- 07 Sep, 2012 2 commits
-
-
Anton Blanchard authored
The pseries firmware currently refuses any non power of two MSI-X request. Unfortunately most network drivers end up asking for that because they want a power of two for RX queues and one or two extra for everything else. This patch rounds up the firmware request to the next power of two if the quota allows it. If this fails we fall back to using the original request size. Signed-off-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org>
-
Gavin Shan authored
When PCI probe flag PCI_REASSIGN_ALL_RSRC has been passed into PCI core, it's hoped that all resources to be reassigned by PCI core. As to particular P2P (PCI-to-PCI) bridge, the size of the corresponding BAR (I/O, MMIO, prefetchable MMIO) is calculated by the resources required by the PCI devices behind the P2P bridge. That means that the information like start/end address retrieved from the hardware registers of the P2P bridge is meainingless in the case. However, we still count that in and the BARs might have been configured by firmware with non-zero size. That leads to space waste. The patch explicitly sets the size of P2P bridge BARs to zero in case that resource reassignment is expected with PCI probe flag PCI_REASSIGN_ALL_RSRC. In the result, it will save overall resource required by the system without waste. Signed-off-by:
Gavin Shan <shangw@linux.vnet.ibm.com> Signed-off-by:
-
- 06 Sep, 2012 3 commits
-
-
Benjamin Herrenschmidt authored
Brings in various bug fixes from 3.6-rcX
-
Ananth N Mavinakayanahalli authored
commit: 8b7b80b9 [24/29] powerpc: Uprobes port to powerpc Caused a clash with the fore200e driver: In file included from drivers/atm/fore200e.c:70:0: drivers/atm/fore200e.h:263:3: error: redefinition of typedef 'opcode_t' with different type arch/powerpc/include/asm/probes.h:25:13: note: previous declaration of 'opcode_t' was here Fix the namespace clash by making opcode_t in probes.h to ppc_opcode_t. Signed-off-by:
Ananth N Mavinakayanahalli <ananth@in.ibm.com> Signed-off-by:
-
Mihai Caraman authored
Critical exception on 64-bit booke uses user-visible SPRG3 as scratch. Restore VDSO information in SPRG3 on exception prolog. Use a common sprg3 field in PACA for all powerpc64 architectures. Signed-off-by:
Mihai Caraman <mihai.caraman@freescale.com> Signed-off-by:
-
- 05 Sep, 2012 35 commits
-
-
Benjamin Herrenschmidt authored
patch_instruction() can be called very early on ppc32, when the kernel isn't yet running at it's linked address. That can cause the ! is_kernel_addr() test in __put_user() to trip and call might_sleep() which is very bad at that point during boot. Use a lower level function instead for now, at least until we get to rework ppc32 boot process to do the code patching later, like ppc64 does. Signed-off-by: -
Paul Mackerras authored
We have been observing hangs, both of KVM guest vcpu tasks and more generally, where a process that is woken doesn't properly wake up and continue to run, but instead sticks in TASK_WAKING state. This happens because the update of rq->wake_list in ttwu_queue_remote() is not ordered with the update of ipi_message in smp_muxed_ipi_message_pass(), and the reading of rq->wake_list in scheduler_ipi() is not ordered with the reading of ipi_message in smp_ipi_demux(). Thus it is possible for the IPI receiver not to see the updated rq->wake_list and therefore conclude that there is nothing for it to do. In order to make sure that anything done before smp_send_reschedule() is ordered before anything done in the resulting call to scheduler_ipi(), this adds barriers in smp_muxed_message_pass() and smp_ipi_demux(). The barrier in smp_muxed_message_pass() is a full barrier to ensure that there is a full ordering between the smp_send_reschedule() caller and scheduler_ipi(). In smp_ipi_demux(), we use xchg() rather than xchg_local() because xchg() includes release and acquire barriers. Using xchg() rather than xchg_local() makes sense given that ipi_message is not just accessed locally. This moves the barrier between setting the message and calling the cause_ipi() function into the individual cause_ipi implementations. Most of them -- those that used outb, out_8 or similar -- already had a full barrier because out_8 etc. include a sync before the MMIO store. This adds an explicit barrier in the two remaining cases. These changes made no measurable difference to the speed of IPIs as measured using a simple ping-pong latency test across two CPUs on different cores of a POWER7 machine. The analysis of the reason why processes were not waking up properly is due to Milton Miller. Cc: stable@vger.kernel.org # v3.0+ Reported-by:
Milton Miller <miltonm@bga.com> Signed-off-by:
Paul Mackerras <paulus@samba.org> Signed-off-by:
-
Anton Blanchard authored
During a context switch we always restore the per thread DSCR value. If we aren't doing explicit DSCR management (ie thread.dscr_inherit == 0) and the default DSCR changed while the process has been sleeping we end up with the wrong value. Check thread.dscr_inherit and select the default DSCR or per thread DSCR as required. This was found with the following test case, when running with more threads than CPUs (ie forcing context switching): http://ozlabs.org/~anton/junkcode/dscr_default_test.c With the four patches applied I can run a combination of all test cases successfully at the same time: http://ozlabs.org/~anton/junkcode/dscr_default_test.c http://ozlabs.org/~anton/junkcode/dscr_explicit_test.c http://ozlabs.org/~anton/junkcode/dscr_inherit_test.cSigned-off-by:
-
Anton Blanchard authored
If the default DSCR is non zero we set thread.dscr_inherit in copy_thread() meaning the new thread and all its children will ignore future updates to the default DSCR. This is not intended and is a change in behaviour that a number of our users have hit. We just need to inherit thread.dscr and thread.dscr_inherit from the parent which ends up being much simpler. This was found with the following test case: http://ozlabs.org/~anton/junkcode/dscr_default_test.cSigned-off-by:
-
Anton Blanchard authored
When we update the DSCR either via emulation of mtspr(DSCR) or via a change to dscr_default in sysfs we don't update thread.dscr. We will eventually update it at context switch time but there is a period where thread.dscr is incorrect. If we fork at this point we will copy the old value of thread.dscr into the child. To avoid this, always keep thread.dscr in sync with reality. This issue was found with the following testcase: http://ozlabs.org/~anton/junkcode/dscr_inherit_test.cSigned-off-by:
-
Anton Blanchard authored
Writing to dscr_default in sysfs doesn't actually change the DSCR - we rely on a context switch on each CPU to do the work. There is no guarantee we will get a context switch in a reasonable amount of time so fire off an IPI to force an immediate change. This issue was found with the following test case: http://ozlabs.org/~anton/junkcode/dscr_explicit_test.cSigned-off-by:
-
Paul Mackerras authored
The CPU hotplug code for the powernv platform currently only puts offline CPUs into nap mode if the powersave_nap variable is set. However, HV-style KVM on this platform requires secondary CPU threads to be offline and in nap mode. Since we know nap mode works just fine on all POWER7 machines, and the only machines that support the powernv platform are POWER7 machines, this changes the code to always put offline CPUs into nap mode, regardless of powersave_nap. Powersave_nap still controls whether or not CPUs go into nap mode when idle, as before. Signed-off-by:
-
Paul Mackerras authored
At the moment the handler for hypervisor decrementer interrupts is the same as for decrementer interrupts, i.e. timer_interrupt(). This is bogus; if we ever do get a hypervisor decrementer interrupt it won't have anything to do with the next timer event. In fact the only time we get hypervisor decrementer interrupts is when one is left pending on exit from a KVM guest. When we get a hypervisor decrementer interrupt we don't need to do anything special to clear it, since they are edge-triggered on the transition of HDEC from 0 to -1. Thus this adds an empty handler function for them. We don't need to have them masked when interrupts are soft-disabled, so we use STD_EXCEPTION_HV instead of MASKABLE_EXCEPTION_HV. Signed-off-by:
-
Jesse Larrew authored
arch_update_cpu_topology() should only return 1 when the topology has actually changed, and should return 0 otherwise. This patch fixes a potential bug where rebuild_sched_domains() would reinitialize the sched domains even when the topology hasn't changed. Signed-off-by:
Jesse Larrew <jlarrew@linux.vnet.ibm.com> Signed-off-by:
-
Mihai Caraman authored
Embedded.Hypervisor category defines GSPRG0..3 physical registers for guests. Avoid SPRG4-7 usage as scratch in host exception handlers, otherwise guest SPRG4-7 registers will be clobbered. For bolted TLB miss exception handlers, which is the version currently supported by KVM, use SPRN_SPRG_GEN_SCRATCH aka SPRG0 instead of SPRN_SPRG_TLB_SCRATCH aka SPRG6. Keep using TLB PACA slots to fit in one 64-byte cache line. For critical exception handlers use SPRG3 instead of SPRG7. Provide a routine to store and restore user-visible SPRGs. This will be subsequently used to restore VDSO information in SPRG3. Add EX_R13 to paca slots to free up SPRG3 and change the critical exception epilog to use it. Signed-off-by:
-
Mihai Caraman authored
Refactor exception prolog to get rid of mfspr srr1 duplicate. This was introduced by KVM integration, with DO_KVM macro logic expecting srr1 value earlier in r11. Reserve r11 to hold srr1's value also required at the end of the prolog and free up r10 to serve as spare in addition macros. For syscalls case this change does not add any performance penalty. For irq soft-disabled case the change adds a store/load of conditional register value to/from a paca slot. Paca slots fit in one 64-byte cache line so these additional operations have little impact on performance. Signed-off-by:
-
Mihai Caraman authored
Hook DO_KVM macro into 64-bit booke for KVM integration. Extend interrupt handlers' parameter list with interrupt vector numbers to accomodate the macro. Only the bolted version of tlb miss handers is addressed now. Signed-off-by:
-
Mihai Caraman authored
Guest Doorbell interrupts use guest save and restore registers. Add a new Guest Doorbell exception type to accommodate GSRR0/1 SPRs usage in exception prolog and fix the exception handler. Signed-off-by:
-
Mihai Caraman authored
Machine check exception handler was using a wrong prolog. Hypervisors like KVM which are called early from the exception handler rely on the interrupt source. Signed-off-by:
-
Ananth N Mavinakayanahalli authored
This is the port of uprobes to powerpc. Usage is similar to x86. [root@xxxx ~]# ./bin/perf probe -x /lib64/libc.so.6 malloc Added new event: probe_libc:malloc (on 0xb4860) You can now use it in all perf tools, such as: perf record -e probe_libc:malloc -aR sleep 1 [root@xxxx ~]# ./bin/perf record -e probe_libc:malloc -aR sleep 20 [ perf record: Woken up 22 times to write data ] [ perf record: Captured and wrote 5.843 MB perf.data (~255302 samples) ] [root@xxxx ~]# ./bin/perf report --stdio ... 69.05% tar libc-2.12.so [.] malloc 28.57% rm libc-2.12.so [.] malloc 1.32% avahi-daemon libc-2.12.so [.] malloc 0.58% bash libc-2.12.so [.] malloc 0.28% sshd libc-2.12.so [.] malloc 0.08% irqbalance libc-2.12.so [.] malloc 0.05% bzip2 libc-2.12.so [.] malloc 0.04% sleep libc-2.12.so [.] malloc 0.03% multipathd libc-2.12.so [.] malloc 0.01% sendmail libc-2.12.so [.] malloc 0.01% automount libc-2.12.so [.] malloc The trap_nr addition patch is a prereq. Signed-off-by: -
Ananth N Mavinakayanahalli authored
Add thread_struct.trap_nr and use it to store the last exception the thread experienced. In this patch, we populate the field at various places where we force_sig_info() to the process. This is also used in uprobes to determine if the probed instruction caused an exception. Signed-off-by:
-
Ananth N Mavinakayanahalli authored
Move is_trap() and relatives to a common file to be shared between kprobes and uprobes. Code movement only; no change in functionality. Suggested by Michael Ellerman. Signed-off-by:
-
Michael Ellerman authored
We have an old FIXME in reg.h which points out that we should standardise on PVR_foo for our PVR #defines. Currently we use PVR_ on 32-bit and PV_ on 64-bit. So do that rename and remove the FIXME. Seeing as we're touching all but one usage of __is_processor(), rename it to something less ugly and more indicative of what it does, which is simply to check the PVR version. Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Benjamin Herrenschmidt authored
The pseries hvterm driver only registers a udbg backend (for xmon and other low level debugging mechanisms) when hvc0 is recognized as the firmware console at boot time, not if it's detected later on, for example because the firmware is using a graphics card. This can make debugging challenging especially under X11, and there's really no good reason for that limitation, so let's hookup udbg whenever hvc0 is detected instead. Signed-off-by: -
Benjamin Herrenschmidt authored
hvc_console has two methods to instanciate the consoles. hvc_instanciate is meant to be called at early boot, while hvc_alloc is called for more dynamically probed objects. Currently, it only deals with adding kernel consoles in the former case, which means for example that if a console only uses dynamic probing, it will never be usable as a kernel console even when specifying console=hvc0 explicitly, which could be considered annoying... More specifically, on pseries, we only do the early instanciate for the console currently used by the firmware, so if you have your firmware configured to go to a video card, for example, you cannot get your kernel console, oops messages, etc... on your serial port or hypervisor console, which would be handy to deal with oopses. This fixes it by checking if hvc_console.flags & CON_ENABLED is set when registering a new dynamic console, and if not, redo the index check and re-register the console if the index matches, allowing console=hvcN to work. Signed-off-by: -
Michael Ellerman authored
It contains no code and is not included by anyone. Signed-off-by:
-
Michael Ellerman authored
It's empty now, apart from other includes. Fixup a few files that were getting things via this header. Signed-off-by:
-
Michael Ellerman authored
Signed-off-by:
-
Michael Ellerman authored
Signed-off-by:
-
Michael Ellerman authored
Signed-off-by:
-
Michael Ellerman authored
virt_to_abs() is just a wrapper around __pa(), call __pa() directly. Signed-off-by:
-
Michael Ellerman authored
These days they are just __va() and __pa() respectively. Signed-off-by:
-
Michael Ellerman authored
These days they are just __va() and __pa() respectively. Signed-off-by:
-
Michael Ellerman authored
These days they are just __va() and __pa() respectively. Signed-off-by:
-
Michael Ellerman authored
abs_to_virt() simply calls __va() and we'd like to get rid of it, so replace all abs_to_virt() uses with __va(). Similarly virt_to_abs() just calls __pa(). Signed-off-by:
-
Michael Ellerman authored
abs_to_virt() is just a wrapper around __va(), call __va() directly. Signed-off-by:
-
Michael Ellerman authored
abs_to_virt() simply calls __va() and we'd like to get rid of it, so replace all abs_to_virt() uses with __va(). __va() returns void *, so when assigning to a pointer there's no need to cast. Similarly virt_to_abs() just calls __pa(). Signed-off-by:
-
Michael Ellerman authored
These days they are just wrappers around __pa() and __va() respectively. Signed-off-by:
-
Michael Ellerman authored
These days they are just wrappers around __pa() and __va() respectively. Signed-off-by:
-
Michael Ellerman authored
abs_to_virt() is just a wrapper around __va(), call __va() directly. Signed-off-by:
-
