headers
fche at redhat dot com | 1 Nov 2011 20:28
Favicon

[Bug server/13370] New: ipv6 support

http://sourceware.org/bugzilla/show_bug.cgi?id=13370

             Bug #: 13370
           Summary: ipv6 support
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: server
        AssignedTo: systemtap <at> sourceware.org
        ReportedBy: fche <at> redhat.com
    Classification: Unclassified

At present, the stap-server and built-in client widgetry appears to get
confused for dual-stack ipv4+ipv6 configurations, to the point that even
avahi-free --use-server=127.0.0.1:NNNN connections don't get made.

--

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
Permalink | Reply | 1 comment |
headers
fche at redhat dot com | 2 Nov 2011 01:11
Favicon

[Bug translator/13338] possible infinite recursion for process wildcard

http://sourceware.org/bugzilla/show_bug.cgi?id=13338

Frank Ch. Eigler <fche at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Frank Ch. Eigler <fche at redhat dot com> 2011-11-02 00:11:31 UTC ---
commit 5750ecc

--

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
Permalink | Reply | 0 comments |
headers
fche at redhat dot com | 2 Nov 2011 01:49
Favicon

[Bug translator/13371] New: parser doubles-up incoming \ characters in string literals

http://sourceware.org/bugzilla/show_bug.cgi?id=13371

             Bug #: 13371
           Summary: parser doubles-up incoming \ characters in string
                    literals
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: translator
        AssignedTo: systemtap <at> sourceware.org
        ReportedBy: fche <at> redhat.com
    Classification: Unclassified

During early parse, string literals containing \ characters are 
doubled, so that later at translate time, the C compiler will turn them
back into normal \'s.  Since we sometimes process the contents of these
strings within the translator, we need the literals to be unmolested with
this sort of doubling, and instead do that only at code-emission time.

% touch 'foo*bar'
% stap -v -p2 -e 'probe process("foo\\*bar").function("main") {}'
[...]
semantic error: glob foo\\*bar error (3) while resolving probe point
process("foo\\*bar").function("main")

It should have successfully matched the 'foo*bar' program without globbing.
(The \\ on the stap command line was for escaping the shell's parser. 
With only a single \, the shell passes * to systemtap.)
(Continue reading)

Permalink | Reply | 0 comments |
headers
mjw at redhat dot com | 2 Nov 2011 13:37
Favicon

[Bug testsuite/13373] New: FAILs for function.label(string) and new plt probe constructs in unprivileged myproc/probes tests

http://sourceware.org/bugzilla/show_bug.cgi?id=13373

             Bug #: 13373
           Summary: FAILs for function.label(string) and new plt probe
                    constructs in unprivileged myproc/probes tests
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: testsuite
        AssignedTo: systemtap <at> sourceware.org
        ReportedBy: mjw <at> redhat.com
    Classification: Unclassified

The unprivileged_myproc.exp and unprivileged_probes.exp tests are smart. They
detect all possible probe constructs and try to test them all. But that means
that newer constructs like function.label(string) and the new plt probe
constructs all generate FAILs. Currently 38 of them. These need to be added to
these testcases.

Running
/home/mark/src/systemtap/testsuite/systemtap.unprivileged/unprivileged_myproc.exp
...
FAIL: unprivileged myproc: not tested:
process.library(string).function(string).label(string)
FAIL: unprivileged myproc: not tested: process.library(string).plt
FAIL: unprivileged myproc: not tested:
process.library(string).plt.statement(number)
FAIL: unprivileged myproc: not tested: process.library(string).plt(string)
(Continue reading)

Permalink | Reply | 1 comment |
headers
mjw at redhat dot com | 2 Nov 2011 17:05
Favicon

[Bug runtime/13376] New: vdso/vma build-id matching fails on s390x

http://sourceware.org/bugzilla/show_bug.cgi?id=13376

             Bug #: 13376
           Summary: vdso/vma build-id matching fails on s390x
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
        AssignedTo: systemtap <at> sourceware.org
        ReportedBy: mjw <at> redhat.com
    Classification: Unclassified

For some (yet unknown) reason, the vma vdso matching doesn't happen on s390x,
which causes vma_vdso.exp to FAIL.

--

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
Permalink | Reply | 0 comments |
headers
mjw at redhat dot com | 2 Nov 2011 17:08
Favicon

[Bug runtime/13377] New: s390x argument access is broken for any argument > 5 (_stp_arg)

http://sourceware.org/bugzilla/show_bug.cgi?id=13377

             Bug #: 13377
           Summary: s390x argument access is broken for any argument > 5
                    (_stp_arg)
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
        AssignedTo: systemtap <at> sourceware.org
        ReportedBy: mjw <at> redhat.com
    Classification: Unclassified

See _stp_arg in tapset/s390/registers.stp:

 * We don't yet support extracting arg #6 and beyond, which are passed
 * on stack

        if (argnum < 1 || argnum > 5) {
                error(sprintf("Cannot access arg(%d)", argnum))
                return 0
        }

This causes random FAILs for a lot of the syscall.exp test cases, since they do
a wide syscall.* probe with argstr printing which will occasionally hit the
above and ERROR out.

--

--
(Continue reading)

Permalink | Reply | 1 comment |
headers
Wade Farnsworth | 3 Nov 2011 19:06
Favicon

[PATCH 0/2] Userspace tap support for ARM

This is a first pass at userspace tap support for ARM.  Required are UTrace
support as well as Roland McGrath's tracehook and user_regset patches found
here:

https://lkml.org/lkml/2009/4/24/383

Comments are welcome.

-Wade Farnsworth

Wade Farnsworth (2):
  Add syscall information for ARM
  ARM uprobes support

 runtime/syscall.h               |   38 +
 runtime/uprobes2/uprobes_arch.c |    2 +
 runtime/uprobes2/uprobes_arch.h |    2 +
 runtime/uprobes2/uprobes_arm.c  | 1914 +++++++++++++++++++++++++++++++++++++++
 runtime/uprobes2/uprobes_arm.h  |  114 +++
 5 files changed, 2070 insertions(+), 0 deletions(-)
 create mode 100644 runtime/uprobes2/uprobes_arm.c
 create mode 100644 runtime/uprobes2/uprobes_arm.h
Permalink | Reply | 8 comments |
headers
Wade Farnsworth | 3 Nov 2011 19:08
Favicon

[PATCH 1/2] Userspace tap support for ARM

Add syscall information for ARM devices.  This allows tapsets to be
compiled on this architecture when CONFIG_UTRACE is enabled.

Signed-off-by: Wade Farnsworth <wade_farnsworth <at> mentor.com>
---
 runtime/syscall.h |   38 ++++++++++++++++++++++++++++++++++++++
 1 files changed, 38 insertions(+), 0 deletions(-)

diff --git a/runtime/syscall.h b/runtime/syscall.h
index 786965f..2adcd3b 100644
--- a/runtime/syscall.h
+++ b/runtime/syscall.h
 <at>  <at>  -85,6 +85,14  <at>  <at> 
 #define MREMAP_SYSCALL_NO(tsk)		163
 #endif

+#if defined(__arm__)
+#define MMAP_SYSCALL_NO(tsk)		90
+#define MMAP2_SYSCALL_NO(tsk)		192
+#define MPROTECT_SYSCALL_NO(tsk)	125
+#define MUNMAP_SYSCALL_NO(tsk)		91
+#define MREMAP_SYSCALL_NO(tsk)		163
+#endif
+
 #if !defined(MMAP_SYSCALL_NO) || !defined(MMAP2_SYSCALL_NO)		\
 	|| !defined(MPROTECT_SYSCALL_NO) || !defined(MUNMAP_SYSCALL_NO)	\
 	|| !defined(MREMAP_SYSCALL_NO)
 <at>  <at>  -146,6 +154,14  <at>  <at>  syscall_get_nr(struct task_struct *task, struct pt_regs *regs)
 }
 #endif
(Continue reading)

Permalink | Reply | 2 comments |
headers
Wade Farnsworth | 3 Nov 2011 19:08
Favicon

[PATCH 2/2] ARM uprobes support

Basic uprobes support for ARM.

Signed-off-by: Wade Farnsworth <wade_farnsworth <at> mentor.com>
---
 runtime/uprobes2/uprobes_arch.c |    2 +
 runtime/uprobes2/uprobes_arch.h |    2 +
 runtime/uprobes2/uprobes_arm.c  | 1914 +++++++++++++++++++++++++++++++++++++++
 runtime/uprobes2/uprobes_arm.h  |  114 +++
 4 files changed, 2032 insertions(+), 0 deletions(-)
 create mode 100644 runtime/uprobes2/uprobes_arm.c
 create mode 100644 runtime/uprobes2/uprobes_arm.h

diff --git a/runtime/uprobes2/uprobes_arch.c b/runtime/uprobes2/uprobes_arch.c
index a91d5b6..1df4fdd 100644
--- a/runtime/uprobes2/uprobes_arch.c
+++ b/runtime/uprobes2/uprobes_arch.c
 <at>  <at>  -1,5 +1,7  <at>  <at> 
 #if defined (__x86_64__) || defined(__i386)
 #include "uprobes_x86.c"
+#elif defined (__arm__)
+#include "uprobes_arm.c"
 #elif defined (__powerpc__)
 #include "../uprobes/uprobes_ppc.c"
 #elif defined (__s390__) || defined (__s390x__)
diff --git a/runtime/uprobes2/uprobes_arch.h b/runtime/uprobes2/uprobes_arch.h
index cce5775..a499279 100644
--- a/runtime/uprobes2/uprobes_arch.h
+++ b/runtime/uprobes2/uprobes_arch.h
 <at>  <at>  -1,5 +1,7  <at>  <at> 
 #if defined (__x86_64__) || defined(__i386)
(Continue reading)

Permalink | Reply | 2 comments |
headers
Wade Farnsworth | 3 Nov 2011 19:10
Favicon

[PATCH] Fix uninitialized variable warning in __stp_call_mmap_callbacks_with_addr

When compiling a tapset, gcc complains that dentry may be used uninitialized in
this function.  Initialize it to NULL to prevent this.

Signed-off-by: Wade Farnsworth <wade_farnsworth <at> mentor.com>
---
 runtime/task_finder.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/runtime/task_finder.c b/runtime/task_finder.c
index 70d5ba1..6be9b07 100644
--- a/runtime/task_finder.c
+++ b/runtime/task_finder.c
 <at>  <at>  -678,7 +678,7  <at>  <at>  __stp_call_mmap_callbacks_with_addr(struct stap_task_finder_target *tgt,
 	struct vm_area_struct *vma;
 	char *mmpath_buf = NULL;
 	char *mmpath = NULL;
-	struct dentry *dentry;
+	struct dentry *dentry = NULL;
 	unsigned long length = 0;
 	unsigned long offset = 0;
 	unsigned long vm_flags = 0;
--

-- 
1.7.0.4
Permalink | Reply | 1 comment |

Gmane