| Message ID | 20220223222002.1085114-1-haoluo@google.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | BPF |
| Headers | show |
| Series | [bpf-next] bpf: Cache the last valid build_id. | expand |
On Wed, Feb 23, 2022 at 2:20 PM Hao Luo <haoluo@google.com> wrote: > > For binaries that are statically linked, consecutive stack frames are > likely to be in the same VMA and therefore have the same build id. > As an optimization for this case, we can cache the previous frame's > VMA, if the new frame has the same VMA as the previous one, reuse the > previous one's build id. We are holding the MM locks as reader across > the entire loop, so we don't need to worry about VMA going away. > > Tested through "stacktrace_build_id" and "stacktrace_build_id_nmi" in > test_progs. > > Suggested-by: Greg Thelen <gthelen@google.com> > Signed-off-by: Hao Luo <haoluo@google.com> > --- > kernel/bpf/stackmap.c | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c > index 22c8ae94e4c1..280b9198af27 100644 > --- a/kernel/bpf/stackmap.c > +++ b/kernel/bpf/stackmap.c > @@ -132,7 +132,8 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, > int i; > struct mmap_unlock_irq_work *work = NULL; > bool irq_work_busy = bpf_mmap_unlock_get_irq_work(&work); > - struct vm_area_struct *vma; > + struct vm_area_struct *vma, *prev_vma = NULL; > + const char *prev_build_id; > > /* If the irq_work is in use, fall back to report ips. Same > * fallback is used for kernel stack (!user) on a stackmap with > @@ -151,6 +152,11 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, > > for (i = 0; i < trace_nr; i++) { > vma = find_vma(current->mm, ips[i]); as a further optimization, shouldn't we first check if ips[i] is within prev_vma and avoid rbtree walk altogether? Would this work: if (prev_vma && range_in_vma(prev_vma, ips[i])) { /* reuse build_id */ } vma = find_vma(current->mm, ips[i]); ? > + if (vma && vma == prev_vma) { > + memcpy(id_offs[i].build_id, prev_build_id, > + BUILD_ID_SIZE_MAX); > + goto build_id_valid; > + } > if (!vma || build_id_parse(vma, id_offs[i].build_id, NULL)) { > /* per entry fall back to ips */ > id_offs[i].status = BPF_STACK_BUILD_ID_IP; > @@ -158,9 +164,12 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, > memset(id_offs[i].build_id, 0, BUILD_ID_SIZE_MAX); > continue; > } > +build_id_valid: > id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i] > - vma->vm_start; > id_offs[i].status = BPF_STACK_BUILD_ID_VALID; > + prev_vma = vma; > + prev_build_id = id_offs[i].build_id; > } > bpf_mmap_unlock_mm(work, current->mm); > } > -- > 2.35.1.473.g83b2b277ed-goog >
Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > On Wed, Feb 23, 2022 at 2:20 PM Hao Luo <haoluo@google.com> wrote: >> >> For binaries that are statically linked, consecutive stack frames are >> likely to be in the same VMA and therefore have the same build id. >> As an optimization for this case, we can cache the previous frame's >> VMA, if the new frame has the same VMA as the previous one, reuse the >> previous one's build id. We are holding the MM locks as reader across >> the entire loop, so we don't need to worry about VMA going away. >> >> Tested through "stacktrace_build_id" and "stacktrace_build_id_nmi" in >> test_progs. >> >> Suggested-by: Greg Thelen <gthelen@google.com> >> Signed-off-by: Hao Luo <haoluo@google.com> >> --- >> kernel/bpf/stackmap.c | 11 ++++++++++- >> 1 file changed, 10 insertions(+), 1 deletion(-) >> >> diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c >> index 22c8ae94e4c1..280b9198af27 100644 >> --- a/kernel/bpf/stackmap.c >> +++ b/kernel/bpf/stackmap.c >> @@ -132,7 +132,8 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, >> int i; >> struct mmap_unlock_irq_work *work = NULL; >> bool irq_work_busy = bpf_mmap_unlock_get_irq_work(&work); >> - struct vm_area_struct *vma; >> + struct vm_area_struct *vma, *prev_vma = NULL; >> + const char *prev_build_id; >> >> /* If the irq_work is in use, fall back to report ips. Same >> * fallback is used for kernel stack (!user) on a stackmap with >> @@ -151,6 +152,11 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, >> >> for (i = 0; i < trace_nr; i++) { >> vma = find_vma(current->mm, ips[i]); > > as a further optimization, shouldn't we first check if ips[i] is > within prev_vma and avoid rbtree walk altogether? Would this work: > > if (prev_vma && range_in_vma(prev_vma, ips[i])) { > /* reuse build_id */ > } > vma = find_vma(current->mm, ips[i]); > > > ? Yes, that's a nice addition. Good idea. >> + if (vma && vma == prev_vma) { >> + memcpy(id_offs[i].build_id, prev_build_id, >> + BUILD_ID_SIZE_MAX); >> + goto build_id_valid; >> + } >> if (!vma || build_id_parse(vma, id_offs[i].build_id, NULL)) { >> /* per entry fall back to ips */ >> id_offs[i].status = BPF_STACK_BUILD_ID_IP; >> @@ -158,9 +164,12 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, >> memset(id_offs[i].build_id, 0, BUILD_ID_SIZE_MAX); >> continue; >> } >> +build_id_valid: >> id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i] >> - vma->vm_start; >> id_offs[i].status = BPF_STACK_BUILD_ID_VALID; >> + prev_vma = vma; >> + prev_build_id = id_offs[i].build_id; >> } >> bpf_mmap_unlock_mm(work, current->mm); >> } >> -- >> 2.35.1.473.g83b2b277ed-goog >>
On Wed, Feb 23, 2022 at 3:16 PM Greg Thelen <gthelen@google.com> wrote: > > Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > > > > > as a further optimization, shouldn't we first check if ips[i] is > > within prev_vma and avoid rbtree walk altogether? Would this work: > > > > if (prev_vma && range_in_vma(prev_vma, ips[i])) { > > /* reuse build_id */ > > } > > vma = find_vma(current->mm, ips[i]); > > > > > > ? > > Yes, that's a nice addition. Good idea. Yes, great point! I noticed range_in_vma() already has a check on the null-ness of prev_vma. I am going to send a v2.
diff --git a/kernel/bpf/stackmap.c b/kernel/bpf/stackmap.c index 22c8ae94e4c1..280b9198af27 100644 --- a/kernel/bpf/stackmap.c +++ b/kernel/bpf/stackmap.c @@ -132,7 +132,8 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, int i; struct mmap_unlock_irq_work *work = NULL; bool irq_work_busy = bpf_mmap_unlock_get_irq_work(&work); - struct vm_area_struct *vma; + struct vm_area_struct *vma, *prev_vma = NULL; + const char *prev_build_id; /* If the irq_work is in use, fall back to report ips. Same * fallback is used for kernel stack (!user) on a stackmap with @@ -151,6 +152,11 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, for (i = 0; i < trace_nr; i++) { vma = find_vma(current->mm, ips[i]); + if (vma && vma == prev_vma) { + memcpy(id_offs[i].build_id, prev_build_id, + BUILD_ID_SIZE_MAX); + goto build_id_valid; + } if (!vma || build_id_parse(vma, id_offs[i].build_id, NULL)) { /* per entry fall back to ips */ id_offs[i].status = BPF_STACK_BUILD_ID_IP; @@ -158,9 +164,12 @@ static void stack_map_get_build_id_offset(struct bpf_stack_build_id *id_offs, memset(id_offs[i].build_id, 0, BUILD_ID_SIZE_MAX); continue; } +build_id_valid: id_offs[i].offset = (vma->vm_pgoff << PAGE_SHIFT) + ips[i] - vma->vm_start; id_offs[i].status = BPF_STACK_BUILD_ID_VALID; + prev_vma = vma; + prev_build_id = id_offs[i].build_id; } bpf_mmap_unlock_mm(work, current->mm); }
For binaries that are statically linked, consecutive stack frames are likely to be in the same VMA and therefore have the same build id. As an optimization for this case, we can cache the previous frame's VMA, if the new frame has the same VMA as the previous one, reuse the previous one's build id. We are holding the MM locks as reader across the entire loop, so we don't need to worry about VMA going away. Tested through "stacktrace_build_id" and "stacktrace_build_id_nmi" in test_progs. Suggested-by: Greg Thelen <gthelen@google.com> Signed-off-by: Hao Luo <haoluo@google.com> --- kernel/bpf/stackmap.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)