[Bug general/3926] New: More detailed stack frame testing.

["npremji at redhat dot com" <sourceware-bugzilla at sourceware dot org>]

We need to test more cases of different stack frame scenarios.

Currently we have a simple c function compiled with debug info, stripped of
debug info, and stripped of all symbols.

Other cases that have been mentioned are:

signal framess.
finished but unjoined threads.
Some combinations of:
 * unprelinked (0x0 basde) / prelinked (only the base binary)
 * before .debug separation / after .debug separation
All the combinations of:
 * -fno-asynchronous-unwind-tables vs. -fasynchronous-unwind-tables
 * without/with -fpie -pie (or use -fpic?)
 * -O0 / -O2
 * -shared / -shared -g / -g
Address cases:
 * Nested (==overlapping) functions (head, nested part, tail)
   (=>outer, inner, outer)
 * Address where the last symbol before is zero-sized and in different section
   (=>unresolved)
 * Address where the last symbol before is non-zero-sized and smaller than the
   distance (=>unresolved)
 * Global function, local function, global function - all three addresses
Unwinding - backtraces:
 * Signal handler unwinds to the first / last function instruction
 * Signal handler unwinds to zero address / zero descriptor
 * Corrupted stack - recovery, corrupted by odd entries number

-- 
           Summary: More detailed stack frame testing.
           Product: frysk
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: general
        AssignedTo: frysk-bugzilla at sourceware dot org
        ReportedBy: npremji at redhat dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=3926

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.