April 2012 - Asg - mailman3.cels.anl.gov

branch, master, updated. c2674d6c8d3aa71ddb5467379279f0fdd32e4ddd
by noreply＠mcs.anl.gov 30 Apr '12

30 Apr '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "". The branch, master has been updated via c2674d6c8d3aa71ddb5467379279f0fdd32e4ddd (commit) from d16694b91f9656d510751073af01584d894e11f3 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit c2674d6c8d3aa71ddb5467379279f0fdd32e4ddd Author: Dries Kimpe <dkimpe(a)mcs.anl.gov> Date: Mon Apr 30 13:55:01 2012 -0500 Update c-utils submodule to latest version ----------------------------------------------------------------------- Summary of changes: src/c-utils | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) Diff of changes: diff --git a/src/c-utils b/src/c-utils index d907964..6bc8c1c 160000 --- a/src/c-utils +++ b/src/c-utils @@ -1 +1 @@ -Subproject commit d907964b6074e401fa34635e956141e26b391087 +Subproject commit 6bc8c1c32a4a0c777847e44d4e0bf12e2b38d20a hooks/post-receive --

1 0

Submission of deliverable
by Richard Brooks 30 Apr '12

30 Apr '12

Since we have a deliverable due today, I wanted to check as to how we are expected to deliver it. Email to Lee? Post on wiki? Thanks, -Richard -- =================== R. R. Brooks Associate Professor Holcombe Department of Electrical and Computer Engineering Clemson University 313-C Riggs Hall PO Box 340915 Clemson, SC 29634-0915 USA Tel. 864-656-0920 Fax. 864-656-5910 email: rrb(a)acm.org web: http://www.clemson.edu/~rrb

1 0

Security System Survey Document
by Benafsh Husain 30 Apr '12

30 Apr '12

All, I apologize for the previous message, it was sent by mistake. I have uploaded the Security System Survey document to the UAB repository. As specified it is mentioned as a draft. If there are any comments or feedback on the document, we would appreciate we get them by Friday, April 27th as we will upload the final draft on April 30th. https://svn.cis.uab.edu/hpcl/Nebula/Security/Documents/ Best regards, Benafsh Husain > > _______________________________________________ > ASG mailing list > ASG(a)lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/asg > >

2 2

branch, master, updated. d16694b91f9656d510751073af01584d894e11f3
by noreply＠mcs.anl.gov 27 Apr '12

27 Apr '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "". The branch, master has been updated via d16694b91f9656d510751073af01584d894e11f3 (commit) via 629648ec5637a49e41be638b92af5b34bc409a11 (commit) from e1dfa6fdf9bbb7e0e36c341186db4356479fc617 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit d16694b91f9656d510751073af01584d894e11f3 Author: Dries Kimpe <dkimpe(a)mcs.anl.gov> Date: Fri Apr 27 22:59:31 2012 -0500 Fix memory leaks & implement finalize for resource - Call the finalize function for activated resources - Memory leak fixes for resource (libev related) - Destroy main loop There should be no more memory leaks now (but valgrind can't handle realloc (p, 0) and shows memory leak) commit 629648ec5637a49e41be638b92af5b34bc409a11 Author: Dries Kimpe <dkimpe(a)mcs.anl.gov> Date: Fri Apr 27 22:48:05 2012 -0500 Switch to explicitly initializing default ev loop Part of mem leak cleanup of resource.c ----------------------------------------------------------------------- Summary of changes: aesop.c | 14 ++++--- resource.c | 124 +++++++++++++++++++++++++++++++++++++++++++++++++++++------ resource.h | 14 +++++++ 3 files changed, 133 insertions(+), 19 deletions(-) Diff of changes: diff --git a/aesop.c b/aesop.c index 05e6b29..5a66f7a 100644 --- a/aesop.c +++ b/aesop.c @@ -68,13 +68,15 @@ int aesop_init(const char* resource_list) void aesop_finalize(void) { - initialized--; - if(initialized == 0) - { - ae_hints_component_finalize(); - } + initialized--; + + if(initialized) + return; + + ae_resource_finalize_active (); + ae_hints_component_finalize(); - return; + ae_resource_cleanup (); } diff --git a/resource.c b/resource.c index 5d9eee5..e3addcb 100644 --- a/resource.c +++ b/resource.c @@ -29,6 +29,7 @@ struct ae_resource_init_data char* name; int (*init)(void); void (*finalize)(void); + unsigned int count; }; #define MAX_RESOURCES 32 static struct ae_resource_init_data ae_resource_init_table[MAX_RESOURCES]; @@ -125,12 +126,27 @@ int ae_resource_init_register(const char* resource_name, ae_resource_init_table[ae_resource_init_table_count].init = init; ae_resource_init_table[ae_resource_init_table_count].finalize = finalize; + ae_resource_init_table[ae_resource_init_table_count].finalize = finalize; ae_resource_init_table_count++; return(AE_SUCCESS); } +static int ae_resource_init_register_cleanup (void) +{ + int i; + + for(i=0; i<ae_resource_init_table_count; i++) + { + free (ae_resource_init_table[i].name); + ae_resource_init_table[i].name = 0; + } + ae_resource_init_table_count = 0; + + return AE_SUCCESS; +} + int ae_resource_register(struct ae_resource *resource, int *newid) { return ae_resource_register_with_data(resource, newid, NULL); @@ -162,7 +178,7 @@ int ae_resource_register_with_data(struct ae_resource *resource, int *newid, * event loop and start a watcher that can be used to break it's * execution */ - eloop = EV_DEFAULT; + eloop = ev_default_loop (EVFLAG_AUTO); ev_async_init(&eloop_breaker, ev_break_cb); ev_async_start(eloop, &eloop_breaker); } @@ -319,8 +335,10 @@ void ae_resource_request_poll(ae_context_t context, int resource_id) /** * ae_cancel_op tries to cancel the operation with the given id. This * function returns AE_SUCCESS if the operation was successfully cancelled, - * the callback for the operation is responsible for returning TRITON_ERR_CANCELLED - * in the callback or somehow notifying through the callback that the operation was + * the callback for the operation is responsible for returning + * TRITON_ERR_CANCELLED + * in the callback or somehow notifying through the callback that the + * operation was * cancelled. */ int ae_cancel_op(ae_context_t context, ae_op_id_t op_id) @@ -345,7 +363,8 @@ int ae_cancel_op(ae_context_t context, ae_op_id_t op_id) ctl = (struct ae_ctl *) tmp; if(!ctl) { - /* No blocking operation associated with this op_id. Nothing to cancel. */ + /* No blocking operation associated with this op_id. Nothing to + * cancel. */ return AE_SUCCESS; } @@ -883,7 +902,39 @@ int ae_check_debug_flag(int resource_id) return(ae_resource_entries[idx].debug); } -int ae_resource_init(const char* resource) +static int ae_resource_init_helper (int i) +{ + int ret = AE_SUCCESS; + + if (!ae_resource_init_table[i].count) + { + ret = ae_resource_init_table[i].init(); + if (ret != AE_SUCCESS) + return ret; + } + + ++ae_resource_init_table[i].count; + + return ret; +} + +static int ae_resource_finalize_helper (int i) +{ + int ret = AE_SUCCESS; + + assert (ae_resource_init_table[i].count); + + --ae_resource_init_table[i].count; + + if(!ae_resource_init_table[i].count) + { + ae_resource_init_table[i].finalize(); + } + + return ret; +} + +static int ae_resource_string_helper (const char* resource, int init) { int i; @@ -891,30 +942,77 @@ int ae_resource_init(const char* resource) { if(!strcmp(resource, ae_resource_init_table[i].name)) { - return(ae_resource_init_table[i].init()); + if (init) + ae_resource_init_helper (i); + else + ae_resource_finalize_helper (i); } } return(AE_ERR_NOT_FOUND); } -int ae_resource_init_all(void) +static int ae_resource_all_helper (int init) { int i; - int ret; + int ret = AE_SUCCESS; for(i=0; i<ae_resource_init_table_count; i++) { - ret = ae_resource_init_table[i].init(); + ret = (init ? ae_resource_init_helper (i) + : ae_resource_finalize_helper (i)); + if(ret != AE_SUCCESS) - { - return(ret); - } + break; } - return(AE_SUCCESS); + return ret; } +int ae_resource_init (const char * s) +{ + return ae_resource_string_helper (s, 1); +} + +int ae_resource_finalize (const char * s) +{ + return ae_resource_string_helper (s, 0); +} + +int ae_resource_finalize_all (void) +{ + return ae_resource_all_helper (0); +} + +int ae_resource_init_all (void) +{ + return ae_resource_all_helper (1); +} + +int ae_resource_finalize_active (void) +{ + int i; + int ret = AE_SUCCESS; + + for(i=0; i<ae_resource_init_table_count; i++) + { + while (ae_resource_init_table[i].count) + { + ret = ae_resource_finalize_helper (i); + if (ret != AE_SUCCESS) + return ret; + } + } + + return ret; +} + +int ae_resource_cleanup (void) +{ + ae_resource_init_register_cleanup (); + ev_async_stop (eloop, &eloop_breaker); + ev_default_destroy (); +} /* * Local variables: diff --git a/resource.h b/resource.h index 5fb30ec..71ef02d 100644 --- a/resource.h +++ b/resource.h @@ -69,12 +69,26 @@ int ae_resource_init_register(const char* resource_name, int (*init)(void), void (*finalize)(void)); +/** + * Cleanup resource + */ +int ae_resource_cleanup (void); + /* initializes the resource named in the argument. This function is called * by aesop itself when it wants to activate a resource. */ int ae_resource_init(const char* resource); int ae_resource_init_all(void); +/** + * Finalize specified/all resources + */ +int ae_resource_finalize_all (void); + +int ae_resource_finalize (const char * resource); + +int ae_resource_finalize_active (void); + /* Called by resources to register themselves to the resource framework once * they are initialized and ready for use by aesop */ int ae_resource_register(struct ae_resource *resource, int *newid); hooks/post-receive --

1 0

The thesis has been uploaded.
by Seok Bae Yun 25 Apr '12

25 Apr '12

All, My master's thesis has been uploaded on ASG wiki public. It was approved by the graduate school yesterday, so I expect that there is no copyright issue for me to upload it. You can find the thesis from the following link: https://wiki.mcs.anl.gov/asg/index.php/Main_Page If you find any problem with it, please let me know. Best regards, Seok Bae Yun

2 1

Reminder for Thursday model call
by Dries Kimpe 23 Apr '12

23 Apr '12

Time: 10am Chicago Time Conference Access Number: 760-984-1000 Conference Passcode: 692481 # (ANL: Room 3162) Please send items for the agenda to me at the latest on Wednesday. Thanks, Dries

1 0

unable to attend our conference meeting
by HAIYING SHEN 23 Apr '12

23 Apr '12

All, I would like to attend a presentation given by Argonne National Lab at 4:00pm, related to the work we are doing. So, I am unable to attend our conference meeting. My student Guoxin and Harrison will report their work last week. Best regards, Haiying

2 1

[aesop] #32: __branch_x_state might be used uninitialized in ___x_worker
by aesop 18 Apr '12

18 Apr '12

#32: __branch_x_state might be used uninitialized in ___x_worker ---------------------------+------------------------------------------------ Reporter: dkimpe | Owner: Type: defect | Status: new Priority: trivial | Component: code generator Version: git repository | Keywords: ---------------------------+------------------------------------------------ {{{ AECC parser/tests/blocking/parallelfor3.ae ./parser/tests/blocking/parallelfor3.ae.s: In function ‘__ae_worker_run_parfor’: ./parser/tests/blocking/parallelfor3.ae.s:2925:43: warning: ‘__ae_pbranch_18_13_state’ may be used uninitialized in this function LD (AETEST) parser/tests/blocking/parallelfor3 }}} I'm not sure if gcc is not wrong here: I've modified the code generator to initialize {{{__ae_pbranch_18_13_state}}} at declaration time, and the warning still happens. i.e. {{{ enum ae_pbranch_state __ae_pbranch_18_13_state = 0; }}} I guess it would be possible to {{{goto}}} to the label following the declaration, bypassing the initialization. Since the variable is on the stack, when resuming the worker (and doing a goto to the label), this could indeed be uninitialized. However, nowhere in the function is the address taken of any label within the scope of the xx_state declaration. I don't see how to fix this... -- Ticket URL: <http://trac.mcs.anl.gov/projects/aesop/ticket/32> aesop <https://trac.mcs.anl.gov/projects/aesop> The Aesop Language

1 0

Security System Document
by Benafsh Husain 16 Apr '12

16 Apr '12

All, The draft document for the security system architecture is uploaded into UAB subversion server. This document contains the survey and related art for the security of the File storage system (Deliverable: April 30 th 2012), along with the proposed system architecture. As the document is a draft we would appreciate comments to improve the document. https://svn.cis.uab.edu/hpcl/Nebula/Security/Documents/ Best regards, Benafsh Husain

2 2

Meeting
by Seok Bae Yun 16 Apr '12

16 Apr '12

All, I am very sorry that I cannot attend todays meeting. I am working on my thesis based on this project. Best regards, Seok Bae Yun

1 0