From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 11851 invoked by alias); 23 Apr 2003 19:47:37 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 11819 invoked from network); 23 Apr 2003 19:47:36 -0000 Received: from unknown (HELO touchme.toronto.redhat.com) (207.219.125.105) by sources.redhat.com with SMTP; 23 Apr 2003 19:47:36 -0000 Received: from redhat.com (toocool.toronto.redhat.com [172.16.14.72]) by touchme.toronto.redhat.com (Postfix) with ESMTP id 452708000F9 for ; Wed, 23 Apr 2003 15:47:36 -0400 (EDT) Message-ID: <3EA6EDD8.8000100@redhat.com> Date: Wed, 23 Apr 2003 21:16:00 -0000 From: "J. Johnston" Organization: Red Hat Inc. User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020823 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: gdb-patches@sources.redhat.com Subject: RFA: fix for thread events in thread-db.c Content-Type: multipart/mixed; boundary="------------040902040102040005050507" X-SW-Source: 2003-04/txt/msg00445.txt.bz2 This is a multi-part message in MIME format. --------------040902040102040005050507 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-length: 866 The attached patch fixes a long-standing problem in thread-db.c: check_event(). When a create or death breakpoint is detected, the td_ta_event_getmsg() interface from libthread_db is used to get "a message" from the queue. This message does not necessarily match the breakpoint just reached. The td_ta_event_getmsg() interface has to be used because we don't the lwp that has been created. The change made is to loop through all available messages and process them. This ensures that the message set for the breakpoint is processed before continuing. This change allows death events to be enabled, if desired. Ok to commit? -- Jeff J. 2003-04-23 Jeff Johnston * thread-db.c (check_event): For create/death event breakpoints, loop through all messages to ensure that we read the message corresponding to the breakpoint we are at. --------------040902040102040005050507 Content-Type: text/plain; name="thread-db.nptl.patch" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="thread-db.nptl.patch" Content-length: 3777 Index: thread-db.c =================================================================== RCS file: /cvs/src/src/gdb/thread-db.c,v retrieving revision 1.30 diff -u -r1.30 thread-db.c --- thread-db.c 17 Apr 2003 17:30:02 -0000 1.30 +++ thread-db.c 23 Apr 2003 19:37:47 -0000 @@ -775,64 +775,73 @@ td_thrinfo_t ti; td_err_e err; CORE_ADDR stop_pc; + int loop = 0; /* Bail out early if we're not at a thread event breakpoint. */ stop_pc = read_pc_pid (ptid) - DECR_PC_AFTER_BREAK; if (stop_pc != td_create_bp_addr && stop_pc != td_death_bp_addr) return; - err = td_ta_event_getmsg_p (thread_agent, &msg); - if (err != TD_OK) + /* If we are at a create breakpoint, we do not know what new lwp + was created and cannot specifically locate the event message for it. + We have to call td_ta_event_getmsg() to get + the latest message. Since we have no way of correlating whether + the event message we get back corresponds to our breakpoint, we must + loop and read all event messages, processing them appropriately. + This guarantees we will process the correct message before continuing + from the breakpoint. + + Currently, death events are not enabled. If they are enabled, + the death event can use the td_thr_event_getmsg() interface to + get the message specifically for that lwp and avoid looping + below. */ + + loop = 1; + + do { - if (err == TD_NOMSG) - return; + err = td_ta_event_getmsg_p (thread_agent, &msg); + if (err != TD_OK) + { + if (err == TD_NOMSG) + return; - error ("Cannot get thread event message: %s", thread_db_err_str (err)); - } + error ("Cannot get thread event message: %s", + thread_db_err_str (err)); + } - err = td_thr_get_info_p (msg.th_p, &ti); - if (err != TD_OK) - error ("check_event: cannot get thread info: %s", - thread_db_err_str (err)); + err = td_thr_get_info_p (msg.th_p, &ti); + if (err != TD_OK) + error ("Cannot get thread info: %s", thread_db_err_str (err)); - ptid = BUILD_THREAD (ti.ti_tid, GET_PID (ptid)); + ptid = BUILD_THREAD (ti.ti_tid, GET_PID (ptid)); - switch (msg.event) - { - case TD_CREATE: -#if 0 - /* FIXME: kettenis/2000-08-26: Since we use td_ta_event_getmsg, - there is no guarantee that the breakpoint will match the - event. Should we use td_thr_event_getmsg instead? */ - - if (stop_pc != td_create_bp_addr) - error ("Thread creation event doesn't match breakpoint."); -#endif - - /* We may already know about this thread, for instance when the - user has issued the `info threads' command before the SIGTRAP - for hitting the thread creation breakpoint was reported. */ - if (!in_thread_list (ptid)) - attach_thread (ptid, msg.th_p, &ti, 1); - return; - - case TD_DEATH: -#if 0 - /* FIXME: See TD_CREATE. */ - - if (stop_pc != td_death_bp_addr) - error ("Thread death event doesn't match breakpoint."); -#endif + switch (msg.event) + { + case TD_CREATE: + + /* We may already know about this thread, for instance when the + user has issued the `info threads' command before the SIGTRAP + for hitting the thread creation breakpoint was reported. */ + if (!in_thread_list (ptid)) + attach_thread (ptid, msg.th_p, &ti, 1); + + break; + + case TD_DEATH: + + if (!in_thread_list (ptid)) + error ("Spurious thread death event."); - if (!in_thread_list (ptid)) - error ("Spurious thread death event."); + detach_thread (ptid, 1); - detach_thread (ptid, 1); - return; + break; - default: - error ("Spurious thread event."); + default: + error ("Spurious thread event."); + } } + while (loop); } static ptid_t --------------040902040102040005050507--