[Bug tdep/29814] [gdb/tdep, powerpc64le] FAIL: gdb.base/msym-bp-shl.exp: debug=0: before run: info breakpoint

["vries at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=29814

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
This is a point fix:
...
diff --git a/gdb/ppc-linux-tdep.c b/gdb/ppc-linux-tdep.c
index cc5a26431ba..a7ee72e24a5 100644
--- a/gdb/ppc-linux-tdep.c
+++ b/gdb/ppc-linux-tdep.c
@@ -1667,9 +1667,11 @@ ppc_elfv2_skip_entrypoint (struct gdbarch *gdbarch,
CORE_ADDR pc)
   if (fun.minsym->target_flag_1 ())
     local_entry_offset = 8;

-  if (fun.value_address () <= pc
-      && pc < fun.value_address () + local_entry_offset)
-    return fun.value_address () + local_entry_offset;
+  struct obj_section *section = fun.obj_section ();
+  CORE_ADDR new_pc = fun.value_address () + local_entry_offset;
+  if (fun.value_address () <= pc && pc < new_pc
+      && (section == nullptr || new_pc < section->endaddr ()))
+    return new_pc;

   return pc;
 }
...

It does fix the test-case, but not the general problem, since this will work
for one or maybe two plt symbols, but if there are more, say foo1@plt, foo2@plt
and foo3@plt, then setting a breakpoint at foo1@plt will end up setting a
breakpoint at foo3@plt.

-- 
You are receiving this mail because:
You are on the CC list for the bug.