[Bug runtime/16096] New: on i686, the nd_syscall tapset is broken

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

https://sourceware.org/bugzilla/show_bug.cgi?id=16096

            Bug ID: 16096
           Summary: on i686, the nd_syscall tapset is broken
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
          Assignee: systemtap at sourceware dot org
          Reporter: dsmith at redhat dot com

Something is broken in the nd_syscall tapset on i686. Testing the nd_syscall
tapset (using nd_syscall.exp) reveals the following:

os     kernel                    passes  failures
====   ========================  ======  ========
rhel6  2.6.32-358.18.1.el6.i686      34         1
f15    2.6.43.8-1.fc15.i686.PAE      34         1
f16    3.6.11-4.fc16.i686.PAE        34         1
f17    3.9.10-100.fc17.i686.PAE       0        35
f18    3.11.4-101.fc18.i686.PAE       0        35
f19    3.11.6-200.fc19.i686.PAE       0        35

So, something happened between 3.6.11-4.fc16.i686.PAE and
3.9.10-100.fc17.i686.PAE. The errors are all similar, and look like the
following (from the nd_syscall.exp access subtest):

====
ERROR: user string copy fault -14 at 0000000b near identifier 'user_string_n'
at
 /usr/local/share/systemtap/tapset/uconversions.stp:120:10
====

The problem here is that we're treating 0xb as an address, where in this case I
believe the 11 (0xb) is really the syscall number of mmap().

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