Doug,
please try this patch also.
пт, 7 сент. 2018 г. в 11:18, Doug Moore <do...@rice.edu>:
> Sergey,
>
> Your patch does remove memory allocation from the file, but oddly, I
> still have problems with malloc being invoked. The second time I
> tested with your patch, I got hung here:
>
> #0 0x0000ffffb5a85d74 in __lll_lock_wait_private () from /lib64/libc.so.6
> #1 0x0000ffffb5a0ea38 in malloc () from /lib64/libc.so.6
> #2 0x0000ffffb59c15bc in qsort_r () from /lib64/libc.so.6
> #3 0x0000ffffb592c8b0 in _ULaarch64_dwarf_find_debug_frame (found=0,
> di_debug=di_debug@entry=0xffffeb09a080, ip=ip@entry=281473734119984,
> segbase=segbase@entry=281473734053888, obj_name=0xffffb67dc440
> "/home/dougm/.local/lib/hpctoolkit/ext-libs/libdwarf.so.1",
> start=<optimized out>,
> end=<optimized out>) at dwarf/Gfind_proc_info-lsb.c:380
> #4 0x0000ffffb592cab0 in _ULaarch64_dwarf_callback
> (info=0xffffeb099ef8, size=<optimized out>, ptr=0xffffeb09a018) at
> dwarf/Gfind_proc_info-lsb.c:667
> #5 0x0000ffffb5abac88 in dl_iterate_phdr () from /lib64/libc.so.6
> #6 0x0000ffffb592cff4 in _ULaarch64_dwarf_find_proc_info
> (as=0xffffb5948230 <local_addr_space>, ip=ip@entry=281473734119984,
> pi=pi@entry=0xffffeb09a530,
> need_unwind_info=1, arg=0xffffeb09a9a0) at
> dwarf/Gfind_proc_info-lsb.c:693
> #7 0x0000ffffb592a158 in fetch_proc_info (c=c@entry=0xffffeb09a1d0,
> ip=281473734119984) at dwarf/Gparser.c:454
> #8 0x0000ffffb592ba08 in _ULaarch64_dwarf_reg_states_iterate
> (c=0xffffeb09a1d0, cb=0xffffb6750844 <dwarf_reg_states_callback>,
> token=0xffffeb09a1b8)
> at dwarf/Gparser.c:1034
>
> So, perhaps qsort isn't safe either, as odd as that seems.
>
> Doug
>
> Quoting Sergey Korolev <korolev.ser...@gmail.com>:
>
> > Doug,
> >
> > Can you try this patch?
> >
> http://lists.nongnu.org/archive/html/libunwind-devel/2018-06/msg00005.html
> >
> > пт, 7 сент. 2018 г. в 1:56, Doug Moore <do...@rice.edu>:
> >
> >> It seems that unw_get_proc_info calls dwarf_make_proc_info, calls
> >> fetch_proc_info, calls dwarf_find_proc_info, calls dl_iterate_phdr,
> >> calls dwarf_callback, calls dwarf_find_debug_frame, calls calloc, and
> >> callloc is not signal safe on the aarch64 device I'm testing on.
> >>
> >> So would there be a problem with replacing the calloc/realloc memory
> >> management of 'tab' in Gfind_proc_info-lsb.c with mmap and munmap and
> >> memcpy?
> >>
> >> Is the memory allocated for 'tab' freed anywhere? It's not obvious that
> >> it is.
> >>
> >> Thanks,
> >>
> >> Doug Moore
> >>
> >> Rice University
> >>
> >>
> >> _______________________________________________
> >> Libunwind-devel mailing list
> >> Libunwind-devel@nongnu.org
> >> https://lists.nongnu.org/mailman/listinfo/libunwind-devel
> >>
>
>
>
>
From e6384fc17f6f18fa605b1aba03edb6268c838a45 Mon Sep 17 00:00:00 2001
From: Sergey Korolev <s.koro...@ndmsystems.com>
Date: Fri, 7 Sep 2018 13:34:57 +0300
Subject: [PATCH] dwarf: do not use qsort to sort the debug frame table
Some qsort implementations (e.g. glibc) may call memory
allocation functions that is not async-signal safe.
---
src/dwarf/Gfind_proc_info-lsb.c | 45 +++++++++++++++++++++++++++--------------
1 file changed, 30 insertions(+), 15 deletions(-)
diff --git a/src/dwarf/Gfind_proc_info-lsb.c b/src/dwarf/Gfind_proc_info-lsb.c
index 18819e6a..a186f870 100644
--- a/src/dwarf/Gfind_proc_info-lsb.c
+++ b/src/dwarf/Gfind_proc_info-lsb.c
@@ -223,19 +223,6 @@ locate_debug_info (unw_addr_space_t as, unw_word_t addr, const char *dlname,
return fdesc;
}
-static int
-debug_frame_index_compare (const void *a, const void *b)
-{
- const struct table_entry *fa = a, *fb = b;
-
- if (fa->start_ip_offset > fb->start_ip_offset)
- return 1;
- else if (fa->start_ip_offset < fb->start_ip_offset)
- return -1;
- else
- return 0;
-}
-
static size_t
debug_frame_index_make (struct unw_debug_frame_list *fdesc)
{
@@ -323,6 +310,35 @@ debug_frame_index_make (struct unw_debug_frame_list *fdesc)
return count;
}
+static void
+debug_frame_index_sort (struct unw_debug_frame_list *fdesc)
+{
+ size_t i, j, k, n = fdesc->index_size / sizeof (*fdesc->index);
+ struct table_entry *a = fdesc->index;
+ struct table_entry t;
+
+ /* Use a simple Shell sort as it relatively fast and
+ * does not require additional memory. */
+
+ for (k = n / 2; k > 0; k /= 2)
+ {
+ for (i = k; i < n; i++)
+ {
+ t = a[i];
+
+ for (j = i; j >= k; j -= k)
+ {
+ if (t.start_ip_offset >= a[j - k].start_ip_offset)
+ break;
+
+ a[j] = a[j - k];
+ }
+
+ a[j] = t;
+ }
+ }
+}
+
int
dwarf_find_debug_frame (int found, unw_dyn_info_t *di_debug, unw_word_t ip,
unw_word_t segbase, const char* obj_name,
@@ -377,8 +393,7 @@ dwarf_find_debug_frame (int found, unw_dyn_info_t *di_debug, unw_word_t ip,
/* Then fill and sort the index. */
debug_frame_index_make (fdesc);
- qsort (fdesc->index, count, sizeof (struct table_entry),
- debug_frame_index_compare);
+ debug_frame_index_sort (fdesc);
/*for (i = 0; i < count; i++)
{
--
2.11.0
_______________________________________________
Libunwind-devel mailing list
Libunwind-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/libunwind-devel
