mirror of
https://github.com/varun-r-mallya/sysprof.git
synced 2026-02-12 16:10:54 +00:00
Merge stackstash-reorg branch into HEAD
2005-10-30 Soren Sandmann <sandmann@redhat.com> * Merge stackstash-reorg branch into HEAD
This commit is contained in:
Soren Sandmann
committed by
Søren Sandmann Pedersen
Søren Sandmann Pedersen
parent
3783be00a8
commit
dff4affaab
140
TODO
140
TODO
@ -11,9 +11,9 @@ Before 1.0.1:
|
||||
|
||||
See also http://www.fedoraproject.org/wiki/Extras/KernelModuleProposal
|
||||
|
||||
Before 1.2:
|
||||
Someone already did create a package - should be googlable.
|
||||
|
||||
* Crash reported by Rudi Chiarito with n_addrs == 0.
|
||||
Before 1.2:
|
||||
|
||||
* Find out why we get hangs with rawhide kernels. This only happens with the
|
||||
'trace "current"' code. See this mail:
|
||||
@ -25,6 +25,54 @@ Before 1.2:
|
||||
|
||||
(Reported by Kjartan Marass).
|
||||
|
||||
- Fix bugs/performance issues:
|
||||
- total should probably be cached so that compute_total() doesn't
|
||||
take 80% of the time to generate a profile.
|
||||
- decorate_node should be done lazily
|
||||
- Find out why we sometimes get completely ridicoulous stacktraces,
|
||||
where main seems to be called from within Xlib etc. This happens
|
||||
even after restarting everything.
|
||||
- It looks like the stackstash-reorg code confuses "main" from
|
||||
unrelated processes. - currently it looks like if multiple
|
||||
"main"s are present, only one gets listed in the object list.
|
||||
- Numbers in caller view are completely screwed up.
|
||||
- It looks like it sometimes gets confused with similar but different
|
||||
processess: Something like:
|
||||
process a spends 80% in foo() called from bar()
|
||||
process b spends 1% in foo() called from baz()
|
||||
we get reports of baz() using > 80% of the time.
|
||||
Or something.
|
||||
|
||||
- make the things we put in a stackstash real
|
||||
objects so that
|
||||
- we can save them
|
||||
- they will know how to delete the presentation
|
||||
names and themselves (through a virtual function)
|
||||
- they can contain markup etc.
|
||||
- a more pragmatic approach might be to just walk the tree and
|
||||
save it.
|
||||
|
||||
- make stasckstash ref counted
|
||||
|
||||
- plug all the leaks
|
||||
- don't leak the presentation strings/objects
|
||||
- loading and saving probably leak right now
|
||||
|
||||
- think about loading and saving. Goals
|
||||
- Can load 1.0 profiles
|
||||
- Don't export too much of stackstashes to the rest of the
|
||||
app
|
||||
|
||||
- Reorganise stackstash and profile
|
||||
|
||||
- Remaining TODO before merging into head:
|
||||
|
||||
- rename profiler->collector
|
||||
|
||||
* Consider renaming profiler.[ch] to collector.[ch]
|
||||
|
||||
* Make sure sysprof-text is not linked to gtk+
|
||||
|
||||
* Figure out how to make sfile.[ch] use less memory.
|
||||
- In general clean sfile.[ch] up a little:
|
||||
- split out dfa in its own generic class
|
||||
@ -192,39 +240,6 @@ http://www.linuxbase.org/spec/booksets/LSB-Embedded/LSB-Embedded/ehframe.html
|
||||
so (can we know the size in advance?))
|
||||
- instead of what we do now: set the busy cursor unconditionally
|
||||
|
||||
- Reorganise stackstash and profile
|
||||
|
||||
- stackstash should just take traces of addresses without knowing
|
||||
anything about what those addresses mean.
|
||||
|
||||
- stacktraces should then begin with a process
|
||||
|
||||
- stackstash should be extended so that the "create_descendant"
|
||||
and "create_ancestor" code in profile.c can use it directly.
|
||||
At that point, get rid of the profile tree, and rename
|
||||
profile.c to analyze.c.
|
||||
|
||||
- the profile tree will then just be a stackstash where the
|
||||
addresses are presentation strings instead.
|
||||
|
||||
- Doing a profile will then amount to converting the raw stash
|
||||
to one where the addresses have been looked up and converted to
|
||||
presentation strings.
|
||||
|
||||
-=-=
|
||||
|
||||
- profile should take traces of pointers to presentation
|
||||
objects without knowing anything about these presentation
|
||||
objects.
|
||||
|
||||
- For each stack node, compute a presentation object
|
||||
(probably need to export opaque stacknode objects
|
||||
with set/get_user_data)
|
||||
|
||||
- Send each stack trace to the profile module, along with
|
||||
presentation objects. Maybe just a map from stack nodes
|
||||
to presentation objects.
|
||||
|
||||
- Charge 'self' properly to processes that don't get any stack trace at all
|
||||
(probably we get that for free with stackstash reorganisation)
|
||||
|
||||
@ -249,7 +264,8 @@ http://www.linuxbase.org/spec/booksets/LSB-Embedded/LSB-Embedded/ehframe.html
|
||||
- Add support for line numbers within functions
|
||||
- Possibly a special "view details" mode, assuming that
|
||||
the details of a function are not that interesting
|
||||
together with a tree.
|
||||
together with a tree. (Could add radio buttons somewhere in
|
||||
in the right pane).
|
||||
- rethink caller list, not terribly useful at the moment. Federico suggested
|
||||
listing all ancestors.
|
||||
|
||||
@ -452,12 +468,64 @@ Later:
|
||||
|
||||
DONE:
|
||||
|
||||
* Crash reported by Rudi Chiarito with n_addrs == 0.
|
||||
|
||||
* Find out what distributions it actually works on
|
||||
(ask for sucess/failure-stories in 1.0 releases)
|
||||
|
||||
* Add note in README about Ubuntu and Debian -dbg packages and how to get
|
||||
debug symbols for X there.
|
||||
|
||||
stackstash reorg:
|
||||
|
||||
- make loading and saving work again.
|
||||
- make stashes loadable and savable.
|
||||
- add a way to convert 1.0 files to stashes
|
||||
|
||||
- Get rid of remaining uses of stack_stash_foreach(), then
|
||||
rename stack_stash_foreach_reversed() to
|
||||
stack_stash_foreach()
|
||||
|
||||
- stackstash should just take traces of addresses without knowing
|
||||
anything about what those addresses mean.
|
||||
|
||||
- stacktraces should then begin with a process
|
||||
|
||||
- stackstash should be extended so that the "create_descendant"
|
||||
and "create_ancestor" code in profile.c can use it directly.
|
||||
At that point, get rid of the profile tree, and rename
|
||||
profile.c to analyze.c.
|
||||
|
||||
- the profile tree will then just be a stackstash where the
|
||||
addresses are presentation strings instead.
|
||||
|
||||
- Doing a profile will then amount to converting the raw stash
|
||||
to one where the addresses have been looked up and converted to
|
||||
presentation strings.
|
||||
|
||||
-=-=
|
||||
|
||||
- profile should take traces of pointers to presentation
|
||||
objects without knowing anything about these presentation
|
||||
objects.
|
||||
|
||||
- For each stack node, compute a presentation object
|
||||
(probably need to export opaque stacknode objects
|
||||
with set/get_user_data)
|
||||
|
||||
- Send each stack trace to the profile module, along with
|
||||
presentation objects. Maybe just a map from stack nodes
|
||||
to presentation objects.
|
||||
|
||||
- Make the Profile class use the stash directly instead of
|
||||
building its own copy.
|
||||
- store a stash in the profile class
|
||||
- make sure descendants and callers can be
|
||||
built from it.
|
||||
- get rid of other stuff in the profile
|
||||
struct
|
||||
|
||||
|
||||
* Before 1.0:
|
||||
|
||||
- Update version numbers in source
|
||||
|
||||
Reference in New Issue
Block a user