From 71eb9ee9596d8df3d5723c3cfc18774c6235e8b1 Mon Sep 17 00:00:00 2001 From: Stephane Eranian Date: Fri, 23 Mar 2018 00:01:47 -0700 Subject: perf/x86/intel: Fix linear IP of PEBS real_ip on Haswell and later CPUs this patch fix a bug in how the pebs->real_ip is handled in the PEBS handler. real_ip only exists in Haswell and later processor. It is actually the eventing IP, i.e., where the event occurred. As opposed to the pebs->ip which is the PEBS interrupt IP which is always off by one. The problem is that the real_ip just like the IP needs to be fixed up because PEBS does not record all the machine state registers, and in particular the code segement (cs). This is why we have the set_linear_ip() function. The problem was that set_linear_ip() was only used on the pebs->ip and not the pebs->real_ip. We have profiles which ran into invalid callstacks because of this. Here is an example: ..... 0: ffffffffffffff80 recent entry, marker kernel v ..... 1: 000000000040044d <= user address in kernel space! ..... 2: fffffffffffffe00 marker enter user v ..... 3: 000000000040044d ..... 4: 00000000004004b6 oldest entry Debugging output in get_perf_callchain(): [ 857.769909] CALLCHAIN: CPU8 ip=40044d regs->cs=10 user_mode(regs)=0 The problem is that the kernel entry in 1: points to a user level address. How can that be? The reason is that with PEBS sampling the instruction that caused the event to occur and the instruction where the CPU was when the interrupt was posted may be far apart. And sometime during that time window, the privilege level may change. This happens, for instance, when the PEBS sample is taken close to a kernel entry point. Here PEBS, eventing IP (real_ip) captured a user level instruction. But by the time the PMU interrupt fired, the processor had already entered kernel space. This is why the debug output shows a user address with user_mode() false. The problem comes from PEBS not recording the code segment (cs) register. The register is used in x86_64 to determine if executing in kernel vs user space. This is okay because the kernel has a software workaround called set_linear_ip(). But the issue in setup_pebs_sample_data() is that set_linear_ip() is never called on the real_ip value when it is available (Haswell and later) and precise_ip > 1. This patch fixes this problem and eliminates the callchain discrepancy. The patch restructures the code around set_linear_ip() to minimize the number of times the IP has to be set. Signed-off-by: Stephane Eranian Cc: Alexander Shishkin Cc: Arnaldo Carvalho de Melo Cc: Jiri Olsa Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Vince Weaver Cc: kan.liang@intel.com Link: http://lkml.kernel.org/r/1521788507-10231-1-git-send-email-eranian@google.com Signed-off-by: Ingo Molnar --- arch/x86/events/intel/ds.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c index d8015235ba76..5e526c54247e 100644 --- a/arch/x86/events/intel/ds.c +++ b/arch/x86/events/intel/ds.c @@ -1153,6 +1153,7 @@ static void setup_pebs_sample_data(struct perf_event *event, if (pebs == NULL) return; + regs->flags &= ~PERF_EFLAGS_EXACT; sample_type = event->attr.sample_type; dsrc = sample_type & PERF_SAMPLE_DATA_SRC; @@ -1197,7 +1198,6 @@ static void setup_pebs_sample_data(struct perf_event *event, */ *regs = *iregs; regs->flags = pebs->flags; - set_linear_ip(regs, pebs->ip); if (sample_type & PERF_SAMPLE_REGS_INTR) { regs->ax = pebs->ax; @@ -1233,13 +1233,22 @@ static void setup_pebs_sample_data(struct perf_event *event, #endif } - if (event->attr.precise_ip > 1 && x86_pmu.intel_cap.pebs_format >= 2) { - regs->ip = pebs->real_ip; - regs->flags |= PERF_EFLAGS_EXACT; - } else if (event->attr.precise_ip > 1 && intel_pmu_pebs_fixup_ip(regs)) - regs->flags |= PERF_EFLAGS_EXACT; - else - regs->flags &= ~PERF_EFLAGS_EXACT; + if (event->attr.precise_ip > 1) { + /* Haswell and later have the eventing IP, so use it: */ + if (x86_pmu.intel_cap.pebs_format >= 2) { + set_linear_ip(regs, pebs->real_ip); + regs->flags |= PERF_EFLAGS_EXACT; + } else { + /* Otherwise use PEBS off-by-1 IP: */ + set_linear_ip(regs, pebs->ip); + + /* ... and try to fix it up using the LBR entries: */ + if (intel_pmu_pebs_fixup_ip(regs)) + regs->flags |= PERF_EFLAGS_EXACT; + } + } else + set_linear_ip(regs, pebs->ip); + if ((sample_type & (PERF_SAMPLE_ADDR | PERF_SAMPLE_PHYS_ADDR)) && x86_pmu.intel_cap.pebs_format >= 1) -- cgit v1.2.3-70-g09d2 From f67b15037a7a50c57f72e69a6d59941ad90a0f0f Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 26 Mar 2018 15:39:07 -1000 Subject: perf/hwbp: Simplify the perf-hwbp code, fix documentation Annoyingly, modify_user_hw_breakpoint() unnecessarily complicates the modification of a breakpoint - simplify it and remove the pointless local variables. Also update the stale Docbook while at it. Signed-off-by: Linus Torvalds Acked-by: Thomas Gleixner Cc: Cc: Alexander Shishkin Cc: Andy Lutomirski Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Cc: Jiri Olsa Cc: Peter Zijlstra Cc: Stephane Eranian Cc: Vince Weaver Signed-off-by: Ingo Molnar --- kernel/events/hw_breakpoint.c | 30 +++++++----------------------- 1 file changed, 7 insertions(+), 23 deletions(-) diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c index 3f8cb1e14588..253ae2da13c3 100644 --- a/kernel/events/hw_breakpoint.c +++ b/kernel/events/hw_breakpoint.c @@ -427,16 +427,9 @@ EXPORT_SYMBOL_GPL(register_user_hw_breakpoint); * modify_user_hw_breakpoint - modify a user-space hardware breakpoint * @bp: the breakpoint structure to modify * @attr: new breakpoint attributes - * @triggered: callback to trigger when we hit the breakpoint - * @tsk: pointer to 'task_struct' of the process to which the address belongs */ int modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *attr) { - u64 old_addr = bp->attr.bp_addr; - u64 old_len = bp->attr.bp_len; - int old_type = bp->attr.bp_type; - int err = 0; - /* * modify_user_hw_breakpoint can be invoked with IRQs disabled and hence it * will not be possible to raise IPIs that invoke __perf_event_disable. @@ -451,27 +444,18 @@ int modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *att bp->attr.bp_addr = attr->bp_addr; bp->attr.bp_type = attr->bp_type; bp->attr.bp_len = attr->bp_len; + bp->attr.disabled = 1; - if (attr->disabled) - goto end; - - err = validate_hw_breakpoint(bp); - if (!err) - perf_event_enable(bp); + if (!attr->disabled) { + int err = validate_hw_breakpoint(bp); - if (err) { - bp->attr.bp_addr = old_addr; - bp->attr.bp_type = old_type; - bp->attr.bp_len = old_len; - if (!bp->attr.disabled) - perf_event_enable(bp); + if (err) + return err; - return err; + perf_event_enable(bp); + bp->attr.disabled = 0; } -end: - bp->attr.disabled = attr->disabled; - return 0; } EXPORT_SYMBOL_GPL(modify_user_hw_breakpoint); -- cgit v1.2.3-70-g09d2