[Bug tapsets/11574] New: probefunc() grudge match with user-space IPs

[Bug tapsets/11574] New: probefunc() grudge match with user-space IPs

[fche at redhat dot com]

probefunc() rejects incoming PC values that indicate spaces outside
the kernel module.  But there is no corresponding uprobefunc() in the
tapset that does the complementary lookup on userspace IPs.  So either
probefunc() should just chill and do a plain _stp_symbol_snprint()
regardless of incoming IP value, or else we need a uprobefunc() proper.

(The pp()-parsing hack in probefunc() should probably also go.  If
we had enough debug data to be able to process foo.function("...")
probes, we should also have enough symbol data to map addresses
back to names.  Maybe kprobe.function("..") is an exception, if
we care a lot.)

-- 
           Summary: probefunc() grudge match with user-space IPs
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: tapsets
        AssignedTo: systemtap at sources dot redhat dot com
        ReportedBy: fche at redhat dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=11574

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

[Bug tapsets/11574] probefunc() grudge match with user-space IPs

[fche at redhat dot com]

------- Additional Comments From fche at redhat dot com  2010-05-06 18:03 -------
We already have symname() and usymname().  probefunc() is a fossil.

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


http://sourceware.org/bugzilla/show_bug.cgi?id=11574

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

Re: [Bug tapsets/11574] probefunc() grudge match with user-space IPs

[Jim Keniston]

On Thu, 2010-05-06 at 18:03 +0000, fche at redhat dot com wrote:
> ------- Additional Comments From fche at redhat dot com  2010-05-06 18:03 -------
> We already have symname() and usymname().  probefunc() is a fossil.

If you say so.  I use probefunc() a lot, and -- unlike symname() and
usymname() -- it's documented in stapfuncs(3stap) and the Language
Reference doc.

Jim

Re: [Bug tapsets/11574] probefunc() grudge match with user-space IPs

[Frank Ch. Eigler]

>> We already have symname() and usymname().  probefunc() is a fossil.
>
> If you say so.  I use probefunc() a lot,

So have I, but more out of inertia.

> and -- unlike symname() and usymname() -- it's documented in
> stapfuncs(3stap)

(Not any more!)

> and the Language Reference doc.

With more and more of the tapset functionality becoming
self-documenting within the tapset reference guide & man pages,
perhaps this sort of content should be removed from the language
reference book.

- FChE