William Bourque [Mon, 19 Apr 2010 22:33:51 +0000 (18:33 -0400)]
Changed freeHandle() to test for empty before calling free
William Bourque [Mon, 19 Apr 2010 22:03:50 +0000 (18:03 -0400)]
Support for multiple trace version in the same experiment (traceset)
William Bourque [Fri, 16 Apr 2010 17:36:44 +0000 (13:36 -0400)]
Some cleaning, modified to respect kernel standard.
William Bourque [Fri, 19 Mar 2010 18:33:16 +0000 (14:33 -0400)]
Added files in spec
William Bourque [Fri, 19 Mar 2010 17:16:38 +0000 (13:16 -0400)]
Add -release tag to library_loader as well
William Bourque [Thu, 11 Mar 2010 22:49:59 +0000 (17:49 -0500)]
Added functions to ease support of multiple version of traces
Mathieu Desnoyers [Tue, 2 Feb 2010 22:04:18 +0000 (17:04 -0500)]
update version 0.29
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Benjamin Poirier [Mon, 1 Feb 2010 16:56:27 +0000 (11:56 -0500)]
Use updated event definitions
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 29 Jan 2010 17:20:43 +0000 (12:20 -0500)]
Remove dependency on kernel headers
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 29 Jan 2010 17:06:41 +0000 (12:06 -0500)]
Add header files in SOURCES
make distcheck agrees
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 29 Jan 2010 17:54:30 +0000 (12:54 -0500)]
William Bourque [Tue, 2 Feb 2010 18:03:10 +0000 (13:03 -0500)]
Comments and pointers style format changed to comply to kernel code standard.
Mathieu Desnoyers [Mon, 1 Feb 2010 14:40:18 +0000 (09:40 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sat, 30 Jan 2010 17:28:56 +0000 (12:28 -0500)]
update manual
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sat, 30 Jan 2010 17:28:05 +0000 (12:28 -0500)]
update manual
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sat, 30 Jan 2010 17:27:08 +0000 (12:27 -0500)]
update lttng manual
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sat, 30 Jan 2010 17:25:43 +0000 (12:25 -0500)]
update
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sat, 30 Jan 2010 17:21:39 +0000 (12:21 -0500)]
update manual with licenses
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 29 Jan 2010 17:41:47 +0000 (12:41 -0500)]
remove sync dependency from gui module
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 29 Jan 2010 17:36:56 +0000 (12:36 -0500)]
update version to 0.12.28
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 29 Jan 2010 17:35:04 +0000 (12:35 -0500)]
Temporarily remove sync trace call from gui
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Wlliam Bourque [Tue, 26 Jan 2010 23:34:52 +0000 (18:34 -0500)]
Fixed a bug on a call to ltt_trace_time_span_get. No API change.
Mathieu Desnoyers [Fri, 29 Jan 2010 13:22:22 +0000 (08:22 -0500)]
Disable sync/ completely, build OK
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 29 Jan 2010 00:16:54 +0000 (19:16 -0500)]
Disable sync/ because it does not build on x86_64
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 29 Jan 2010 00:03:27 +0000 (19:03 -0500)]
update version
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Thu, 28 Jan 2010 23:57:36 +0000 (18:57 -0500)]
Fix tracecontext seek backward underflow
Fix a problem with the detailed event list "seek backward". In the following
condition:
- Long interval between events (e.g. generated with power management suspend).
- Happening close to trace start.
- Trace start near 0s 0ns.
The substraction could underflow. Fix this by comparing the time to substract
and floor to trace start time if it would underflow.
The visible effect was that the detailed event list is seeked to the end of the
trace rather than the previous event when going "up" one event prior to the
suspend begin.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
CC: Viktor Rosendahl <viktor.rosendahl@nokia.com>
Pierre-Marc Fournier [Wed, 27 Jan 2010 20:09:32 +0000 (15:09 -0500)]
Add COPYING file
Mathieu Desnoyers [Sat, 23 Jan 2010 00:23:09 +0000 (19:23 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 20 Jan 2010 23:10:03 +0000 (18:10 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Wlliam Bourque [Wed, 6 Jan 2010 21:49:57 +0000 (16:49 -0500)]
(Hopefully) Merge of the patches.
Mathieu Desnoyers [Mon, 4 Jan 2010 18:53:58 +0000 (13:53 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 23 Dec 2009 18:54:45 +0000 (13:54 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 23 Dec 2009 16:17:39 +0000 (11:17 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Sun, 20 Dec 2009 12:35:30 +0000 (07:35 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Benjamin Poirier [Fri, 18 Dec 2009 19:15:11 +0000 (14:15 -0500)]
Add glib options to help out valgrind
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 18 Dec 2009 19:15:10 +0000 (14:15 -0500)]
Partial revert "runlttv: change valgrind args to give more details"
This reverts commit
2a65ba6ead92f40eded9946c3413b39d81a338ba.
As the manual says "Performance overhead: origin tracking is expensive."
Conflicts:
runlttv
Benjamin Poirier [Fri, 18 Dec 2009 22:37:56 +0000 (17:37 -0500)]
Fixes for building without libglpk
This fixes some compilation bugs and one execution bug when generating
--eval-graphs without libglpk
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 18 Dec 2009 15:41:26 +0000 (10:41 -0500)]
Fix compilation warnings
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 17 Dec 2009 16:28:29 +0000 (11:28 -0500)]
Rebuild traceset contexts after performing synchronization
This fixes an integration bug with the state system that caused the control
flow view display to become corrupted when zooming in closely to synchronized
traces. It also caused many messages like
WARNING **: Cannot find pin_in in schedchange 5
to be displayed.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 14 Dec 2009 16:44:49 +0000 (11:44 -0500)]
Correct the README regarding eval mode usage
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Sun, 6 Dec 2009 20:20:44 +0000 (15:20 -0500)]
Change a test file to create hulls that are more spaced out
This way they are easier to see.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 27 Nov 2009 18:43:05 +0000 (13:43 -0500)]
Fix leaks and invalid memory accesses
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 27 Nov 2009 16:09:44 +0000 (11:09 -0500)]
Add a processing module that reads events from a text file
This is particularly useful for unitests.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 26 Nov 2009 22:03:06 +0000 (17:03 -0500)]
Add sync_chain_unittest
Makes the former unitest program more compliant with the current architecture.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 26 Nov 2009 20:57:03 +0000 (15:57 -0500)]
Use a common function to print stats of all modules
Removes code duplication.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 26 Nov 2009 20:39:42 +0000 (15:39 -0500)]
Initialize traceNb in the processing modules
Removes manipulation of structure fields by sync_chain.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 26 Nov 2009 18:43:30 +0000 (13:43 -0500)]
Make option type OPTIONAL_ARG available
These options are not be supported by Lttv and will be ignored (with a warning
message).
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 26 Nov 2009 16:12:58 +0000 (11:12 -0500)]
initProcessing takes a va_arg rather than an lttv-specific type
Makes the framework more general.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 20:57:01 +0000 (15:57 -0500)]
Move some functions around to improve overall structure
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 20:14:41 +0000 (15:14 -0500)]
Clean debug statements
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 20:06:03 +0000 (15:06 -0500)]
Remove unused g_info definitions
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 20:02:48 +0000 (15:02 -0500)]
Remove leftover references to processing_lttv
It is now called processing_lttng.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 19:58:01 +0000 (14:58 -0500)]
Don't access stats if they're not activated
Fixes a bug.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 19:41:51 +0000 (14:41 -0500)]
Update README with info about algorithms, glpk and more
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 25 Nov 2009 19:41:34 +0000 (14:41 -0500)]
Print approximation type from a constant instead of hardcoded
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 24 Nov 2009 22:18:20 +0000 (17:18 -0500)]
Make the accuracy area easier to see on the broadcast graphs
Replots the contour of the area over the broadcast points.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 24 Nov 2009 21:17:51 +0000 (16:17 -0500)]
Have callbacks for "background" and "foreground" graphs
Allows easier composition of many functions on a single graph.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 17 Nov 2009 19:35:48 +0000 (14:35 -0500)]
Generate graphs of synchronization accuracy
This depends on libglpk. Automatically enabled if present at configure time.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 16 Nov 2009 22:04:54 +0000 (17:04 -0500)]
Store graph callbacks in a structure
Also support two classes of graphs: with "trace-trace" scale (both axes
present timestamp data); with "trace-time" scale (horizontal axis presents
timestamp data, vertical axis presents difference between timstamps)
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 13 Nov 2009 20:30:07 +0000 (15:30 -0500)]
Generate graphs of broadcasts
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 16 Nov 2009 16:17:24 +0000 (11:17 -0500)]
Use a timescale without correction in the graphs
... since graphs present timestamp data without correction.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 13 Nov 2009 17:06:09 +0000 (12:06 -0500)]
Do not use LttCycleCount elsewhere than in processing_lttv
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 6 Nov 2009 21:56:40 +0000 (16:56 -0500)]
Print a sum row in the evaluation statistics
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 6 Nov 2009 19:22:31 +0000 (14:22 -0500)]
Show the latency histograms by IP address rather than by trace
A trace may contain traffic between more than just one pair of hosts.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 5 Nov 2009 20:27:54 +0000 (15:27 -0500)]
Add graphStream field to syncState
So that modules can create their own individual graphs, independently of the
write[...]GraphsPlots()/Options() mechanism.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 4 Nov 2009 22:15:06 +0000 (17:15 -0500)]
Use histograms with logarithmicly-sized bins
These are very cool and fancy! In a single pass you get good resolution for
small values (not all lumped in one bin) without wasting many small bins for
large values. You also do no loose any values at all thanks to underflow and
overflow bins at each end. Number of bins is configurable.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 3 Nov 2009 18:57:36 +0000 (13:57 -0500)]
Call the stats and graph functions from sync_chain
Versus the former daisy chain method, this avoids having the analysis stats
and graph functions called twice when there are many matching modules (via
matching_distributor).
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 3 Nov 2009 16:31:07 +0000 (11:31 -0500)]
Add a module to distribute messages to many analysis modules
Also perform RTT analysis in the eval module.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 2 Nov 2009 19:58:49 +0000 (14:58 -0500)]
Adds wall time field to events
Also analyze more events in the eval module.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 30 Oct 2009 23:09:26 +0000 (19:09 -0400)]
Add support for module options
This is independant (down to sync_chain) of lttv's option mechanism.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 28 Oct 2009 19:10:53 +0000 (15:10 -0400)]
Add a module to evaluate the quality of synchronization
Prints information about message inversions.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 27 Oct 2009 21:08:28 +0000 (17:08 -0400)]
Add a broadcast matching module
Matches UDP packets based on their first 8 bytes of content (this also depends
on the kernel instrumentation).
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 27 Oct 2009 17:56:25 +0000 (13:56 -0400)]
Add a batchanalysis module to build and run a sync chain
This is mostly to build a sync chain with an analysis module that evaluates
the quality of synchronization. It does not modifiy the time correction
factors in the traces.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 26 Oct 2009 20:59:42 +0000 (16:59 -0400)]
Processing of UDP events
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Tue, 20 Oct 2009 18:25:06 +0000 (14:25 -0400)]
Adjust the marker names used for clock synchronization
... according to the patches posted to the ltt-dev list on 2009-10-21. There
are now regular and _extended version of the markers. Synchronization needs
the extended version.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 8 Oct 2009 20:04:37 +0000 (16:04 -0400)]
Add a README for the clock synchronization code
This README describes the principles, usage and design of clock
synchronization in LTT.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 8 Oct 2009 20:03:07 +0000 (16:03 -0400)]
Make the synchronization module interfaces more generic
Instead of taking NetEvents and Packets, public interfaces take Events,
Messages and Exchanges. These are specialized into other structures for TCP.
This is to support the eventual integration of algorithms based on other event
types, like UDP.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 4 Sep 2009 18:44:33 +0000 (14:44 -0400)]
Add a unittest program for clock synchronization modules
Allows to test matching and analysis modules with data read from text files.
Includes some sample data for simple good and bad (unsynchronizable) cases.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 14 Aug 2009 19:54:45 +0000 (15:54 -0400)]
Add convex hull algorithm-based synchronization
This analysis module implements an algorithm that provides a garantee that the
synchronization will not result in inverted messages. It is now the default
algorithm, over linear regression.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 10 Aug 2009 20:13:40 +0000 (16:13 -0400)]
Do not use pkfree_skb events for synchronization
Don't rely on events indicating when sk_buff structures are freed. After a
receive, we wait for another event indicating that this receive was for TCP
data. In the case where the data was not TCP, instead of keeping information
about the receive, we used to discard it when the skb was freed. It turns out
that it faster (and simpler) not to look at pkfree_skb events and keep the
information around anyways. Since sk_buff's are allocated in a pool, the
information will get overwritten and the size of pendingRecv will not grow
infinitely.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 23 Jul 2009 21:59:24 +0000 (17:59 -0400)]
Split the synchronization code into modules
Separates synchronization code in processing, matching and analysis modules.
There is also a "sync-chain" that coordinates module creation.
Adds a "--sync-null" option, mainly for performance evaluation.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Wed, 22 Jul 2009 18:21:32 +0000 (14:21 -0400)]
Graphical mode synchronization
Hooks the trace synchronization code in lttvwindow. This allows to use trace
synchronization in graphical mode when lttv is started with the "--sync"
option. Unfortunately, the viewer interface "freezes" while the
synchronization code is running. This can take a noticeable amount of time
(more than a minute) for large traces.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Fri, 10 Jul 2009 15:31:47 +0000 (11:31 -0400)]
Text mode clock synchronization
Adds linear regression based clock correction of a group of traces. This is
mostly a reimplementation of the work presented in
[1] Clément, E.: Synchronisation de traces dans un réseau distribué, École
Polytechnique de montréal, 2006
This implementation does not use a synchronization "window" and timeouts as
described in figure 3.3 of [1]. Not using that algorithm, we don't have to
rely on the system time to be loosely synchronized; events can be processed
regardless of the order the traces are merge-sorted. The downside is that
there is no limit to the number of events that could be queued up, waiting for
their matching event from another trace. The solution to this would be to
insert extra logic (like a plugable module) that controls the merge-sort;
upstream of event delivery to the syncrhonization code. Many possibilities
come to mind:
1) reimplement something like what is found in [1] that relies on system time
for a weak pre-synchronization
2) use a heuristic based on the size of the current queue for each trace in
the synchronization code. If a trace's queue is growing and growing, deliver
events to other traces.
This implementation also does not need to keep track of the IP adresses
assigned to local interfaces. If a packet is processed by the network
subsystem it can be used for synchronization. There is no need to know if the
node really the packet's final destination.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Thu, 9 Jul 2009 19:13:56 +0000 (15:13 -0400)]
Add members to LttTrace to support time adjustments
Adds a function to convert tsc values to nanoseconds.
Adds drift and offset members for linear clock correction.
Signed-off-by: Benjamin Poirier <benjamin.poirier@polymtl.ca>
Benjamin Poirier [Mon, 31 Aug 2009 21:11:13 +0000 (17:11 -0400)]
Revert "sync frequency of all traces with the first loaded"
This reverts commit
9c731a505cf0d0ded02f8e90a823fa90084c7740.
Conflicts:
ltt/branches/poly/ltt/tracefile.c
Mathieu Desnoyers [Wed, 16 Dec 2009 22:38:30 +0000 (17:38 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 16 Dec 2009 22:35:13 +0000 (17:35 -0500)]
update version
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 16 Dec 2009 22:33:45 +0000 (17:33 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Wed, 16 Dec 2009 21:23:28 +0000 (16:23 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Tue, 15 Dec 2009 18:50:40 +0000 (13:50 -0500)]
Temporary fix for large IRQ tables
IRQ tables, trap tables and softirq tables can grow. The copy mechanism for
these is not correct when saving/restoring state, because it uses always the
name table size (which grows) to copy the saved/restored state snapshots.
We should change the g_new allocations for g_arrays, and use the array length as
boundary for the copy rather than the name table len.
TODO / FIXME !
I currently increased the initial irq name table size to 512 to deal correctly
with ARM. This is a dumb temporary fix.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 4 Dec 2009 16:13:29 +0000 (11:13 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 4 Dec 2009 15:59:33 +0000 (10:59 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Tue, 1 Dec 2009 12:04:38 +0000 (07:04 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 27 Nov 2009 17:43:40 +0000 (12:43 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 27 Nov 2009 17:33:55 +0000 (12:33 -0500)]
update version
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Fri, 27 Nov 2009 17:33:24 +0000 (12:33 -0500)]
state.c coding style fix
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Chris Smowton [Fri, 27 Nov 2009 17:18:31 +0000 (12:18 -0500)]
wakeup wait for cpu display
Here's a slightly more constructive patch: this one adds support in
LTTV's generic state monitoring code and its control flow visualiser to
note when a process is woken by another process using the
sched_try_wakeup event.
Previously the woken process would continue to register its old state
until such time as it got scheduled; here the process transitions to
WAIT_CPU state (like preempted processes, indicating it is ready to run
but not currently scheduled). Ordinarily we see this state exist very
briefly, in between the device driver IRQ (typically) clearing it to run
and the scheduler invocation after irq_exit, but on a heavily loaded
system we might see a large stripe of dark yellow indicating the process
is ready but cannot yet be allocated a core.
The new code in controlflow/eventhooks.c is essentially a copy of the
second half of before_schedchange -- it would be nice to factor these
two and before_execmode, all of which basically identify a process,
create state objects if necessary, and draw his line up to a certain
time.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Mon, 23 Nov 2009 04:52:20 +0000 (23:52 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Mon, 23 Nov 2009 01:09:46 +0000 (20:09 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Mathieu Desnoyers [Thu, 19 Nov 2009 05:46:33 +0000 (00:46 -0500)]
update compat
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
This page took 0.044726 seconds and 4 git commands to generate.