path needs to be __cpuinit annotated for CPU hotplug. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: Tidy CPU probing and fixup section annotations.Paul Mundt2010-04-218-22/+9
| | | | | | | | This does a detect_cpu_and_cache_system() -> cpu_probe() rename, tidies up the unused return value, and stuffs it under __cpuinit in preparation for CPU hotplug. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: hw-breakpoints: Kill off stub unthrottle callback.Paul Mundt2010-04-212-6/+0
| | | | | | | This follows the x86 change and kills off the unthrottle stub. As the x86 change killed off the generic callback it isn't used anymore anyways. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: Zero out aliases counter when using SH-X3 hardware assistance.Paul Mundt2010-04-201-1/+10
| | | | | | | | | This zeroes out the number of cache aliases in the cache info descriptors when hardware alias avoidance is enabled. This cuts down on the amount of flushing taken care of by common code, and also permits coherency control to be disabled for the single CPU and 4k page size case. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: dwarf unwinder needs linux/module.h.Paul Mundt2010-04-201-0/+1
| | | | | | | | Previously the struct module definition was pulled in from other headers, but we want the reference to be explicit. Fixes up randconfig build issues. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: Enable SH-X3 hardware synonym avoidance handling.Paul Mundt2010-04-193-0/+43
| | | | | | | | | This enables support for the hardware synonym avoidance handling on SH-X3 CPUs for the case where dcache aliases are possible. icache handling is retained, but we flip on broadcasting of the block invalidations due to the lack of coherency otherwise on SMP. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: mach-sdk7786: pm_power_off support.Paul Mundt2010-04-192-0/+23
| | | | | | This wires up power-off support for the SDK7786 board. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: intc: IRQ auto-distribution support.Paul Mundt2010-04-156-79/+244
| | | | | | | | | | | | | | This implements support for hardware-managed IRQ balancing as implemented by SH-X3 cores (presently only hooked up for SH7786, but can probably be carried over to other SH-X3 cores, too). CPUs need to specify their distribution register along with the mask definitions, as these follow the same format. Peripheral IRQs that don't opt out of balancing will be automatically distributed at the whim of the hardware block, while each CPU needs to verify whether it is handling the IRQ or not, especially before clearing the mask. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: Disable IRQ balancing for timer and IPI IRQs.Paul Mundt2010-04-154-5/+8
| | | | | | | | Make sure that the timer IRQs and IPIs aren't enabled for IRQ balancing. IPIs are disabled as a result of being percpu while the timers simply disable balancing outright. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: intc: userimask support.Paul Mundt2010-04-135-2/+94
| | | | | | | | | | | This adds support for hardware-assisted userspace irq masking for special priority levels. Due to the SR.IMASK interactivity, only some platforms implement this in hardware (including but not limited to SH-4A interrupt controllers, and ARM-based SH-Mobile CPUs). Each CPU needs to wire this up on its own, for now only SH7786 is wired up as an example. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: intc: Tidy up loglevel mismatches.Paul Mundt2010-04-131-4/+7
| | | | | | | The printk loglevels are all over the place, make them a bit more coherent, and add some registration notification while we're at it. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: intc: Provide sysdev name for intc controllers.Paul Mundt2010-04-131-0/+16
| | | | | | | | | Presently the sysdevs are simply numbered based on the list position, without having any direct way of figuring out which controller these are actually mapping to. This provides a name attr for mapping out the chip name. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: dmaengine support for SH7786 DMAC0.Paul Mundt2010-04-132-2/+83
| | | | | | Hook up DMAC0 on SH7786. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* sh: Disable MMUCR_AT for SH-4 nommu.Paul Mundt2010-04-071-2/+9
| | | | | | | | Presently address translation is default-enabled regardless of whether CONFIG_MMU is set or not in the SH-4 case, this fixes it up, and also makes the control init word a bit more readable in the process. Signed-off-by: Paul Mundt <lethal@linux-sh.org>
* Merge branch 'sh/stable-updates'Paul Mundt2010-04-05351-3734/+5901
|\
| * Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6Linus Torvalds2010-04-047-21/+120
| |\ | | | | | | | | | | | | | | | | | | | | | | | | | | | * git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6: sunxvr500: Ignore secondary output PCI devices. sparc64: Implement perf_arch_fetch_caller_regs sparc64: Update defconfig. sparc64: Fix array size reported by vmemmap_populate() sparc: Fix regset register window handling. drivers/serial/sunsu.c: Correct use after free
| | * sunxvr500: Ignore secondary output PCI devices.David S. Miller2010-04-041-7/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These just represent the secondary and further heads attached to the card, and they have different sets of PCI bar registers to map. So don't try to drive them in the main driver. Reported-by: Frans van Berckel <fberckel@xs4all.nl> Tested-by: Frans van Berckel <fberckel@xs4all.nl> Signed-off-by: David S. Miller <davem@davemloft.net>
| | * sparc64: Implement perf_arch_fetch_caller_regsDavid S. Miller2010-04-041-0/+75
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We provide regs->tstate, regs->tpc, regs->tnpc and regs->u_regs[UREG_FP]. regs->tstate is necessary for: user_mode() (via perf_exclude_event()) perf_misc_flags() (via perf_prepare_sample()) regs->tpc is necessary for: perf_instruction_pointer() (via perf_prepare_sample()) and regs->u_regs[UREG_FP] is necessary for: perf_callchain() (via perf_prepare_sample()) The regs->tnpc value is provided just to be tidy. Signed-off-by: David S. Miller <davem@davemloft.net>
| | * sparc64: Update defconfig.David S. Miller2010-04-041-12/+16
| | | | | | | | | | | | Signed-off-by: David S. Miller <davem@davemloft.net>
| | * Merge branch 'master' of /home/davem/src/GIT/linux-2.6/David S. Miller2010-04-044501-121780/+251528
| | |\
| | * | sparc64: Fix array size reported by vmemmap_populate()Ben Hutchings2010-04-031-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | vmemmap_populate() attempts to report the used index and total size of vmemmap_table, but it wrongly shifts the total size so that it is always shown as 0. Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
| | * | sparc: Fix regset register window handling.David S. Miller2010-04-012-0/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | We have to adjust 'reg_window' down by 16 becuase the 'pos' iterator we'll use to index into the stack slots will be between 16 and 32. Signed-off-by: David S. Miller <davem@davemloft.net>
| | * | drivers/serial/sunsu.c: Correct use after freeJulia Lawall2010-03-301-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The of_iounmap is at the out_unmap label, but at that point up has already been freed. The free cannot be moved to the out_unmap label, because that label is reachable from cases where up should not be freed. So the call to of_iounmap is just duplicated, and the goto converted to a return. A simplified version of the semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression x,e; identifier f; iterator I; statement S; @@ *kfree(x); ... when != &x when != x = e when != I(x,...) S *x->f // </smpl> Signed-off-by: Julia Lawall <julia@diku.dk> Signed-off-by: David S. Miller <davem@davemloft.net>
| * | | Merge branch 'perf-fixes-for-linus' of ↵Linus Torvalds2010-04-0411-75/+153
| |\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: perf: Always build the powerpc perf_arch_fetch_caller_regs version perf: Always build the stub perf_arch_fetch_caller_regs version perf, probe-finder: Build fix on Debian perf/scripts: Tuple was set from long in both branches in python_process_event() perf: Fix 'perf sched record' deadlock perf, x86: Fix callgraphs of 32-bit processes on 64-bit kernels perf, x86: Fix AMD hotplug & constraint initialization x86: Move notify_cpu_starting() callback to a later stage x86,kgdb: Always initialize the hw breakpoint attribute perf: Use hot regs with software sched switch/migrate events perf: Correctly align perf event tracing buffer
| | * | | perf: Always build the powerpc perf_arch_fetch_caller_regs versionFrederic Weisbecker2010-04-031-2/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now that software events use perf_arch_fetch_caller_regs() too, we need the powerpc version to be always built. Fixes the following build error: (.text+0x3210): undefined reference to `perf_arch_fetch_caller_regs' (.text+0x3324): undefined reference to `perf_arch_fetch_caller_regs' (.text+0x33bc): undefined reference to `perf_arch_fetch_caller_regs' (.text+0x33ec): undefined reference to `perf_arch_fetch_caller_regs' (.text+0xd4a0): undefined reference to `perf_arch_fetch_caller_regs' arch/powerpc/kernel/built-in.o:(.text+0xd528): more undefined references to `perf_arch_fetch_caller_regs' follow make[1]: *** [.tmp_vmlinux1] Error 1 make: *** [sub-make] Error 2 Reported-by: Michael Ellerman <michael@ellerman.id.au> Reported-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Paul Mackerras <paulus@samba.org>
| | * | | perf: Always build the stub perf_arch_fetch_caller_regs versionFrederic Weisbecker2010-04-031-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now that software events use perf_arch_fetch_caller_regs() too, we need the stub version to be always built in for archs that don't implement it. Fixes the following build error in PARISC: kernel/built-in.o: In function `perf_event_task_sched_out': (.text.perf_event_task_sched_out+0x54): undefined reference to `perf_arch_fetch_caller_regs' Reported-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Paul Mackerras <paulus@samba.org>
| | * | | perf, probe-finder: Build fix on DebianBorislav Petkov2010-04-021-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Building chokes with: In file included from /usr/include/gelf.h:53, from /usr/include/elfutils/libdw.h:53, from util/probe-finder.h:61, from util/probe-finder.c:39: /usr/include/libelf.h:98: error: expected specifier-qualifier-list before 'off64_t' [...] Signed-off-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> LKML-Reference: <20100329164755.GA16034@aftab> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | perf/scripts: Tuple was set from long in both branches in python_process_event()Tom Zanussi2010-04-021-5/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is a fix to the signed/unsigned field handling in the Python scripting engine, based on a patch from Roel Kluin. Basically, Python wants to use a PyInt (which is internally a long) if it can i.e. if the value will fit into that type. If not, it stores it into a PyLong, which isn't actually a long, but an arbitrary-precision integer variable. The code below is similar to to what Python does internally, and it seems to work as expected on the x86 and x86_64 sytems I tested it on. Signed-off-by: Tom Zanussi <tzanussi@gmail.com> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Roel Kluin <roel.kluin@gmail.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: rostedt@goodmis.org LKML-Reference: <1270184305.6422.10.camel@tropicana> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | perf: Fix 'perf sched record' deadlockMike Galbraith2010-04-021-3/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | perf sched record can deadlock a box should the holder of handle->data->lock take an interrupt, and then attempt to acquire an rq lock held by a CPU trying to acquire the same lock. Disable interrupts. CPU0 CPU1 sched event with rq->lock held grab handle->data->lock spin on handle->data->lock interrupt try to grab rq->lock Reported-by: Li Zefan <lizf@cn.fujitsu.com> Signed-off-by: Mike Galbraith <efault@gmx.de> Tested-by: Li Zefan <lizf@cn.fujitsu.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> LKML-Reference: <1269598293.6174.8.camel@marge.simson.net> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | perf, x86: Fix callgraphs of 32-bit processes on 64-bit kernelsTorok Edwin2010-04-022-5/+44
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When profiling a 32-bit process on a 64-bit kernel, callgraph tracing stopped after the first function, because it has seen a garbage memory address (tried to interpret the frame pointer, and return address as a 64-bit pointer). Fix this by using a struct stack_frame with 32-bit pointers when the TIF_IA32 flag is set. Note that TIF_IA32 flag must be used, and not is_compat_task(), because the latter is only set when the 32-bit process is executing a syscall, which may not always be the case (when tracing page fault events for example). Signed-off-by: Török Edwin <edwintorok@gmail.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Paul Mackerras <paulus@samba.org> Cc: x86@kernel.org Cc: linux-kernel@vger.kernel.org LKML-Reference: <1268820436-13145-1-git-send-email-edwintorok@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | perf, x86: Fix AMD hotplug & constraint initializationPeter Zijlstra2010-04-022-36/+52
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit 3f6da39 ("perf: Rework and fix the arch CPU-hotplug hooks") moved the amd northbridge allocation from CPUS_ONLINE to CPUS_PREPARE_UP however amd_nb_id() doesn't work yet on prepare so it would simply bail basically reverting to a state where we do not properly track node wide constraints - causing weird perf results. Fix up the AMD NorthBridge initialization code by allocating from CPU_UP_PREPARE and installing it from CPU_STARTING once we have the proper nb_id. It also properly deals with the allocation failing. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> [ robustify using amd_has_nb() ] Signed-off-by: Stephane Eranian <eranian@google.com> LKML-Reference: <1269353485.5109.48.camel@twins> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | x86: Move notify_cpu_starting() callback to a later stagePeter Zijlstra2010-04-021-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Because we need to have cpu identification things done by the time we run CPU_STARTING notifiers. ( This init ordering will be relied on by the next fix. ) Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <1269353485.5109.48.camel@twins> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | Merge branch 'perf/urgent' of ↵Ingo Molnar2010-04-025-15/+25
| | |\ \ \ | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing into perf/urgent
| | | * | | x86,kgdb: Always initialize the hw breakpoint attributeJason Wessel2010-04-011-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | It is required to call hw_breakpoint_init() on an attr before using it in any other calls. This fixes the problem where kgdb will sometimes fail to initialize on x86_64. Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: 2.6.33 <stable@kernel.org> LKML-Reference: <1269975907-27602-1-git-send-email-jason.wessel@windriver.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
| | | * | | perf: Use hot regs with software sched switch/migrate eventsFrederic Weisbecker2010-04-013-12/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Scheduler's task migration events don't work because they always pass NULL regs perf_sw_event(). The event hence gets filtered in perf_swevent_add(). Scheduler's context switches events use task_pt_regs() to get the context when the event occured which is a wrong thing to do as this won't give us the place in the kernel where we went to sleep but the place where we left userspace. The result is even more wrong if we switch from a kernel thread. Use the hot regs snapshot for both events as they belong to the non-interrupt/exception based events family. Unlike page faults or so that provide the regs matching the exact origin of the event, we need to save the current context. This makes the task migration event working and fix the context switch callchains and origin ip. Example: perf record -a -e cs Before: 10.91% ksoftirqd/0 0 [k] 0000000000000000 | --- (nil) perf_callchain perf_prepare_sample __perf_event_overflow perf_swevent_overflow perf_swevent_add perf_swevent_ctx_event do_perf_sw_event __perf_sw_event perf_event_task_sched_out schedule run_ksoftirqd kthread kernel_thread_helper After: 23.77% hald-addon-stor [kernel.kallsyms] [k] schedule | --- schedule | |--60.00%-- schedule_timeout | wait_for_common | wait_for_completion | blk_execute_rq | scsi_execute | scsi_execute_req | sr_test_unit_ready | | | |--66.67%-- sr_media_change | | media_changed | | cdrom_media_changed | | sr_block_media_changed | | check_disk_change | | cdrom_open v2: Always build perf_arch_fetch_caller_regs() now that software events need that too. They don't need it from modules, unlike trace events, so we keep the EXPORT_SYMBOL in trace_event_perf.c Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net>
| | | * | | perf: Correctly align perf event tracing bufferFrederic Weisbecker2010-04-011-2/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The trace event buffer used by perf to record raw sample events is typed as an array of char and may then not be aligned to 8 by alloc_percpu(). But we need it to be aligned to 8 in sparc64 because we cast this buffer into a random structure type built by the TRACE_EVENT() macro to store the traces. So if a random 64 bits field is accessed inside, it may be not under an expected good alignment. Use an array of long instead to force the appropriate alignment, and perform a compile time check to ensure the size in byte of the buffer is a multiple of sizeof(long) so that its actual size doesn't get shrinked under us. This fixes unaligned accesses reported while using perf lock in sparc 64. Suggested-by: David Miller <davem@davemloft.net> Suggested-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: David Miller <davem@davemloft.net> Cc: Steven Rostedt <rostedt@goodmis.org>
| * | | | | Merge branch 'sched-fixes-for-linus' of ↵Linus Torvalds2010-04-042-5/+1
| |\ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: sched: set_cpus_allowed_ptr(): Don't use rq->migration_thread after unlock sched: Fix proc_sched_set_task()
| | * | | | | sched: set_cpus_allowed_ptr(): Don't use rq->migration_thread after unlockOleg Nesterov2010-04-021-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Trivial typo fix. rq->migration_thread can be NULL after task_rq_unlock(), this is why we have "mt" which should be used instead. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <20100330165829.GA18284@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| | * | | | | sched: Fix proc_sched_set_task()Mike Galbraith2010-04-021-4/+0
| | |/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Latencytop clearing sum_exec_runtime via proc_sched_set_task() breaks task_times(). Other places in kernel use nvcsw and nivcsw, which are being cleared as well, Clear task statistics only. Reported-by: Török Edwin <edwintorok@gmail.com> Signed-off-by: Mike Galbraith <efault@gmx.de> Cc: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com> Cc: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <1269940193.19286.14.camel@marge.simson.net> Signed-off-by: Ingo Molnar <mingo@elte.hu>
| * | | | | Merge branch 'tracing-fixes-for-linus' of ↵Linus Torvalds2010-04-042-5/+7
| |\ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'tracing-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: ring-buffer: Add missing unlock tracing: Fix lockdep warning in global_clock()
| | * | | | | ring-buffer: Add missing unlockJulia Lawall2010-03-291-3/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In some error handling cases the lock is not unlocked. The return is converted to a goto, to share the unlock at the end of the function. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @r exists@ expression E1; identifier f; @@ f (...) { <+... * spin_lock_irq (E1,...); ... when != E1 * return ...; ...+> } // </smpl> Signed-off-by: Julia Lawall <julia@diku.dk> LKML-Reference: <Pine.LNX.4.64.1003291736440.21896@ask.diku.dk> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| | * | | | | tracing: Fix lockdep warning in global_clock()Li Zefan2010-03-291-2/+2
| | | |/ / / | | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | # echo 1 > events/enable # echo global > trace_clock ------------[ cut here ]------------ WARNING: at kernel/lockdep.c:3162 check_flags+0xb2/0x190() ... ---[ end trace 3f86734a89416623 ]--- possible reason: unannotated irqs-on. ... There's no reason to use the raw_local_irq_save() in trace_clock_global. The local_irq_save() version is fine, and does not cause the bug in lockdep. Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> LKML-Reference: <4BA97FA1.7030606@cn.fujitsu.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
generated by cgit v1.2.3 (git 2.39.1) at 2025-01-04 09:35:19 +0100