Frederic Suter [Sun, 9 Jul 2017 19:46:15 +0000 (21:46 +0200)]
cleanups
Martin Quinson [Sun, 9 Jul 2017 14:16:23 +0000 (16:16 +0200)]
update the insider doc on multiple build trees
Martin Quinson [Sun, 9 Jul 2017 08:33:07 +0000 (10:33 +0200)]
codacy
Frederic Suter [Sun, 9 Jul 2017 19:14:46 +0000 (21:14 +0200)]
codacy
Martin Quinson [Sat, 8 Jul 2017 21:50:11 +0000 (23:50 +0200)]
codacy
Christian Heinrich [Fri, 7 Jul 2017 12:58:14 +0000 (14:58 +0200)]
[SMPI] Remove 2nd include of private.hpp
Martin Quinson [Fri, 7 Jul 2017 07:50:04 +0000 (09:50 +0200)]
fix MC builds
Christian Heinrich [Thu, 6 Jul 2017 17:49:51 +0000 (19:49 +0200)]
[MC] Missing include of smpi/private.hpp
Christian Heinrich [Thu, 6 Jul 2017 17:29:42 +0000 (19:29 +0200)]
[SMPI] Make spelling uniform: privatisation -> privatization
Christian Heinrich [Thu, 6 Jul 2017 17:08:02 +0000 (19:08 +0200)]
[SMPI] Move SMPI_PRIVATIZATION_* macros to enums
I didn't see any reason this should be a macro. The naming is still
the same and certainly ugly, but this should be better for the moment.
Christian Heinrich [Tue, 27 Jun 2017 08:20:55 +0000 (10:20 +0200)]
[ENERGY] Instead of using plain integers, use constants
Martin Quinson [Thu, 6 Jul 2017 14:36:42 +0000 (16:36 +0200)]
please sonar: remove an unused parameters down below
Also: fix jedule build (sorry for mixing)
Martin Quinson [Thu, 6 Jul 2017 13:12:52 +0000 (15:12 +0200)]
ok, last try here: use uint64_t for size of sent data
All this is ridiculous since we use doubles internally for the fluid models.
But some users don't like to be given the ability to send 0.5 bytes.
Martin Quinson [Thu, 6 Jul 2017 12:33:14 +0000 (14:33 +0200)]
use double for the sizes, so that 32bits simulators can send more than 4G
Martin Quinson [Thu, 6 Jul 2017 12:29:10 +0000 (14:29 +0200)]
Revert "s4u: use size_t for payload size instead of int and double"
This reverts commit
0322a73bb9bb8b3285c6d7fce7e19a40e5025823.
Frederic Suter [Thu, 6 Jul 2017 11:43:51 +0000 (13:43 +0200)]
data is now userdata for process/actor
Frederic Suter [Thu, 6 Jul 2017 11:18:52 +0000 (13:18 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Martin Quinson [Thu, 6 Jul 2017 11:14:31 +0000 (13:14 +0200)]
please sonar
Martin Quinson [Thu, 6 Jul 2017 10:06:54 +0000 (12:06 +0200)]
s4u: implement Engine::setConfig()
Frederic Suter [Thu, 6 Jul 2017 10:01:41 +0000 (12:01 +0200)]
still playing with ActorImpl
Frederic Suter [Thu, 6 Jul 2017 09:39:15 +0000 (11:39 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 6 Jul 2017 09:24:50 +0000 (11:24 +0200)]
make sleep and suspend methods of ActorImpl
Martin Quinson [Thu, 6 Jul 2017 09:24:17 +0000 (11:24 +0200)]
s4u: use size_t for payload size instead of int and double
Martin Quinson [Thu, 6 Jul 2017 08:54:06 +0000 (10:54 +0200)]
s4u::Actor::execute() dont need to return a state
It could be used to say if the host failed during that execution, but
the actor would be killed on such failure anyway. So it can only
return "no problem". Not that useful.
Martin Quinson [Thu, 6 Jul 2017 08:50:45 +0000 (10:50 +0200)]
tend to the google coding standards in all S4U API
Frederic Suter [Thu, 6 Jul 2017 08:27:27 +0000 (10:27 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Thu, 6 Jul 2017 08:26:58 +0000 (10:26 +0200)]
various cleanups in actors/processes
Martin Quinson [Thu, 6 Jul 2017 08:21:23 +0000 (10:21 +0200)]
implement getter s4u::Comm::mailbox()
Martin Quinson [Thu, 6 Jul 2017 08:18:01 +0000 (10:18 +0200)]
document last changes
henricasanova [Thu, 6 Jul 2017 07:49:53 +0000 (09:49 +0200)]
Removed the "state_ != finished" assert in s4u::Comm::wait(...) since
waiting on a finished communication should be a feature, not a bug
Martin Quinson [Thu, 6 Jul 2017 07:39:55 +0000 (09:39 +0200)]
You don't send a mailbox, but instead put stuff on it
this_actor::send(mailbox) is now mailbox->put()
Martin Quinson [Thu, 6 Jul 2017 05:59:56 +0000 (07:59 +0200)]
codacy
Martin Quinson [Thu, 6 Jul 2017 05:51:13 +0000 (07:51 +0200)]
Fix clang build
I've no idea of why g++ don't see the deprecation marker on this
function...
Martin Quinson [Thu, 6 Jul 2017 00:46:16 +0000 (02:46 +0200)]
Rework the OO design of S4U comms
This sounds much better this way:
- simgrid::s4u::this_actor::send(mailbox, &pid, comm_size);
+ mailbox->send(&pid, comm_size);
That's intrusive for our (early) users, but the old way should still
work. Only, it raises a deprecation warning.
Martin Quinson [Wed, 5 Jul 2017 23:46:54 +0000 (01:46 +0200)]
fix the API to isend detach() and speak about it in ChangeLog
Martin Quinson [Wed, 5 Jul 2017 23:18:41 +0000 (01:18 +0200)]
change the prototype of s4u::Comm::wait_any to mimick the C code
Returning an iterator was not working as we returned every elements in
the input collection, without any sorting (fix #170).
Returning a collection or an iterator is difficult, as is. We would
have to build the collection in the backstage, iterating over the
provided comms to see which ones are done.
Plus, the user would probably want to then find these comms in the
provided vector to remove them from there.
In the end, that would be a lot of useless and potentially unefficient
code. It seems much better to provide the index in the vector provided
by the user. If she wants, she can iterate from there to see if the
comms are to be removed, or she can just remove this one and fire
another wait_any.
Maybe less C++ish, but more efficient and not that ugly either.
Martin Quinson [Wed, 5 Jul 2017 23:14:55 +0000 (01:14 +0200)]
move a symbol decl to a better header file
Martin Quinson [Wed, 5 Jul 2017 22:45:40 +0000 (00:45 +0200)]
revalidate the tesh files after fixing #195
Martin Quinson [Wed, 5 Jul 2017 22:39:17 +0000 (00:39 +0200)]
all actors are unified, no need to virtualize between APIs
In addition, the virtualization was not done in s4u, so the %i
formatter of the logs was broken. (fix #195).
Martin Quinson [Wed, 5 Jul 2017 22:35:37 +0000 (00:35 +0200)]
always out of tree builds are annoying
Martin Quinson [Wed, 5 Jul 2017 22:23:23 +0000 (00:23 +0200)]
don't fail on sg_host_energy_update_all() in presence of VMs (fix #192)
Frederic Suter [Wed, 5 Jul 2017 21:39:51 +0000 (23:39 +0200)]
simcall to kernelImmediate
Frederic Suter [Wed, 5 Jul 2017 20:46:13 +0000 (22:46 +0200)]
another move to FileImpl
Martin Quinson [Wed, 5 Jul 2017 16:04:58 +0000 (18:04 +0200)]
revalidate a last example after the fix of loopback' latency
Martin Quinson [Wed, 5 Jul 2017 15:49:41 +0000 (17:49 +0200)]
kill the msg-chord example: the s4u one is much better
Martin Quinson [Wed, 5 Jul 2017 15:31:44 +0000 (17:31 +0200)]
make it clear to static analyzers that this mem is not leaked
Martin Quinson [Wed, 5 Jul 2017 15:15:43 +0000 (17:15 +0200)]
revalidate some tesh after the loopback latency fix
Martin Quinson [Wed, 5 Jul 2017 15:07:00 +0000 (17:07 +0200)]
fiddeling with BuildSimGrid.sh
Martin Quinson [Wed, 5 Jul 2017 15:06:25 +0000 (17:06 +0200)]
codacy
Martin Quinson [Wed, 5 Jul 2017 15:06:16 +0000 (17:06 +0200)]
create the loopback after the right configuration is set
Martin Quinson [Wed, 5 Jul 2017 14:37:53 +0000 (16:37 +0200)]
this header is C++ only
Frederic Suter [Wed, 5 Jul 2017 15:15:29 +0000 (17:15 +0200)]
revalide tesh
Frederic Suter [Wed, 5 Jul 2017 13:17:21 +0000 (15:17 +0200)]
don't use old values in doc and alternate models
Martin Quinson [Wed, 5 Jul 2017 13:06:14 +0000 (15:06 +0200)]
further improve the entry page of the doc
Frederic Suter [Wed, 5 Jul 2017 12:22:56 +0000 (14:22 +0200)]
add missing new file
Frederic Suter [Wed, 5 Jul 2017 12:16:29 +0000 (14:16 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Wed, 5 Jul 2017 12:16:10 +0000 (14:16 +0200)]
modernize one more simcall
Frederic Suter [Wed, 5 Jul 2017 12:08:38 +0000 (14:08 +0200)]
free the engine at the end of test
Martin Quinson [Wed, 5 Jul 2017 08:17:06 +0000 (10:17 +0200)]
please codacy
Frederic Suter [Tue, 4 Jul 2017 13:40:03 +0000 (15:40 +0200)]
add daemonize for s4u actors too
Frederic Suter [Tue, 4 Jul 2017 12:25:41 +0000 (14:25 +0200)]
another try to solve weird travis issues
Martin Quinson [Tue, 4 Jul 2017 10:24:12 +0000 (12:24 +0200)]
FIX THE VM SHARING
Martin Quinson [Tue, 4 Jul 2017 10:00:38 +0000 (12:00 +0200)]
refactor: be explicit on what an action's priority is
Martin Quinson [Tue, 4 Jul 2017 09:42:01 +0000 (11:42 +0200)]
maxmin: rename a variable for clarity
Frederic Suter [Tue, 4 Jul 2017 10:12:07 +0000 (12:12 +0200)]
fix include
Frederic Suter [Tue, 4 Jul 2017 10:07:14 +0000 (12:07 +0200)]
address a todo
Frederic Suter [Tue, 4 Jul 2017 09:41:55 +0000 (11:41 +0200)]
modernize some simcalls
Frederic Suter [Tue, 4 Jul 2017 09:25:12 +0000 (11:25 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid
Frederic Suter [Tue, 4 Jul 2017 09:16:46 +0000 (11:16 +0200)]
get rid of smx_file_t
henricasanova [Tue, 4 Jul 2017 09:15:58 +0000 (11:15 +0200)]
Replaced:
if (state_ == finished)
xbt_die("Don't call test on a finished comm.");
by:
if (state_ == finished)
return true;
Since testing a finished comm should be a feature, not a bug
Martin Quinson [Tue, 4 Jul 2017 00:11:30 +0000 (02:11 +0200)]
further reduce the intrusive_ptr churn
Martin Quinson [Tue, 4 Jul 2017 00:04:19 +0000 (02:04 +0200)]
fix non-MC builds
Martin Quinson [Tue, 4 Jul 2017 00:04:05 +0000 (02:04 +0200)]
please clang
Martin Quinson [Mon, 3 Jul 2017 23:58:04 +0000 (01:58 +0200)]
finish removing simix.h from C files, and avoid the use of Ptr when not needed
Martin Quinson [Mon, 3 Jul 2017 19:48:08 +0000 (21:48 +0200)]
try to speed up the refcounting madness by using std::move
Frederic Suter [Mon, 3 Jul 2017 15:58:45 +0000 (17:58 +0200)]
please henri, add recv with timeout.
Frederic Suter [Mon, 3 Jul 2017 08:13:52 +0000 (10:13 +0200)]
please sonar
Martin Quinson [Sun, 2 Jul 2017 21:33:31 +0000 (23:33 +0200)]
Merge branch 'master' of github.com:simgrid/simgrid
Martin Quinson [Sun, 2 Jul 2017 21:32:43 +0000 (23:32 +0200)]
remove simix.h from one more C file
Martin Quinson [Sat, 1 Jul 2017 21:20:53 +0000 (23:20 +0200)]
remove simix.h from two more C files
Martin Quinson [Sat, 1 Jul 2017 21:12:27 +0000 (23:12 +0200)]
rename SIMIX_host_self() into sg_host_self(), and make it public
Martin Quinson [Sat, 1 Jul 2017 08:27:46 +0000 (10:27 +0200)]
only take the public part of MC within the mallocators
I'm trying to not load simix.h from C code anymore, to ease the work
on ActivityImpl.
Martin Quinson [Fri, 30 Jun 2017 23:40:38 +0000 (01:40 +0200)]
regenerate the simcall code -- sorry for the noisy commit
Martin Quinson [Fri, 30 Jun 2017 23:39:28 +0000 (01:39 +0200)]
try to reduce the amount of casts by using the right type for simcalls
Martin Quinson [Fri, 30 Jun 2017 23:06:33 +0000 (01:06 +0200)]
generate code that does not drive clang-format crazy
Martin Quinson [Fri, 30 Jun 2017 23:03:42 +0000 (01:03 +0200)]
please sonar
Martin Quinson [Fri, 30 Jun 2017 20:38:10 +0000 (22:38 +0200)]
Merge pull request #193 from Takishipp/signals
get sg_instr_new_router through a signal
Martin Quinson [Fri, 30 Jun 2017 15:29:17 +0000 (17:29 +0200)]
well, v3.8 was released a while ago, now :)
This paragraph kept away from us
https://theses.ncl.ac.uk/dspace/bitstream/10443/2981/1/Alabdulhafez,%20A%202015.pdf
Takishipp [Fri, 30 Jun 2017 10:18:49 +0000 (12:18 +0200)]
get sg_instr_new_router through a signal
Martin Quinson [Thu, 29 Jun 2017 20:56:31 +0000 (22:56 +0200)]
make it possible to destroy the engine
Martin Quinson [Wed, 28 Jun 2017 11:57:36 +0000 (13:57 +0200)]
VM migration: do not precompute dp_rate
Before, the rate was computed at VM creation, which would have lead to
wrong results if the host_speed changes between the time at with the
VM is created and the migration time.
(fix #191)
Martin Quinson [Wed, 28 Jun 2017 11:20:30 +0000 (13:20 +0200)]
include build path before source path to ease out of source builds
Martin Quinson [Wed, 28 Jun 2017 11:19:39 +0000 (13:19 +0200)]
cmake/java: don't rebuild the libs before building the jar if we won't pack them
Martin Quinson [Wed, 28 Jun 2017 09:34:13 +0000 (11:34 +0200)]
hardlinks target now recreate the links
A file got desynch'ed on my disk. Maybe git did remove and recreate it?
Martin Quinson [Wed, 28 Jun 2017 09:25:27 +0000 (11:25 +0200)]
use a msg_barrier instead of smx synchro in this example
Martin Quinson [Tue, 27 Jun 2017 21:23:22 +0000 (23:23 +0200)]
use MSG_bar instead of xbt_mutex in this example
Martin Quinson [Tue, 27 Jun 2017 20:57:06 +0000 (22:57 +0200)]
cosmetics
Martin Quinson [Tue, 27 Jun 2017 20:44:07 +0000 (22:44 +0200)]
MC: compute if a COMM_WAIT is enabled with the activity only, w/o the request
Frederic Suter [Wed, 28 Jun 2017 08:51:54 +0000 (10:51 +0200)]
please sonar and simplify simcalls (before killing them)
Frederic Suter [Wed, 28 Jun 2017 08:28:15 +0000 (10:28 +0200)]
Merge branch 'master' of git+ssh://scm.gforge.inria.fr//gitroot/simgrid/simgrid