On Tue, Nov 24, 2009 at 12:12 PM, Arun Sharma <[email protected]> wrote:
> You're right. This part is broken for dwarf. We'll need to fix it along the
> lines of src/ia64/Gscript.c
Tricky (unless we drastically reduce cache size): current
sizeof(local_addr_space) on x86_64 is 43664, and adding this much
TLS to each thread may run into glibc issue: glibc does not
automatically add required TLS space at pthread_create, so if you
specify e.g. 64K stack, pthread_create will fail with EINVAL just
because you linked to libunwind.
> Agree that fix#1 is suboptimal. The choice is really between fix #2 and a
> fixed UNW_CACHE_PER_THREAD.
Hmm, I don't think that's the choice: libunwind *should* work with either!
Attached is a fix#2A, which removes the race for UNW_CACHE_GLOBAL,
but doesn't penalize UNW_CACHE_PER_THREAD (except for a single branch).
Thanks,
--
Paul Pluzhnikov
diff --git a/src/dwarf/Gparser.c b/src/dwarf/Gparser.c
index b77bde2..2c082d5 100644
--- a/src/dwarf/Gparser.c
+++ b/src/dwarf/Gparser.c
@@ -784,12 +784,13 @@ HIDDEN int
dwarf_find_save_locs (struct dwarf_cursor *c)
{
dwarf_state_record_t sr;
- dwarf_reg_state_t *rs;
+ dwarf_reg_state_t *rs, rs_copy;
struct dwarf_rs_cache *cache;
int ret = 0;
intrmask_t saved_mask;
+ const unw_caching_policy_t caching_policy = c->as->caching_policy;
- if (c->as->caching_policy == UNW_CACHE_NONE)
+ if (caching_policy == UNW_CACHE_NONE)
return uncached_dwarf_find_save_locs (c);
cache = get_rs_cache(c->as, &saved_mask);
@@ -816,6 +817,12 @@ dwarf_find_save_locs (struct dwarf_cursor *c)
put_unwind_info (c, &c->pi);
}
+ if (likely (caching_policy == UNW_CACHE_GLOBAL))
+ {
+ memcpy (&rs_copy, rs, sizeof (rs_copy));
+ rs = &rs_copy;
+ }
+
put_rs_cache (c->as, cache, &saved_mask);
if ((ret = apply_reg_state (c, rs)) < 0)
return ret;
_______________________________________________
Libunwind-devel mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/libunwind-devel
