From patchwork Wed Jun 7 10:01:13 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70717 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 71DA0385660B for ; Wed, 7 Jun 2023 10:02:01 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 71DA0385660B DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132121; bh=mm+/XWRQ+DsnGO4nNmVLBFhfnxnpv17BL9oqvmNRQ7s=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=NlvM8NkEJd6vODhnbtMFOfJnRY3IgKg+X75aSzhy1dhiguljpG47o+VBbBz+QEDCH YxmQpO9FCMWzfNuVFZpQ0A2dgk0OricVBl9kbMUEcSeNpJ0VpKDU7FslBpQWJXVXhn oSVAOmI4S3UG479lFoKB1MT6y5AyCB7e2UX8wUGE= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by sourceware.org (Postfix) with ESMTPS id 1882E385842C for ; Wed, 7 Jun 2023 10:01:27 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 1882E385842C Received: from mail-wm1-f69.google.com (mail-wm1-f69.google.com [209.85.128.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-178-EuHabKshNs64qLSp4pJthQ-1; Wed, 07 Jun 2023 06:01:25 -0400 X-MC-Unique: EuHabKshNs64qLSp4pJthQ-1 Received: by mail-wm1-f69.google.com with SMTP id 5b1f17b1804b1-3f7f4a7b289so392165e9.2 for ; Wed, 07 Jun 2023 03:01:25 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132084; x=1688724084; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=mm+/XWRQ+DsnGO4nNmVLBFhfnxnpv17BL9oqvmNRQ7s=; b=LzPfQYiotSG59NdOc4Twxv1j33mBI9+DLbR1KQAYSnHVGgVBblC5WIpMMrYvnBDJHB tGYhp43RBhGoZ4MRr70U0kL2tSKjfPnq9zIVQbeOUJnVvQ0KCPwtqumR1Vyri9ApLgQ5 Ik7/w0oTCjn9B6gOE1UwXY/8vPuu1TqY6hEnAY3nvQe1HL4+jrWzwBqsKOw995tmp0Rp fs+tujOQfOhgW8wUNz0Dss4gKUwH9LaHYcScCd/i44dgjrEOJBIofhYOBRxsIuEq1hx7 SAyg9JiuXH6SPmCNEH7Sjhby9MGuGBVuJMxeIFsyTFda4nM4btd7oRHUHEGGFthmtER5 J98A== X-Gm-Message-State: AC+VfDyHzy4yyZQMvUeWIzVASBi1TzBB3njU0ddkNLtj5FPJuQBs4knC gMggpzouRHNpDXYEFRkHU+qNs1vWzMdqhWnqaLmESWkYFSkujJw6YSPdGnVkTgdn9OpYG69utvO Q0u1bbIvku0Pp5PnZIHn0xbqMJdDG1dOYogOgcd446LRt3h7h+DJf86VUum5IeWpa7vX3B0TTok IlyTq/+Q== X-Received: by 2002:a1c:7215:0:b0:3f5:d0b8:4a53 with SMTP id n21-20020a1c7215000000b003f5d0b84a53mr4779947wmc.34.1686132084346; Wed, 07 Jun 2023 03:01:24 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ7d0OA62foZud7gtjLJFjp3yCBnmLqpg9vxYI8gPtWWO3fx7tUJpjjmnADG+2OkgSSuXdpPPw== X-Received: by 2002:a1c:7215:0:b0:3f5:d0b8:4a53 with SMTP id n21-20020a1c7215000000b003f5d0b84a53mr4779932wmc.34.1686132084052; Wed, 07 Jun 2023 03:01:24 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id k11-20020a5d428b000000b0030aded83385sm15119245wrq.27.2023.06.07.03.01.23 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:23 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess , Tankut Baris Aktemur Subject: [PATCHv8 1/6] Revert "gdb: remove unnecessary parameter wait_ptid from do_target_wait" Date: Wed, 7 Jun 2023 11:01:13 +0100 Message-Id: <487f03fdc1ff646e069e50a839c03b622d270a49.1686131880.git.aburgess@redhat.com> X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" This reverts commit ac0d67ed1dcf470bad6a3bc4800c2ddc9bedecca. There was nothing wrong with the commit which I'm reverting here, but it removed some functionality that will be needed for a later commit; that is, the ability for GDB to ask for events from a specific ptid_t via the do_target_wait function. In a follow up commit, this functionality will be used to implement inferior function calls in multi-threaded inferiors. This is not a straight revert of the above commit. Reverting the above commit replaces a 'nullptr' with 'NULL', I've gone in and changed that, preserving the 'nullptr'. Reviewed-By: Tankut Baris Aktemur --- gdb/infrun.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/gdb/infrun.c b/gdb/infrun.c index 58da1cef29e..db8b918986c 100644 --- a/gdb/infrun.c +++ b/gdb/infrun.c @@ -3881,7 +3881,8 @@ do_target_wait_1 (inferior *inf, ptid_t ptid, more events. Polls for events from all inferiors/targets. */ static bool -do_target_wait (execution_control_state *ecs, target_wait_flags options) +do_target_wait (ptid_t wait_ptid, execution_control_state *ecs, + target_wait_flags options) { int num_inferiors = 0; int random_selector; @@ -3891,9 +3892,10 @@ do_target_wait (execution_control_state *ecs, target_wait_flags options) polling the rest of the inferior list starting from that one in a circular fashion until the whole list is polled once. */ - auto inferior_matches = [] (inferior *inf) + auto inferior_matches = [&wait_ptid] (inferior *inf) { - return inf->process_target () != nullptr; + return (inf->process_target () != nullptr + && ptid_t (inf->pid).matches (wait_ptid)); }; /* First see how many matching inferiors we have. */ @@ -3932,7 +3934,7 @@ do_target_wait (execution_control_state *ecs, target_wait_flags options) auto do_wait = [&] (inferior *inf) { - ecs->ptid = do_target_wait_1 (inf, minus_one_ptid, &ecs->ws, options); + ecs->ptid = do_target_wait_1 (inf, wait_ptid, &ecs->ws, options); ecs->target = inf->process_target (); return (ecs->ws.kind () != TARGET_WAITKIND_IGNORE); }; @@ -4362,7 +4364,7 @@ fetch_inferior_event () the event. */ scoped_disable_commit_resumed disable_commit_resumed ("handling event"); - if (!do_target_wait (&ecs, TARGET_WNOHANG)) + if (!do_target_wait (minus_one_ptid, &ecs, TARGET_WNOHANG)) { infrun_debug_printf ("do_target_wait returned no event"); disable_commit_resumed.reset_and_commit (); From patchwork Wed Jun 7 10:01:14 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70723 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id D1AED3856629 for ; Wed, 7 Jun 2023 10:03:55 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org D1AED3856629 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132235; bh=tbyjVf1wWAntVaSerosYUNGZZO2v3mmmjWInenpv+P0=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=DWvVLHSTO+rFq9YrlIF7rt7NaIGc1qv6Nl1JG8Vve0BY6oB87ZfkDUUa6lUXJW/rD 0BHZHkv6ErD4IMvKjenFaTfd9uio8Kz3afAQNoHGfNqV6exNEybJVG3KkimxqNhuVm yUiTGyzlxQTF5OX0gNiktOtvVoEjGuwQHSqEy7uQ= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by sourceware.org (Postfix) with ESMTPS id A09923857BA4 for ; Wed, 7 Jun 2023 10:01:30 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org A09923857BA4 Received: from mail-wr1-f69.google.com (mail-wr1-f69.google.com [209.85.221.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-394-zbRgjqMxM7a0xNlIU75SjQ-1; Wed, 07 Jun 2023 06:01:29 -0400 X-MC-Unique: zbRgjqMxM7a0xNlIU75SjQ-1 Received: by mail-wr1-f69.google.com with SMTP id ffacd0b85a97d-30479b764f9so3108851f8f.0 for ; Wed, 07 Jun 2023 03:01:28 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132087; x=1688724087; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=tbyjVf1wWAntVaSerosYUNGZZO2v3mmmjWInenpv+P0=; b=ByEGGU7neplIiXs2zDLMJpws5ql1Cz/t9Pc13RJr87A391hX2S2+TZhK2o5OgCBrk5 djH8pnFSlU3W7AarKtPx+8vo0V4gxGPPYkUq1WppSA4GdbUXYRyf8+RJS2hERBrb1bYO NR29jEgewEln4FQJE0YkMpXB4y+FIfn46m+VbdFx9jV71nIADEiUur3kxOrexp8MjEuL ZjIohdnJik59Pu5yqHYXausAlVR3RA5o3CZZDly4ALAFL/HjdIKJXpTKGAgkbzaiQmK1 9U1ncWmcic6Jtpla6u7sqge+jcVSJ2ZYQenhoGjkY1SrZqOehLcf7+UZf5VXsu8xDNia fKKg== X-Gm-Message-State: AC+VfDxp3nZRMQT78Ot3W1cazcuDbTxIHY2rL4z9urvZr8/hNCxEqxHA WCQtiKNX9Rb78QfiyC4brP4NZzdpcAxad4XyDmt7oXq9YbFVvsUfayCrG2sicmEhOfuP7vWPCoZ N465CLYS/w4mcBAyupcOQ0Xdib7Q5s5LXHVSJLudyIidBzwCIh5FSSMhYMndzDxfkaCHl0h+WwI yWmANAGA== X-Received: by 2002:adf:dcc4:0:b0:30a:e66d:25b7 with SMTP id x4-20020adfdcc4000000b0030ae66d25b7mr3836679wrm.13.1686132086871; Wed, 07 Jun 2023 03:01:26 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ7x3Lj6gHONfs4mBLaRc+aOqSjG6WhibvZkeXAyRBC87Jo1OSJkX2LE/j6az8kJvUbsDtxSCg== X-Received: by 2002:adf:dcc4:0:b0:30a:e66d:25b7 with SMTP id x4-20020adfdcc4000000b0030ae66d25b7mr3836617wrm.13.1686132085760; Wed, 07 Jun 2023 03:01:25 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id z6-20020a7bc7c6000000b003f4fb5532a1sm1552034wmk.43.2023.06.07.03.01.24 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:25 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess , Natalia Saiapova , Tankut Baris Aktemur Subject: [PATCHv8 2/6] gdb: fix b/p conditions with infcalls in multi-threaded inferiors Date: Wed, 7 Jun 2023 11:01:14 +0100 Message-Id: X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" This commit fixes bug PR 28942, that is, creating a conditional breakpoint in a multi-threaded inferior, where the breakpoint condition includes an inferior function call. Currently, when a user tries to create such a breakpoint, then GDB will fail with: (gdb) break infcall-from-bp-cond-single.c:61 if (return_true ()) Breakpoint 2 at 0x4011fa: file /tmp/build/gdb/testsuite/../../../src/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.c, line 61. (gdb) continue Continuing. [New Thread 0x7ffff7c5d700 (LWP 2460150)] [New Thread 0x7ffff745c700 (LWP 2460151)] [New Thread 0x7ffff6c5b700 (LWP 2460152)] [New Thread 0x7ffff645a700 (LWP 2460153)] [New Thread 0x7ffff5c59700 (LWP 2460154)] Error in testing breakpoint condition: Couldn't get registers: No such process. An error occurred while in a function called from GDB. Evaluation of the expression containing the function (return_true) will be abandoned. When the function is done executing, GDB will silently stop. Selected thread is running. (gdb) Or, in some cases, like this: (gdb) break infcall-from-bp-cond-simple.c:56 if (is_matching_tid (arg, 1)) Breakpoint 2 at 0x401194: file /tmp/build/gdb/testsuite/../../../src/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.c, line 56. (gdb) continue Continuing. [New Thread 0x7ffff7c5d700 (LWP 2461106)] [New Thread 0x7ffff745c700 (LWP 2461107)] ../../src.release/gdb/nat/x86-linux-dregs.c:146: internal-error: x86_linux_update_debug_registers: Assertion `lwp_is_stopped (lwp)' failed. A problem internal to GDB has been detected, further debugging may prove unreliable. The precise error depends on the exact thread state; so there's race conditions depending on which threads have fully started, and which have not. But the underlying problem is always the same; when GDB tries to execute the inferior function call from within the breakpoint condition, GDB will, incorrectly, try to resume threads that are already running - GDB doesn't realise that some threads might already be running. The solution proposed in this patch requires an additional member variable thread_info::in_cond_eval. This flag is set to true (in breakpoint.c) when GDB is evaluating a breakpoint condition. In user_visible_resume_ptid (infrun.c), when the in_cond_eval flag is true, then GDB will only try to resume the current thread, that is, the thread for which the breakpoint condition is being evaluated. This solves the problem of GDB trying to resume threads that are already running. The next problem is that inferior function calls are assumed to be synchronous, that is, GDB doesn't expect to start an inferior function call in thread #1, then receive a stop from thread #2 for some other, unrelated reason. To prevent GDB responding to an event from another thread, we update fetch_inferior_event and do_target_wait in infrun.c, so that, when an inferior function call (on behalf of a breakpoint condition) is in progress, we only wait for events from the current thread (the one evaluating the condition). In do_target_wait I had to change the inferior_matches lambda function, which is used to select which inferior to wait on. Previously the logic was this: auto inferior_matches = [&wait_ptid] (inferior *inf) { return (inf->process_target () != nullptr && ptid_t (inf->pid).matches (wait_ptid)); }; This compares the pid of the inferior against the complete ptid we want to wait on. Before this commit wait_ptid was only ever minus_one_ptid (which is special, and means any process), and so every inferior would match. After this commit though wait_ptid might represent a specific thread in a specific inferior. If we compare the pid of the inferior to a specific ptid then these will not match. The fix is to compare against the pid extracted from the wait_ptid, not against the complete wait_ptid itself. In fetch_inferior_event, after receiving the event, we only want to stop all the other threads, and call inferior_event_handler with INF_EXEC_COMPLETE, if we are not evaluating a conditional breakpoint. If we are, then all the other threads should be left doing whatever they were before. The inferior_event_handler call will be performed once the breakpoint condition has finished being evaluated, and GDB decides to stop or not. The final problem that needs solving relates to GDB's commit-resume mechanism, which allows GDB to collect resume requests into a single packet in order to reduce traffic to a remote target. The problem is that the commit-resume mechanism will not send any resume requests for an inferior if there are already events pending on the GDB side. Imagine an inferior with two threads. Both threads hit a breakpoint, maybe the same conditional breakpoint. At this point there are two pending events, one for each thread. GDB selects one of the events and spots that this is a conditional breakpoint, GDB evaluates the condition. The condition includes an inferior function call, so GDB sets up for the call and resumes the one thread, the resume request is added to the commit-resume queue. When the commit-resume queue is committed GDB sees that there is a pending event from another thread, and so doesn't send any resume requests to the actual target, GDB is assuming that when we wait we will select the event from the other thread. However, as this is an inferior function call for a condition evaluation, we will not select the event from the other thread, we only care about events from the thread that is evaluating the condition - and the resume for this thread was never sent to the target. And so, GDB hangs, waiting for an event from a thread that was never fully resumed. To fix this issue I have added the concept of "forcing" the commit-resume queue. When enabling commit resume, if the force flag is true, then any resumes will be committed to the target, even if there are other threads with pending events. A note on authorship: this patch was based on some work done by Natalia Saiapova and Tankut Baris Aktemur from Intel[1]. I have made some changes to their work in this version. Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28942 [1] https://sourceware.org/pipermail/gdb-patches/2020-October/172454.html Co-authored-by: Natalia Saiapova Co-authored-by: Tankut Baris Aktemur Reviewed-By: Tankut Baris Aktemur --- gdb/breakpoint.c | 2 + gdb/gdbthread.h | 3 + gdb/infcall.c | 6 + gdb/infrun.c | 64 +++-- gdb/infrun.h | 3 +- .../infcall-from-bp-cond-other-thread-event.c | 135 ++++++++++ ...nfcall-from-bp-cond-other-thread-event.exp | 174 +++++++++++++ .../gdb.threads/infcall-from-bp-cond-simple.c | 89 +++++++ .../infcall-from-bp-cond-simple.exp | 235 ++++++++++++++++++ .../gdb.threads/infcall-from-bp-cond-single.c | 139 +++++++++++ .../infcall-from-bp-cond-single.exp | 117 +++++++++ 11 files changed, 952 insertions(+), 15 deletions(-) create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.c create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.exp create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.c create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.c create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.exp diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index da6c8de9d14..592aa29d949 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -5540,6 +5540,8 @@ bpstat_check_breakpoint_conditions (bpstat *bs, thread_info *thread) { try { + scoped_restore reset_in_cond_eval + = make_scoped_restore (&thread->control.in_cond_eval, true); condition_result = breakpoint_cond_eval (cond); } catch (const gdb_exception_error &ex) diff --git a/gdb/gdbthread.h b/gdb/gdbthread.h index 7135515bf45..897752f2691 100644 --- a/gdb/gdbthread.h +++ b/gdb/gdbthread.h @@ -171,6 +171,9 @@ struct thread_control_state command. This is used to decide whether "set scheduler-locking step" behaves like "on" or "off". */ int stepping_command = 0; + + /* True if the thread is evaluating a BP condition. */ + bool in_cond_eval = false; }; /* Inferior thread specific part of `struct infcall_suspend_state'. */ diff --git a/gdb/infcall.c b/gdb/infcall.c index ddf325a62a5..b055fbfbd64 100644 --- a/gdb/infcall.c +++ b/gdb/infcall.c @@ -642,6 +642,12 @@ run_inferior_call (std::unique_ptr sm, proceed (real_pc, GDB_SIGNAL_0); + /* Enable commit resume, but pass true for the force flag. This + ensures any thread we set running in proceed will actually be + committed to the target, even if some other thread in the current + target has a pending event. */ + scoped_enable_commit_resumed enable ("infcall", true); + infrun_debug_show_threads ("non-exited threads after proceed for inferior-call", all_non_exited_threads ()); diff --git a/gdb/infrun.c b/gdb/infrun.c index db8b918986c..d2c17433104 100644 --- a/gdb/infrun.c +++ b/gdb/infrun.c @@ -2276,6 +2276,14 @@ user_visible_resume_ptid (int step) mode. */ resume_ptid = inferior_ptid; } + else if (inferior_ptid != null_ptid + && inferior_thread ()->control.in_cond_eval) + { + /* The inferior thread is evaluating a BP condition. Other threads + might be stopped or running and we do not want to change their + state, thus, resume only the current thread. */ + resume_ptid = inferior_ptid; + } else if (!sched_multi && target_supports_multi_process ()) { /* Resume all threads of the current process (and none of other @@ -3004,12 +3012,24 @@ schedlock_applies (struct thread_info *tp) execution_direction))); } -/* Set process_stratum_target::COMMIT_RESUMED_STATE in all target - stacks that have threads executing and don't have threads with - pending events. */ +/* When FORCE_P is false, set process_stratum_target::COMMIT_RESUMED_STATE + in all target stacks that have threads executing and don't have threads + with pending events. + + When FORCE_P is true, set process_stratum_target::COMMIT_RESUMED_STATE + in all target stacks that have threads executing regardless of whether + there are pending events or not. + + Passing FORCE_P as false makes sense when GDB is going to wait for + events from all threads and will therefore spot the pending events. + However, if GDB is only going to wait for events from select threads + (i.e. when performing an inferior call) then a pending event on some + other thread will not be spotted, and if we fail to commit the resume + state for the thread performing the inferior call, then the inferior + call will never complete (or even start). */ static void -maybe_set_commit_resumed_all_targets () +maybe_set_commit_resumed_all_targets (bool force_p) { scoped_restore_current_thread restore_thread; @@ -3038,7 +3058,7 @@ maybe_set_commit_resumed_all_targets () status to report, handle it before requiring the target to commit its resumed threads: handling the status might lead to resuming more threads. */ - if (proc_target->has_resumed_with_pending_wait_status ()) + if (!force_p && proc_target->has_resumed_with_pending_wait_status ()) { infrun_debug_printf ("not requesting commit-resumed for target %s, a" " thread has a pending waitstatus", @@ -3048,7 +3068,7 @@ maybe_set_commit_resumed_all_targets () switch_to_inferior_no_thread (inf); - if (target_has_pending_events ()) + if (!force_p && target_has_pending_events ()) { infrun_debug_printf ("not requesting commit-resumed for target %s, " "target has pending events", @@ -3141,7 +3161,7 @@ scoped_disable_commit_resumed::reset () { /* This is the outermost instance, re-enable COMMIT_RESUMED_STATE on the targets where it's possible. */ - maybe_set_commit_resumed_all_targets (); + maybe_set_commit_resumed_all_targets (false); } else { @@ -3174,7 +3194,7 @@ scoped_disable_commit_resumed::reset_and_commit () /* See infrun.h. */ scoped_enable_commit_resumed::scoped_enable_commit_resumed - (const char *reason) + (const char *reason, bool force_p) : m_reason (reason), m_prev_enable_commit_resumed (enable_commit_resumed) { @@ -3186,7 +3206,7 @@ scoped_enable_commit_resumed::scoped_enable_commit_resumed /* Re-enable COMMIT_RESUMED_STATE on the targets where it's possible. */ - maybe_set_commit_resumed_all_targets (); + maybe_set_commit_resumed_all_targets (force_p); maybe_call_commit_resumed_all_targets (); } @@ -3892,10 +3912,11 @@ do_target_wait (ptid_t wait_ptid, execution_control_state *ecs, polling the rest of the inferior list starting from that one in a circular fashion until the whole list is polled once. */ - auto inferior_matches = [&wait_ptid] (inferior *inf) + ptid_t wait_ptid_pid {wait_ptid.pid ()}; + auto inferior_matches = [&wait_ptid_pid] (inferior *inf) { return (inf->process_target () != nullptr - && ptid_t (inf->pid).matches (wait_ptid)); + && ptid_t (inf->pid).matches (wait_ptid_pid)); }; /* First see how many matching inferiors we have. */ @@ -4364,7 +4385,17 @@ fetch_inferior_event () the event. */ scoped_disable_commit_resumed disable_commit_resumed ("handling event"); - if (!do_target_wait (minus_one_ptid, &ecs, TARGET_WNOHANG)) + /* Is the current thread performing an inferior function call as part + of a breakpoint condition evaluation? */ + bool in_cond_eval = (inferior_ptid != null_ptid + && inferior_thread ()->control.in_cond_eval); + + /* If the thread is in the middle of the condition evaluation, wait for + an event from the current thread. Otherwise, wait for an event from + any thread. */ + ptid_t waiton_ptid = in_cond_eval ? inferior_ptid : minus_one_ptid; + + if (!do_target_wait (waiton_ptid, &ecs, TARGET_WNOHANG)) { infrun_debug_printf ("do_target_wait returned no event"); disable_commit_resumed.reset_and_commit (); @@ -4420,7 +4451,12 @@ fetch_inferior_event () bool should_notify_stop = true; bool proceeded = false; - stop_all_threads_if_all_stop_mode (); + /* If the thread that stopped just completed an inferior + function call as part of a condition evaluation, then we + don't want to stop all the other threads. */ + if (ecs.event_thread == nullptr + || !ecs.event_thread->control.in_cond_eval) + stop_all_threads_if_all_stop_mode (); clean_up_just_stopped_threads_fsms (&ecs); @@ -4435,7 +4471,7 @@ fetch_inferior_event () proceeded = normal_stop (); } - if (!proceeded) + if (!proceeded && !in_cond_eval) { inferior_event_handler (INF_EXEC_COMPLETE); cmd_done = 1; diff --git a/gdb/infrun.h b/gdb/infrun.h index a343d27f72d..d8db153ecc6 100644 --- a/gdb/infrun.h +++ b/gdb/infrun.h @@ -406,7 +406,8 @@ extern void maybe_call_commit_resumed_all_targets (); struct scoped_enable_commit_resumed { - explicit scoped_enable_commit_resumed (const char *reason); + explicit scoped_enable_commit_resumed (const char *reason, + bool force_p = false); ~scoped_enable_commit_resumed (); DISABLE_COPY_AND_ASSIGN (scoped_enable_commit_resumed); diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.c b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.c new file mode 100644 index 00000000000..e2a8ccb4ebe --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.c @@ -0,0 +1,135 @@ +/* Copyright 2022-2023 Free Software Foundation, Inc. + + This file is part of GDB. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include +#include +#include +#include + +#define NUM_THREADS 2 + +pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; + +/* Some global variables to poke, just for something to do. */ +volatile int global_var_0 = 0; +volatile int global_var_1 = 0; + +/* This flag is updated from GDB. */ +volatile int raise_signal = 0; + +/* Implement the breakpoint condition function. Release the other thread + and try to give the other thread a chance to run. Then return ANSWER. */ +int +condition_core_func (int answer) +{ + /* This unlock should release the other thread. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + + /* And this yield and sleep should (hopefully) give the other thread a + chance to run. This isn't guaranteed of course, but once the other + thread does run it should hit a breakpoint, which GDB should + (temporarily) ignore, so there's no easy way for us to know the other + thread has done what it needs to, thus, yielding and sleeping is the + best we can do. */ + sched_yield (); + sleep (2); + + return answer; +} + +void +stop_marker () +{ + int a = 100; /* Final breakpoint here. */ +} + +/* A breakpoint condition function that always returns true. */ +int +condition_true_func () +{ + return condition_core_func (1); +} + +/* A breakpoint condition function that always returns false. */ +int +condition_false_func () +{ + return condition_core_func (0); +} + +void * +worker_func (void *arg) +{ + volatile int *ptr = 0; + int tid = *((int *) arg); + + switch (tid) + { + case 0: + global_var_0 = 11; /* First thread breakpoint. */ + break; + + case 1: + if (pthread_mutex_lock (&mutex) != 0) + abort (); + if (raise_signal) + global_var_1 = *ptr; /* Signal here. */ + else + global_var_1 = 99; /* Other thread breakpoint. */ + break; + + default: + abort (); + } + + return NULL; +} + +int +main () +{ + pthread_t threads[NUM_THREADS]; + int args[NUM_THREADS]; + + /* Set an alarm, just in case the test deadlocks. */ + alarm (300); + + /* We want the mutex to start locked. */ + if (pthread_mutex_lock (&mutex) != 0) + abort (); + + for (int i = 0; i < NUM_THREADS; i++) + { + args[i] = i; + pthread_create (&threads[i], NULL, worker_func, &args[i]); + } + + for (int i = 0; i < NUM_THREADS; i++) + { + void *retval; + pthread_join (threads[i], &retval); + } + + /* Unlock once we're done, just for cleanliness. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + + stop_marker (); + + return 0; +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.exp new file mode 100644 index 00000000000..1453575e4ad --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-other-thread-event.exp @@ -0,0 +1,174 @@ +# Copyright 2022-2023 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# Test for conditional breakpoints where the breakpoint condition includes +# an inferior function call. +# +# The tests in this script are testing what happens when an event arrives in +# another thread while GDB is waiting for the inferior function call (in the +# breakpoint condition) to finish. +# +# The expectation is that GDB will queue events for other threads and wait +# for the inferior function call to complete, if the condition is true, then +# the conditional breakpoint should be reported first. The other thread +# event should of course, not get lost, and should be reported as soon as +# the user tries to continue the inferior. +# +# If the conditional breakpoint ends up not being taken (the condition is +# false), then the other thread event should be reported immediately. +# +# This script tests what happens when the other thread event is (a) the +# other thread hitting a breakpoint, and (b) the other thread taking a +# signal (SIGSEGV in this case). + +standard_testfile + +if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ + {debug pthreads}] == -1 } { + return +} + +set cond_bp_line [gdb_get_line_number "First thread breakpoint"] +set other_bp_line [gdb_get_line_number "Other thread breakpoint"] +set final_bp_line [gdb_get_line_number "Final breakpoint here"] +set signal_line [gdb_get_line_number "Signal here"] + +# Start GDB based on TARGET_ASYNC and TARGET_NON_STOP, and then runto main. +proc start_gdb_and_runto_main { target_async target_non_stop } { + save_vars { ::GDBFLAGS } { + append ::GDBFLAGS \ + " -ex \"maint set target-non-stop $target_non_stop\"" + append ::GDBFLAGS \ + " -ex \"maintenance set target-async ${target_async}\"" + + clean_restart ${::binfile} + } + + if { ![runto_main] } { + return -1 + } + + return 0 +} + +# Run a test of GDB's conditional breakpoints, where the conditions include +# inferior function calls. While the inferior function call is executing +# another thread will hit a breakpoint (when OTHER_THREAD_SIGNAL is false), +# or receive a signal (when OTHER_THREAD_SIGNAL is true). GDB should report +# the conditional breakpoint first (if the condition is true), and then +# report the second thread event once the inferior is continued again. +# +# When STOP_AT_COND is true then the conditional breakpoint will have a +# condition that evaluates to true (and GDB will stop at the breakpoint), +# otherwise, the condition will evaluate to false (and GDB will not stop at +# the breakpoint). +proc run_condition_test { stop_at_cond other_thread_signal \ + target_async target_non_stop } { + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # Setup the conditional breakpoint. + if { $stop_at_cond } { + set cond_func "condition_true_func" + } else { + set cond_func "condition_false_func" + } + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (${cond_func} ())" + set cond_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for conditional breakpoint"] + + if { $other_thread_signal } { + # Arrange for the other thread to raise a signal while GDB is + # evaluating the breakpoint condition. + gdb_test_no_output "set raise_signal = 1" + } else { + # And a breakpoint that will be hit by another thread only once the + # breakpoint condition starts to be evaluated. + gdb_breakpoint "${::srcfile}:${::other_bp_line}" + set other_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for other breakpoint"] + } + + # A final breakpoint once the test has completed. + gdb_breakpoint "${::srcfile}:${::final_bp_line}" + set final_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for final breakpoint"] + + if { $stop_at_cond } { + # Continue. The first breakpoint we hit should be the conditional + # breakpoint. The other thread will have hit its breakpoint, but + # that will have been deferred until the conditional breakpoint is + # reported. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${cond_bp_num}, worker_func \[^\r\n\]+:${::cond_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+First thread breakpoint\[^\r\n\]+"] \ + "hit the conditional breakpoint" + } + + if { $other_thread_signal } { + # Now continue again, the other thread will now report that it + # received a signal. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" received signal SIGSEGV, Segmentation fault\\." \ + "\\\[Switching to Thread \[^\r\n\]+\\\]" \ + "${::hex} in worker_func \[^\r\n\]+:${::signal_line}" \ + "${::decimal}\\s+\[^\r\n\]+Signal here\[^\r\n\]+"] \ + "received signal in other thread" + } else { + # Now continue again, the other thread will now report its + # breakpoint. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${other_bp_num}, worker_func \[^\r\n\]+:${::other_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+Other thread breakpoint\[^\r\n\]+"] \ + "hit the breakpoint in other thread" + + # Run to the stop marker. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${final_bp_num}, stop_marker \[^\r\n\]+:${::final_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+Final breakpoint here\[^\r\n\]+"] \ + "hit the final breakpoint" + } + + gdb_exit +} + +foreach_with_prefix target_async { "on" "off" } { + foreach_with_prefix target_non_stop { "on" "off" } { + foreach_with_prefix other_thread_signal { true false } { + foreach_with_prefix stop_at_cond { true false } { + run_condition_test $stop_at_cond $other_thread_signal \ + $target_async $target_non_stop + } + } + } +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.c b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.c new file mode 100644 index 00000000000..9d746d8be49 --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.c @@ -0,0 +1,89 @@ +/* Copyright 2022-2023 Free Software Foundation, Inc. + + This file is part of GDB. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include +#include + +#define NUM_THREADS 3 + +int +is_matching_tid (int *tid_ptr, int tid_value) +{ + return *tid_ptr == tid_value; +} + +int +return_true () +{ + return 1; +} + +int +return_false () +{ + return 0; +} + +int +function_that_segfaults () +{ + int *p = 0; + *p = 1; /* Segfault happens here. */ +} + +int +function_with_breakpoint () +{ + return 1; /* Nested breakpoint. */ +} + +void * +worker_func (void *arg) +{ + int a = 42; /* Breakpoint here. */ +} + +void +stop_marker () +{ + int b = 99; /* Stop marker. */ +} + +int +main () +{ + pthread_t threads[NUM_THREADS]; + int args[NUM_THREADS]; + + alarm (300); + + for (int i = 0; i < NUM_THREADS; i++) + { + args[i] = i; + pthread_create (&threads[i], NULL, worker_func, &args[i]); + } + + for (int i = 0; i < NUM_THREADS; i++) + { + void *retval; + pthread_join (threads[i], &retval); + } + + stop_marker (); + + return 0; +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp new file mode 100644 index 00000000000..c46f8cdb981 --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp @@ -0,0 +1,235 @@ +# Copyright 2022-2023 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# Some simple tests of inferior function calls from breakpoint +# conditions, in multi-threaded inferiors. +# +# This test sets up a multi-threaded inferior, and places a breakpoint +# at a location that many of the threads will reach. We repeat the +# test with different conditions, sometimes a single thread should +# stop at the breakpoint, sometimes multiple threads should stop, and +# sometimes no threads should stop. + +standard_testfile + +if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ + {debug pthreads}] == -1 } { + return +} + +set cond_bp_line [gdb_get_line_number "Breakpoint here"] +set stop_bp_line [gdb_get_line_number "Stop marker"] +set nested_bp_line [gdb_get_line_number "Nested breakpoint"] +set segv_line [gdb_get_line_number "Segfault happens here"] + +# Start GDB based on TARGET_ASYNC and TARGET_NON_STOP, and then runto main. +proc start_gdb_and_runto_main { target_async target_non_stop } { + save_vars { ::GDBFLAGS } { + append ::GDBFLAGS \ + " -ex \"maint set target-non-stop $target_non_stop\"" + append ::GDBFLAGS \ + " -ex \"maintenance set target-async ${target_async}\"" + + clean_restart ${::binfile} + } + + if { ![runto_main] } { + return -1 + } + + return 0 +} + +# Run a test of GDB's conditional breakpoints, where the conditions include +# inferior function calls. +# +# CONDITION is the expression to be used as the breakpoint condition. +# +# N_EXPECTED_HITS is the number of threads that we expect to stop due to +# CONDITON. +# +# MESSAGE is used as a test name prefix. +proc run_condition_test { message n_expected_hits condition \ + target_async target_non_stop } { + with_test_prefix $message { + + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # Use this convenience variable to track how often the + # breakpoint condition has been evaluated. This should be + # once per thread. + gdb_test "set \$n_cond_eval = 0" + + # Setup the conditional breakpoint. + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if ((++\$n_cond_eval) && (${condition}))" + + # And a breakpoint that we hit when the test is over, this one is + # not conditional. Only the main thread gets here once all the + # other threads have finished. + gdb_breakpoint "${::srcfile}:${::stop_bp_line}" + + # The number of times we stop at the conditional breakpoint. + set n_hit_condition 0 + + # Now keep 'continue'-ing GDB until all the threads have finished + # and we reach the stop_marker breakpoint. + gdb_test_multiple "continue" "spot all breakpoint hits" { + -re " worker_func \[^\r\n\]+${::srcfile}:${::cond_bp_line}\r\n${::decimal}\\s+\[^\r\n\]+Breakpoint here\[^\r\n\]+\r\n${::gdb_prompt} $" { + incr n_hit_condition + send_gdb "continue\n" + exp_continue + } + + -re " stop_marker \[^\r\n\]+${::srcfile}:${::stop_bp_line}\r\n${::decimal}\\s+\[^\r\n\]+Stop marker\[^\r\n\]+\r\n${::gdb_prompt} $" { + pass $gdb_test_name + } + } + + gdb_assert { $n_hit_condition == $n_expected_hits } \ + "stopped at breakpoint the expected number of times" + + # Ensure the breakpoint condition was evaluated once per thread. + gdb_test "print \$n_cond_eval" "= 3" \ + "condition was evaluated in each thread" + } +} + +# Check that after handling a conditional breakpoint (where the condition +# includes an inferior call), it is still possible to kill the running +# inferior, and then restart the inferior. +# +# At once point doing this would result in GDB giving an assertion error. +proc_with_prefix run_kill_and_restart_test { target_async target_non_stop } { + # This test relies on the 'start' command, which is not possible with + # the plain 'remote' target. + if { [target_info gdb_protocol] == "remote" } { + return + } + + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # Setup the conditional breakpoint. + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (is_matching_tid (arg, 1))" + gdb_continue_to_breakpoint "worker_func" + + # Now kill the program being debugged. + gdb_test "kill" "" "kill process" \ + "Kill the program being debugged.*y or n. $" "y" + + # Check we can restart the inferior. At one point this would trigger an + # assertion. + gdb_start_cmd +} + +# Create a conditional breakpoint which includes a call to a function that +# segfaults. Run GDB and check what happens when the inferior segfaults +# during the inferior call. +proc_with_prefix run_bp_cond_segfaults { target_async target_non_stop } { + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # This test relies on the inferior segfaulting when trying to + # access address zero. + if { [is_address_zero_readable] } { + return + } + + # Setup the conditional breakpoint, include a call to + # 'function_that_segfaults', which triggers the segfault. + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (is_matching_tid (arg, 0) && function_that_segfaults ())" + set bp_1_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number of conditional breakpoint"] + + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "Thread ${::decimal} \"infcall-from-bp\" received signal SIGSEGV, Segmentation fault\\." \ + "${::hex} in function_that_segfaults \\(\\) at \[^\r\n\]+:${::segv_line}" \ + "${::decimal}\\s+\[^\r\n\]+Segfault happens here\[^\r\n\]+" \ + "Error in testing condition for breakpoint ${bp_1_num}:" \ + "The program being debugged was signaled while in a function called from GDB\\." \ + "GDB remains in the frame where the signal was received\\." \ + "To change this behavior use \"set unwindonsignal on\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(function_that_segfaults\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] +} + +# Create a conditional breakpoint which includes a call to a function that +# itself has a breakpoint set within it. Run GDB and check what happens +# when GDB hits the nested breakpoint. +proc_with_prefix run_bp_cond_hits_breakpoint { target_async target_non_stop } { + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # Setup the conditional breakpoint, include a call to + # 'function_with_breakpoint' in which we will shortly place a + # breakpoint. + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (is_matching_tid (arg, 0) && function_with_breakpoint ())" + set bp_1_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number of conditional breakpoint"] + + gdb_breakpoint "${::srcfile}:${::nested_bp_line}" + set bp_2_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number of nested breakpoint"] + + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "Thread ${::decimal} \"infcall-from-bp\" hit Breakpoint ${bp_2_num}, function_with_breakpoint \\(\\) at \[^\r\n\]+:${::nested_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+Nested breakpoint\[^\r\n\]+" \ + "Error in testing condition for breakpoint ${bp_1_num}:" \ + "The program being debugged stopped while in a function called from GDB\\." \ + "Evaluation of the expression containing the function" \ + "\\(function_with_breakpoint\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] +} + +foreach_with_prefix target_async { "on" "off" } { + foreach_with_prefix target_non_stop { "on" "off" } { + run_condition_test "exactly one thread is hit" \ + 1 "is_matching_tid (arg, 1)" \ + $target_async $target_non_stop + run_condition_test "exactly two threads are hit" \ + 2 "(is_matching_tid (arg, 0) || is_matching_tid (arg, 2))" \ + $target_async $target_non_stop + run_condition_test "all three threads are hit" \ + 3 "return_true ()" \ + $target_async $target_non_stop + run_condition_test "no thread is hit" \ + 0 "return_false ()" \ + $target_async $target_non_stop + + run_kill_and_restart_test $target_async $target_non_stop + run_bp_cond_segfaults $target_async $target_non_stop + run_bp_cond_hits_breakpoint $target_async $target_non_stop + } +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.c b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.c new file mode 100644 index 00000000000..835c72f03cf --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.c @@ -0,0 +1,139 @@ +/* Copyright 2022-2023 Free Software Foundation, Inc. + + This file is part of GDB. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include +#include +#include +#include + +#define NUM_THREADS 5 + +/* Semaphores, used to track when threads have started, and to control + when the threads finish. */ +sem_t startup_semaphore; +sem_t finish_semaphore; + +/* Mutex to control when the first worker thread hit a breakpoint + location. */ +pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; + +/* Global variable to poke, just so threads have something to do. */ +volatile int global_var = 0; + +int +return_true () +{ + return 1; +} + +int +return_false () +{ + return 0; +} + +void * +worker_func (void *arg) +{ + int tid = *((int *) arg); + + switch (tid) + { + case 0: + /* Wait for MUTEX to become available, then pass through the + conditional breakpoint location. */ + if (pthread_mutex_lock (&mutex) != 0) + abort (); + global_var = 99; /* Conditional breakpoint here. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + break; + + default: + /* Notify the main thread that the thread has started, then wait for + the main thread to tell us to finish. */ + sem_post (&startup_semaphore); + if (sem_wait (&finish_semaphore) != 0) + abort (); + break; + } +} + +void +stop_marker () +{ + global_var = 99; /* Stop marker. */ +} + +int +main () +{ + pthread_t threads[NUM_THREADS]; + int args[NUM_THREADS]; + void *retval; + + /* An alarm, just in case the thread deadlocks. */ + alarm (300); + + /* Semaphore initialization. */ + if (sem_init (&startup_semaphore, 0, 0) != 0) + abort (); + if (sem_init (&finish_semaphore, 0, 0) != 0) + abort (); + + /* Lock MUTEX, this prevents the first worker thread from rushing ahead. */ + if (pthread_mutex_lock (&mutex) != 0) + abort (); + + /* Worker thread creation. */ + for (int i = 0; i < NUM_THREADS; i++) + { + args[i] = i; + pthread_create (&threads[i], NULL, worker_func, &args[i]); + } + + /* Wait for every thread (other than the first) to tell us it has started + up. */ + for (int i = 1; i < NUM_THREADS; i++) + { + if (sem_wait (&startup_semaphore) != 0) + abort (); + } + + /* Unlock the first thread so it can proceed. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + + /* Wait for the first thread only. */ + pthread_join (threads[0], &retval); + + /* Now post FINISH_SEMAPHORE to allow all the other threads to finish. */ + for (int i = 1; i < NUM_THREADS; i++) + sem_post (&finish_semaphore); + + /* Now wait for the remaining threads to complete. */ + for (int i = 1; i < NUM_THREADS; i++) + pthread_join (threads[i], &retval); + + /* Semaphore cleanup. */ + sem_destroy (&finish_semaphore); + sem_destroy (&startup_semaphore); + + stop_marker (); + + return 0; +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.exp new file mode 100644 index 00000000000..787dee3aa8e --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-single.exp @@ -0,0 +1,117 @@ +# Copyright 2022-2023 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# This test reprocuces bug gdb/28942, performing an inferior function +# call from a breakpoint condition in a multi-threaded inferior. +# +# The important part of this test is that, when the conditional +# breakpoint is hit, and the condition (which includes an inferior +# function call) is evaluated, the other threads are running. + +standard_testfile + +if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ + {debug pthreads}] == -1 } { + return +} + +set cond_bp_line [gdb_get_line_number "Conditional breakpoint here"] +set final_bp_line [gdb_get_line_number "Stop marker"] + +# Start GDB based on TARGET_ASYNC and TARGET_NON_STOP, and then runto main. +proc start_gdb_and_runto_main { target_async target_non_stop } { + save_vars { ::GDBFLAGS } { + append ::GDBFLAGS \ + " -ex \"maint set target-non-stop $target_non_stop\"" + append ::GDBFLAGS \ + " -ex \"maintenance set target-async ${target_async}\"" + + clean_restart ${::binfile} + } + + if { ![runto_main] } { + return -1 + } + + return 0 +} + +# Run a test of GDB's conditional breakpoints, where the conditions include +# inferior function calls. +# +# TARGET_ASYNC and TARGET_NON_STOP are used when starting up GDB. +# +# When STOP_AT_COND is true the breakpoint condtion will evaluate to +# true, and GDB will stop at the breakpoint. Otherwise, the +# breakpoint condition will evaluate to false and GDB will not stop at +# the breakpoint. +proc run_condition_test { stop_at_cond \ + target_async target_non_stop } { + if { [start_gdb_and_runto_main $target_async \ + $target_non_stop] == -1 } { + return + } + + # Setup the conditional breakpoint. + if { $stop_at_cond } { + set cond_func "return_true" + } else { + set cond_func "return_false" + } + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (${cond_func} ())" + set cond_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for conditional breakpoint"] + + # And a breakpoint that we hit when the test is over, this one is + # not conditional. + gdb_breakpoint "${::srcfile}:${::final_bp_line}" + set final_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for final breakpoint"] + + if { $stop_at_cond } { + # Continue. The first breakpoint we hit should be the conditional + # breakpoint. The other thread will have hit its breakpoint, but + # that will have been deferred until the conditional breakpoint is + # reported. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${cond_bp_num}, worker_func \[^\r\n\]+:${::cond_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+Conditional breakpoint here\[^\r\n\]+"] \ + "hit the conditional breakpoint" + } + + # Run to the stop marker. + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${final_bp_num}, stop_marker \[^\r\n\]+:${::final_bp_line}" \ + "${::decimal}\\s+\[^\r\n\]+Stop marker\[^\r\n\]+"] \ + "hit the final breakpoint" +} + +foreach_with_prefix target_async { "on" "off" } { + foreach_with_prefix target_non_stop { "on" "off" } { + foreach_with_prefix stop_at_cond { true false } { + run_condition_test $stop_at_cond \ + $target_async $target_non_stop + } + } +} From patchwork Wed Jun 7 10:01:15 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70721 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 6CB3538555A1 for ; Wed, 7 Jun 2023 10:03:08 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 6CB3538555A1 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132188; bh=yU3L8PiaClQk74plCQ29L17VKrv0PArZFfdYSrw8yAk=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=Hpb8nAwKhIPxy/0MtyFCdC0VhV/WKyBO+1uAdIQnKm0lBiqp3zg2aUG8Wq29GH9rv SGFc886FAsYxTnCG7aH6YeJo2ymUglHC+IWK0eF96V/qa5B9lpXxFLP7/WC9TMY8HN vEzOcwDnUZt397++kYeNwRVnvavei3XqMDzocq+w= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id A0449385772E for ; Wed, 7 Jun 2023 10:01:32 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org A0449385772E Received: from mail-wr1-f69.google.com (mail-wr1-f69.google.com [209.85.221.69]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-595-ID__NiK9NMORdoiN18pSzg-1; Wed, 07 Jun 2023 06:01:31 -0400 X-MC-Unique: ID__NiK9NMORdoiN18pSzg-1 Received: by mail-wr1-f69.google.com with SMTP id ffacd0b85a97d-30d4704fcbaso2591764f8f.0 for ; Wed, 07 Jun 2023 03:01:30 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132089; x=1688724089; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=yU3L8PiaClQk74plCQ29L17VKrv0PArZFfdYSrw8yAk=; b=cL1gQktnQzQAqy9HAZRYvgU6x+QFs5udS3JJkmKjBgvJuNrxSNYOQF5f/fpZ3RLEyn oUcTloQBp/pTvo+pMVheLyTYxv+FIKhYF5UZn8qCp9fuAAmT0vLi6foVoNRYfniUEeVh pmygyeVImAYRvkZVVo4ch+86gimNwMBzmRpmqSZcSt9QikYYUMQrS3cgv6M8hI9qFfMX /vlrsBHWbxXPl5redQxYQO6y4FaWiJGja+OpC4GA3cDsYA9S6/Cmtb8c3Oqb0mHPaFRV Ss9V99iaU1We12XqCLPqJGH8MlNVkW3+Fctn41JlbrJ/dgWDoodnjJuUhG/1wvX+vwmr l3jA== X-Gm-Message-State: AC+VfDz6DT5RoqiHnmQhImsUA/KeoAfYN9kKdXFVhR2Xpq8fMTxWCRNj h9nPNbekqtMczXYYmxjJIlzoVSEqg7GBUpK94phbruVTV4cjhr1wvw1+KbsdJehA4m8VC87/p+o ffM7RstZfz2/1PemZJ1KCEvHYApFk2wUGpuME5eFuDG7QTDYApvKyVp4UE0LZ6wmaelh83McMIh L+Qp9NEw== X-Received: by 2002:adf:efc9:0:b0:30d:1df7:f145 with SMTP id i9-20020adfefc9000000b0030d1df7f145mr6357139wrp.64.1686132088500; Wed, 07 Jun 2023 03:01:28 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ6VXItsh4t9Me6KfC9kKblK39IKd5S/RiuBd6aG+QthGLEQ3bA3YqYyKI9kdXpsitjWEpKIpA== X-Received: by 2002:adf:efc9:0:b0:30d:1df7:f145 with SMTP id i9-20020adfefc9000000b0030d1df7f145mr6357059wrp.64.1686132087244; Wed, 07 Jun 2023 03:01:27 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id i8-20020adfdec8000000b0030796e103a1sm15145823wrn.5.2023.06.07.03.01.26 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:26 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess , Tankut Baris Aktemur Subject: [PATCHv8 3/6] gdb: add timeouts for inferior function calls Date: Wed, 7 Jun 2023 11:01:15 +0100 Message-Id: <3e0f4872a0f6a6ff4c1d3e61d69ef5988f684e06.1686131880.git.aburgess@redhat.com> X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" In the previous commits I have been working on improving inferior function call support. One thing that worries me about using inferior function calls from a conditional breakpoint is: what happens if the inferior function call fails? If the failure is obvious, e.g. the thread performing the call crashes, or hits a breakpoint, then this case is already well handled, and the error is reported to the user. But what if the thread performing the inferior call just deadlocks? If the user made the call from a 'print' or 'call' command, then the user might have some expectation of when the function call should complete, and, when this time limit is exceeded, the user will (hopefully) interrupt GDB and regain control of the debug session. But, when the inferior function call is from a breakpoint condition it is much harder to understand that GDB is deadlocked within an inferior call. Maybe the breakpoint hasn't been hit yet? Or maybe the condition was always false? Or maybe GDB is deadlocked in an inferior call? The only way to know for sure is for the user to periodically interrupt the inferior, check on the state of all the threads, and then continue. Additionally, the focus of the previous commit was inferior function calls, from a conditional breakpoint, in a multi-threaded inferior. This opens up a whole new set of potential failure conditions. For example, what if the function called relies on interaction with some other thread, and the other thread crashes? Or hits a breakpoint? Given how inferior function calls work (in a synchronous manner), a stop event in some other thread is going to be ignored while the inferior function call is being executed as part of a breakpoint condition, and this means that GDB could get stuck waiting for the original condition thread, which will now never complete. In this commit I propose a solution to this problem. A timeout. For targets that support async-mode we can install an event-loop timer before starting the inferior function call. When the timer expires we will stop the thread performing the inferior function call. With this mechanism in place a user can be sure that any inferior call they make will either complete, or timeout eventually. Adding a timer like this is obviously a change in behaviour for the more common 'call' and 'print' uses of inferior function calls, so, in this patch, I propose having two different timers. One I call the 'direct-call-timeout', which is used for 'call' and 'print' commands. This timeout is by default set to unlimited, which, not surprisingly, means there is no timeout in place. A second timer, which I've called 'indirect-call-timeout', is used for inferior function calls from breakpoint conditions. This timeout has a default value of 30 seconds. This is a reasonably long time to wait, and hopefully should be enough in most cases to allow the inferior call to complete. An inferior call that takes more than 30 seconds, which is installed on a breakpoint condition is really going to slow down the debug session, so hopefully this is not a common use case. The user is, of course, free to reduce, or increase the timeout value, and can always use Ctrl-c to interrupt an inferior function call, but this timeout will ensure that GDB will stop at some point. The new commands added by this commit are: set direct-call-timeout SECONDS show direct-call-timeout set indirect-call-timeout SECONDS show indirect-call-timeout These new timeouts do depend on async-mode, so, if async-mode is disabled (maint set target-async off), or not supported (e.g. target sim), then the timeout is treated as unlimited (that is, no timeout is set). For targets that "fake" non-async mode, e.g. Linux native, where non-async mode is really just async mode, but then we park the target in a sissuspend, we could easily fix things so that the timeouts still work, however, for targets that really are not async aware, like the simulator, fixing things so that timeouts work correctly would be a much bigger task - that effort would be better spent just making the target async-aware. And so, I'm happy for now that this feature will only work on async targets. The two new show commands will display slightly different text if the current target is a non-async target, which should allow users to understand what's going on. There's a somewhat random test adjustment needed in gdb.base/help.exp, the test uses a regexp with the apropos command, and expects to find a single result. Turns out the new settings I added also matched the regexp, which broke the test. I've updated the regexp a little to exclude my new settings. In infcall.c you'll notice the thread_info::stop_requested flag being set when a timeout occurs. This flag setting is not required as part of this commit, but will be needed in a later commit. However, it seemed like setting this flag fitted better with this commit, which is why the change is added here. Reviewed-By: Tankut Baris Aktemur --- gdb/NEWS | 18 ++ gdb/doc/gdb.texinfo | 66 ++++++ gdb/infcall.c | 221 +++++++++++++++++- gdb/testsuite/gdb.base/help.exp | 2 +- gdb/testsuite/gdb.base/infcall-timeout.c | 36 +++ gdb/testsuite/gdb.base/infcall-timeout.exp | 83 +++++++ .../infcall-from-bp-cond-timeout.c | 169 ++++++++++++++ .../infcall-from-bp-cond-timeout.exp | 157 +++++++++++++ 8 files changed, 747 insertions(+), 5 deletions(-) create mode 100644 gdb/testsuite/gdb.base/infcall-timeout.c create mode 100644 gdb/testsuite/gdb.base/infcall-timeout.exp create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.c create mode 100644 gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp diff --git a/gdb/NEWS b/gdb/NEWS index 649a3a9824a..4ddc0a047f0 100644 --- a/gdb/NEWS +++ b/gdb/NEWS @@ -111,6 +111,24 @@ info main $2 = 1 (gdb) break func if $_shell("some command") == 0 +set direct-call-timeout SECONDS +show direct-call-timeout +set indirect-call-timeout SECONDS +show indirect-call-timeout + These new settings can be used to limit how long GDB will wait for + an inferior function call to complete. The direct timeout is used + for inferior function calls from e.g. 'call' and 'print' commands, + while the indirect timeout is used for inferior function calls from + within a conditional breakpoint expression. + + The default for the direct timeout is unlimited, while the default + for the indirect timeout is 30 seconds. + + These timeouts will only have an effect for targets that are + operating in async mode. For non-async targets the timeouts are + ignored, GDB will wait indefinitely for an inferior function to + complete, unless interrupted by the user using Ctrl-C. + * MI changes ** mi now reports 'no-history' as a stop reason when hitting the end of the diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo index 40d1f244417..f27ec719d11 100644 --- a/gdb/doc/gdb.texinfo +++ b/gdb/doc/gdb.texinfo @@ -20920,6 +20920,72 @@ @code{step}, etc). In this case, when the inferior finally returns to the dummy-frame, @value{GDBN} will once again halt the inferior. +On targets that support asynchronous execution (@pxref{Background +Execution}) @value{GDBN} can place a timeout on any functions called +from @value{GDBN}. If the timeout expires and the function call is +still ongoing, then @value{GDBN} will interrupt the program. + +For targets that don't support asynchronous execution +(@pxref{Background Execution}) then timeouts for functions called from +@value{GDBN} are not supported, the timeout settings described below +will be treated as @code{unlimited}, meaning @value{GDBN} will wait +indefinitely for function call to complete, unless interrupted by the +user using @kbd{Ctrl-C}. + +@table @code +@item set direct-call-timeout @var{seconds} +@kindex set direct-call-timeout +@cindex timeout for called functions +Set the timeout used when calling functions in the program to +@var{seconds}, which should be an integer greater than zero, or the +special value @code{unlimited}, which indicates no timeout should be +used. The default for this setting is @code{unlimited}. + +This setting is used when the user calls a function directly from the +command prompt, for example with a @code{call} or @code{print} +command. + +This setting only works for targets that support asynchronous +execution (@pxref{Background Execution}), for any other target the +setting is treated as @code{unlimited}. + +@item show direct-call-timeout +@kindex show direct-call-timeout +@cindex timeout for called functions +Show the timeout used when calling functions in the program with a +@code{call} or @code{print} command. +@end table + +It is also possible to call functions within the program from the +condition of a conditional breakpoint (@pxref{Conditions, ,Break +Conditions}). A different setting controls the timeout used for +function calls made from a breakpoint condition. + +@table @code +@item set indirect-call-timeout @var{seconds} +@kindex set indirect-call-timeout +@cindex timeout for called functions +Set the timeout used when calling functions in the program from a +breakpoint or watchpoint condition to @var{seconds}, which should be +an integer greater than zero, or the special value @code{unlimited}, +which indicates no timeout should be used. The default for this +setting is @code{30} seconds. + +This setting only works for targets that support asynchronous +execution (@pxref{Background Execution}), for any other target the +setting is treated as @code{unlimited}. + +If a function called from a breakpoint or watchpoint condition times +out, then @value{GDBN} will stop at the point where the timeout +occurred. The breakpoint condition evaluation will be abandoned. + +@item show indirect-call-timeout +@kindex show indirect-call-timeout +@cindex timeout for called functions +Show the timeout used when calling functions in the program from a +breakpoint or watchpoint condition. +@end table + @subsection Calling functions with no debug info @cindex no debug info functions diff --git a/gdb/infcall.c b/gdb/infcall.c index b055fbfbd64..162659ca8e6 100644 --- a/gdb/infcall.c +++ b/gdb/infcall.c @@ -96,6 +96,53 @@ show_may_call_functions_p (struct ui_file *file, int from_tty, value); } +/* A timeout (in seconds) for direct inferior calls. A direct inferior + call is one the user triggers from the prompt, e.g. with a 'call' or + 'print' command. Compare with the definition of indirect calls below. */ + +static unsigned int direct_call_timeout = UINT_MAX; + +/* Implement 'show direct-call-timeout'. */ + +static void +show_direct_call_timeout (struct ui_file *file, int from_tty, + struct cmd_list_element *c, const char *value) +{ + if (target_has_execution () && !target_can_async_p ()) + gdb_printf (file, _("Current target does not support async mode, timeout " + "for direct inferior calls is \"unlimited\".\n")); + else if (direct_call_timeout == UINT_MAX) + gdb_printf (file, _("Timeout for direct inferior function calls " + "is \"unlimited\".\n")); + else + gdb_printf (file, _("Timeout for direct inferior function calls " + "is \"%s seconds\".\n"), value); +} + +/* A timeout (in seconds) for indirect inferior calls. An indirect inferior + call is one that originates from within GDB, for example, when + evaluating an expression for a conditional breakpoint. Compare with + the definition of direct calls above. */ + +static unsigned int indirect_call_timeout = 30; + +/* Implement 'show indirect-call-timeout'. */ + +static void +show_indirect_call_timeout (struct ui_file *file, int from_tty, + struct cmd_list_element *c, const char *value) +{ + if (target_has_execution () && !target_can_async_p ()) + gdb_printf (file, _("Current target does not support async mode, timeout " + "for indirect inferior calls is \"unlimited\".\n")); + else if (indirect_call_timeout == UINT_MAX) + gdb_printf (file, _("Timeout for indirect inferior function calls " + "is \"unlimited\".\n")); + else + gdb_printf (file, _("Timeout for indirect inferior function calls " + "is \"%s seconds\".\n"), value); +} + /* How you should pass arguments to a function depends on whether it was defined in K&R style or prototype style. If you define a function using the K&R syntax that takes a `float' argument, then @@ -590,6 +637,86 @@ call_thread_fsm::should_notify_stop () return true; } +/* A class to control creation of a timer that will interrupt a thread + during an inferior call. */ +struct infcall_timer_controller +{ + /* Setup an event-loop timer that will interrupt PTID if the inferior + call takes too long. DIRECT_CALL_P is true when this inferior call is + a result of the user using a 'print' or 'call' command, and false when + this inferior call is a result of e.g. a conditional breakpoint + expression, this is used to select which timeout to use. */ + infcall_timer_controller (thread_info *thr, bool direct_call_p) + : m_thread (thr) + { + unsigned int timeout + = direct_call_p ? direct_call_timeout : indirect_call_timeout; + if (timeout < UINT_MAX && target_can_async_p ()) + { + int ms = timeout * 1000; + int id = create_timer (ms, infcall_timer_controller::timed_out, this); + m_timer_id.emplace (id); + infcall_debug_printf ("Setting up infcall timeout timer for " + "ptid %s: %d milliseconds", + m_thread->ptid.to_string ().c_str (), ms); + } + } + + /* Destructor. Ensure that the timer is removed from the event loop. */ + ~infcall_timer_controller () + { + /* If the timer has already triggered, then it will have already been + deleted from the event loop. If the timer has not triggered, then + delete it now. */ + if (m_timer_id.has_value () && !m_triggered) + delete_timer (*m_timer_id); + + /* Just for clarity, discard the timer id now. */ + m_timer_id.reset (); + } + + /* Return true if there was a timer in place, and the timer triggered, + otherwise, return false. */ + bool triggered_p () + { + gdb_assert (!m_triggered || m_timer_id.has_value ()); + return m_triggered; + } + +private: + /* The thread we should interrupt. */ + thread_info *m_thread; + + /* Set true when the timer is triggered. */ + bool m_triggered = false; + + /* Given a value when a timer is in place. */ + gdb::optional m_timer_id; + + /* Callback for the timer, forwards to ::trigger below. */ + static void + timed_out (gdb_client_data context) + { + infcall_timer_controller *ctrl + = static_cast (context); + ctrl->trigger (); + } + + /* Called when the timer goes off. Stop thread m_thread. */ + void + trigger () + { + m_triggered = true; + + scoped_disable_commit_resumed disable_commit_resumed ("infcall timeout"); + + infcall_debug_printf ("Stopping thread %s", + m_thread->ptid.to_string ().c_str ()); + target_stop (m_thread->ptid); + m_thread->stop_requested = true; + } +}; + /* Subroutine of call_function_by_hand to simplify it. Start up the inferior and wait for it to stop. Return the exception if there's an error, or an exception with @@ -600,13 +727,15 @@ call_thread_fsm::should_notify_stop () static struct gdb_exception run_inferior_call (std::unique_ptr sm, - struct thread_info *call_thread, CORE_ADDR real_pc) + struct thread_info *call_thread, CORE_ADDR real_pc, + bool *timed_out_p) { INFCALL_SCOPED_DEBUG_ENTER_EXIT; struct gdb_exception caught_error; ptid_t call_thread_ptid = call_thread->ptid; int was_running = call_thread->state == THREAD_RUNNING; + *timed_out_p = false; infcall_debug_printf ("call function at %s in thread %s, was_running = %d", core_addr_to_string (real_pc), @@ -618,6 +747,16 @@ run_inferior_call (std::unique_ptr sm, scoped_restore restore_in_infcall = make_scoped_restore (&call_thread->control.in_infcall, 1); + /* If the thread making the inferior call stops with a timeout then the + stop_requested flag will be set. However, we don't want changes to + this flag to leak back to our caller, we might be here to handle an + inferior call from a breakpoint condition, so leaving this flag set + would appear that the breakpoint stop was actually a requested stop, + which is not true, and will cause GDB to print extra messages to the + output. */ + scoped_restore restore_stop_requested + = make_scoped_restore (&call_thread->stop_requested, false); + clear_proceed_status (0); /* Associate the FSM with the thread after clear_proceed_status @@ -651,11 +790,23 @@ run_inferior_call (std::unique_ptr sm, infrun_debug_show_threads ("non-exited threads after proceed for inferior-call", all_non_exited_threads ()); + /* Setup a timer (if possible, and if the settings allow) to prevent + the inferior call running forever. */ + bool direct_call_p = !call_thread->control.in_cond_eval; + infcall_timer_controller infcall_timer (call_thread, direct_call_p); + /* Inferior function calls are always synchronous, even if the target supports asynchronous execution. */ wait_sync_command_done (); - infcall_debug_printf ("inferior call completed successfully"); + /* If the timer triggered then the inferior call failed. */ + if (infcall_timer.triggered_p ()) + { + infcall_debug_printf ("inferior call timed out"); + *timed_out_p = true; + } + else + infcall_debug_printf ("inferior call completed successfully"); } catch (gdb_exception &e) { @@ -1309,6 +1460,10 @@ call_function_by_hand_dummy (struct value *function, scoped_restore restore_stopped_by_random_signal = make_scoped_restore (&stopped_by_random_signal, 0); + /* Set to true by the call to run_inferior_call below if the inferior + call is artificially interrupted by GDB due to taking too long. */ + bool timed_out_p = false; + /* - SNIP - SNIP - SNIP - SNIP - SNIP - SNIP - SNIP - SNIP - SNIP - If you're looking to implement asynchronous dummy-frames, then just below is the place to chop this function in two.. */ @@ -1335,7 +1490,8 @@ call_function_by_hand_dummy (struct value *function, struct_addr); { std::unique_ptr sm_up (sm); - e = run_inferior_call (std::move (sm_up), call_thread.get (), real_pc); + e = run_inferior_call (std::move (sm_up), call_thread.get (), real_pc, + &timed_out_p); } if (e.reason < 0) @@ -1487,7 +1643,10 @@ When the function is done executing, GDB will silently stop."), std::string name = get_function_name (funaddr, name_buf, sizeof (name_buf)); - if (stopped_by_random_signal) + /* If the inferior call timed out then it will have been interrupted + by a signal, but we want to report this differently to the user, + which is done later in this function. */ + if (stopped_by_random_signal && !timed_out_p) { /* We stopped inside the FUNCTION because of a random signal. Further execution of the FUNCTION is not @@ -1531,6 +1690,36 @@ GDB remains in the frame where the signal was received.\n\ To change this behavior use \"set unwindonsignal on\".\n\ Evaluation of the expression containing the function\n\ (%s) will be abandoned.\n\ +When the function is done executing, GDB will silently stop."), + name.c_str ()); + } + } + + if (timed_out_p) + { + /* A timeout results in a signal being sent to the inferior. */ + gdb_assert (stopped_by_random_signal); + + /* Indentation is weird here. A later patch is going to move the + following block into an if/else, so I'm leaving the indentation + here to minimise the later patch. + + Also, the error message used below refers to 'set + unwind-on-timeout' which doesn't exist yet. This will be added + in a later commit, I'm leaving this in for now to minimise the + churn caused by the commit that adds unwind-on-timeout. */ + { + /* The user wants to stay in the frame where we stopped + (default). Discard inferior status, we're not at the same + point we started at. */ + discard_infcall_control_state (inf_status.release ()); + + error (_("\ +The program being debugged timed out while in a function called from GDB.\n\ +GDB remains in the frame where the timeout occurred.\n\ +To change this behavior use \"set unwind-on-timeout on\".\n\ +Evaluation of the expression containing the function\n\ +(%s) will be abandoned.\n\ When the function is done executing, GDB will silently stop."), name.c_str ()); } @@ -1644,6 +1833,30 @@ The default is to unwind the frame."), show_unwind_on_terminating_exception_p, &setlist, &showlist); + add_setshow_uinteger_cmd ("direct-call-timeout", no_class, + &direct_call_timeout, _("\ +Set the timeout, for direct calls to inferior function calls."), _("\ +Show the timeout, for direct calls to inferior function calls."), _("\ +If running on a target that supports, and is running in, async mode\n\ +then this timeout is used for any inferior function calls triggered\n\ +directly from the prompt, i.e. from a 'call' or 'print' command. The\n\ +timeout is specified in seconds."), + nullptr, + show_direct_call_timeout, + &setlist, &showlist); + + add_setshow_uinteger_cmd ("indirect-call-timeout", no_class, + &indirect_call_timeout, _("\ +Set the timeout, for indirect calls to inferior function calls."), _("\ +Show the timeout, for indirect calls to inferior function calls."), _("\ +If running on a target that supports, and is running in, async mode\n\ +then this timeout is used for any inferior function calls triggered\n\ +indirectly, i.e. being made as part of a breakpoint, or watchpoint,\n\ +condition expression. The timeout is specified in seconds."), + nullptr, + show_indirect_call_timeout, + &setlist, &showlist); + add_setshow_boolean_cmd ("infcall", class_maintenance, &debug_infcall, _("Set inferior call debugging."), diff --git a/gdb/testsuite/gdb.base/help.exp b/gdb/testsuite/gdb.base/help.exp index 87919a819ab..504bf90cc15 100644 --- a/gdb/testsuite/gdb.base/help.exp +++ b/gdb/testsuite/gdb.base/help.exp @@ -121,7 +121,7 @@ gdb_test "help info bogus-gdb-command" "Undefined info command: \"bogus-gdb-comm gdb_test "help gotcha" "Undefined command: \"gotcha\"\. Try \"help\"\." # Test apropos regex. -gdb_test "apropos \\\(print\[\^\[ bsiedf\\\".-\]\\\)" "handle -- Specify how to handle signals\." +gdb_test "apropos \\\(print\[\^\[ bsiedf\\\"'.-\]\\\)" "handle -- Specify how to handle signals\." # Test apropos >1 word string. gdb_test "apropos handle signal" "handle -- Specify how to handle signals\." # Test apropos apropos. diff --git a/gdb/testsuite/gdb.base/infcall-timeout.c b/gdb/testsuite/gdb.base/infcall-timeout.c new file mode 100644 index 00000000000..12774ca2599 --- /dev/null +++ b/gdb/testsuite/gdb.base/infcall-timeout.c @@ -0,0 +1,36 @@ +/* Copyright 2022-2023 Free Software Foundation, Inc. + + This file is part of GDB. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include + +/* This function is called from GDB. */ +int +function_that_never_returns () +{ + while (1) + sleep (1); + + return 0; +} + +int +main () +{ + alarm (300); + + return 0; +} diff --git a/gdb/testsuite/gdb.base/infcall-timeout.exp b/gdb/testsuite/gdb.base/infcall-timeout.exp new file mode 100644 index 00000000000..729d8d8734a --- /dev/null +++ b/gdb/testsuite/gdb.base/infcall-timeout.exp @@ -0,0 +1,83 @@ +# Copyright 2022-2023 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# Test GDB's direct-call-timeout setting, that is, ensure that if an +# inferior function call, invoked from e.g. a 'print' command, takes +# too long, then GDB can interrupt it, and return control to the user. + +standard_testfile + +if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ + {debug}] == -1 } { + return +} + +# Start GDB according to TARGET_ASYNC and TARGET_NON_STOP, then adjust +# the direct-call-timeout, and make an inferior function call that +# will never return. GDB should eventually timeout and stop the +# inferior. +proc_with_prefix run_test { target_async target_non_stop } { + save_vars { ::GDBFLAGS } { + append ::GDBFLAGS \ + " -ex \"maint set target-non-stop $target_non_stop\"" + append ::GDBFLAGS \ + " -ex \"maintenance set target-async ${target_async}\"" + + clean_restart ${::binfile} + } + + if {![runto_main]} { + return + } + + gdb_test_no_output "set direct-call-timeout 5" + + # When non-stop mode is off we get slightly different output from GDB. + if { ([target_info gdb_protocol] == "remote" + || [target_info gdb_protocol] == "extended-remote") + && !$target_non_stop } { + set stopped_line_pattern "Program received signal SIGINT, Interrupt\\." + } else { + set stopped_line_pattern "Program stopped\\." + } + + gdb_test "print function_that_never_returns ()" \ + [multi_line \ + $stopped_line_pattern \ + ".*" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB remains in the frame where the timeout occurred\\." \ + "To change this behavior use \"set unwind-on-timeout on\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(function_that_never_returns\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] + + gdb_test "bt" ".* function_that_never_returns .*.*" +} + +foreach_with_prefix target_async { "on" "off" } { + + if { !$target_async } { + # GDB can't timeout while waiting for a thread if the target + # runs with async-mode turned off; once the target is running + # GDB is effectively blocked until the target stops for some + # reason. + continue + } + + foreach_with_prefix target_non_stop { "on" "off" } { + run_test $target_async $target_non_stop + } +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.c b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.c new file mode 100644 index 00000000000..4da4245746e --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.c @@ -0,0 +1,169 @@ +/* This testcase is part of GDB, the GNU debugger. + + Copyright 2022-2023 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . */ + +#include +#include +#include +#include +#include +#include + +#define NUM_THREADS 5 + +/* Semaphores, used to track when threads have started, and to control + when the threads finish. */ +sem_t startup_semaphore; +sem_t finish_semaphore; +sem_t thread_1_semaphore; +sem_t thread_2_semaphore; + +/* Mutex to control when the first worker thread hit a breakpoint + location. */ +pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; + +/* Global variable to poke, just so threads have something to do. */ +volatile int global_var = 0; + +int +condition_func () +{ + /* Let thread 2 run. */ + if (sem_post (&thread_2_semaphore) != 0) + abort (); + + /* Wait for thread 2 to complete its actions. */ + if (sem_wait (&thread_1_semaphore) != 0) + abort (); + + return 1; +} + +void +do_segfault () +{ + volatile int *p = 0; + *p = 0; /* Segfault here. */ +} + +void * +worker_func (void *arg) +{ + int tid = *((int *) arg); + + /* Let the main thread know that this worker has started. */ + if (sem_post (&startup_semaphore) != 0) + abort (); + + switch (tid) + { + case 0: + /* Wait for MUTEX to become available, then pass through the + conditional breakpoint location. */ + if (pthread_mutex_lock (&mutex) != 0) + abort (); + global_var = 99; /* Conditional breakpoint here. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + break; + + case 1: + if (sem_wait (&thread_2_semaphore) != 0) + abort (); + do_segfault (); + if (sem_post (&thread_1_semaphore) != 0) + abort (); + + /* Fall through. */ + default: + /* Wait until we are allowed to finish. */ + if (sem_wait (&finish_semaphore) != 0) + abort (); + break; + } +} + +void +stop_marker () +{ + global_var = 99; /* Stop marker. */ +} + +/* The main program entry point. */ + +int +main () +{ + pthread_t threads[NUM_THREADS]; + int args[NUM_THREADS]; + void *retval; + + /* An alarm, just in case the thread deadlocks. */ + alarm (300); + + /* Semaphore initialization. */ + if (sem_init (&startup_semaphore, 0, 0) != 0) + abort (); + if (sem_init (&finish_semaphore, 0, 0) != 0) + abort (); + if (sem_init (&thread_1_semaphore, 0, 0) != 0) + abort (); + if (sem_init (&thread_2_semaphore, 0, 0) != 0) + abort (); + + /* Lock MUTEX, this prevents the first worker thread from rushing ahead. */ + if (pthread_mutex_lock (&mutex) != 0) + abort (); + + /* Worker thread creation. */ + for (int i = 0; i < NUM_THREADS; i++) + { + args[i] = i; + pthread_create (&threads[i], NULL, worker_func, &args[i]); + } + + /* Wait for every thread to start. */ + for (int i = 0; i < NUM_THREADS; i++) + { + if (sem_wait (&startup_semaphore) != 0) + abort (); + } + + /* Unlock the first thread so it can proceed. */ + if (pthread_mutex_unlock (&mutex) != 0) + abort (); + + /* Wait for the first thread only. */ + pthread_join (threads[0], &retval); + + /* Now post FINISH_SEMAPHORE to allow all the other threads to finish. */ + for (int i = 1; i < NUM_THREADS; i++) + sem_post (&finish_semaphore); + + /* Now wait for the remaining threads to complete. */ + for (int i = 1; i < NUM_THREADS; i++) + pthread_join (threads[i], &retval); + + /* Semaphore cleanup. */ + sem_destroy (&finish_semaphore); + sem_destroy (&startup_semaphore); + sem_destroy (&thread_1_semaphore); + sem_destroy (&thread_2_semaphore); + + stop_marker (); + + return 0; +} diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp new file mode 100644 index 00000000000..3e74958589c --- /dev/null +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp @@ -0,0 +1,157 @@ +# Copyright 2022-2023 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# Tests inferior calls executed from a breakpoint condition in +# a multi-threaded program. +# +# This test has the inferior function call timeout, and checks how GDB +# handles this situation. + +standard_testfile + +if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ + {debug pthreads}] } { + return +} + +set cond_bp_line [gdb_get_line_number "Conditional breakpoint here"] +set final_bp_line [gdb_get_line_number "Stop marker"] +set segfault_line [gdb_get_line_number "Segfault here"] + +# Setup GDB based on TARGET_ASYNC and TARGET_NON_STOP. Setup some +# breakpoints in the inferior, one of which has an inferior call +# within its condition. +# +# Continue GDB, the breakpoint with inferior call will be hit, but the +# inferior call will never return. We expect GDB to timeout. +# +# The reason that the inferior call never completes is that a second +# thread, on which the inferior call relies, either hits a breakpoint +# (when OTHER_THREAD_BP is true), or crashes (when OTHER_THREAD_BP is +# false). +proc run_test { target_async target_non_stop other_thread_bp } { + save_vars { ::GDBFLAGS } { + append ::GDBFLAGS " -ex \"maint set target-non-stop $target_non_stop\"" + append ::GDBFLAGS " -ex \"maintenance set target-async ${target_async}\"" + + clean_restart ${::binfile} + } + + if {![runto_main]} { + return + } + + # The default timeout for indirect inferior calls (e.g. inferior + # calls for conditional breakpoint expressions) is pretty high. + # We don't want the test to take too long, so reduce this. + # + # However, the test relies on a second thread hitting some event + # (either a breakpoint or signal) before this timeout expires. + # + # There is a chance that on a really slow system this might not + # happen, in which case the test might fail. + # + # However, we still allocate 5 seconds, which feels like it should + # be enough time in most cases, but maybe we need to do something + # smarter here? Possibly we could have some initial run where the + # inferior doesn't timeout, but does perform the same interaction + # between threads, we could time that, and use that as the basis + # for this timeout. For now though, we just hope 5 seconds is + # enough. + gdb_test_no_output "set indirect-call-timeout 5" + + gdb_breakpoint \ + "${::srcfile}:${::cond_bp_line} if (condition_func ())" + set bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for conditional breakpoint"] + + gdb_breakpoint "${::srcfile}:${::final_bp_line}" + set final_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for final breakpoint"] + + # The thread performing an inferior call relies on a second + # thread. The second thread will segfault unless it hits a + # breakpoint first. In either case the initial thread will not + # complete its inferior call. + if { $other_thread_bp } { + gdb_breakpoint "${::srcfile}:${::segfault_line}" + set segfault_bp_num [get_integer_valueof "\$bpnum" "*UNKNOWN*" \ + "get number for segfault breakpoint"] + } + + # When non-stop mode is off we get slightly different output from GDB. + if { ([target_info gdb_protocol] == "remote" + || [target_info gdb_protocol] == "extended-remote") + && !$target_non_stop} { + set stopped_line_pattern "Thread ${::decimal} \"\[^\r\n\"\]+\" received signal SIGINT, Interrupt\\." + } else { + set stopped_line_pattern "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." + } + + gdb_test "continue" \ + [multi_line \ + $stopped_line_pattern \ + ".*" \ + "Error in testing condition for breakpoint ${bp_num}:" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB remains in the frame where the timeout occurred\\." \ + "To change this behavior use \"set unwind-on-timeout on\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(condition_func\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] \ + "expected timeout waiting for inferior call to complete" + + # Remember that other thread that either crashed (with a segfault) + # or hit a breakpoint? Now that the inferior call has timed out, + # if we try to resume then we should see the pending event from + # that other thread. + if { $other_thread_bp } { + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "" \ + "Thread ${::decimal} \"\[^\"\r\n\]+\" hit Breakpoint ${segfault_bp_num}, do_segfault \[^\r\n\]+:${::segfault_line}" \ + "${::decimal}\\s+\[^\r\n\]+Segfault here\[^\r\n\]+"] \ + "hit the segfault breakpoint" + } else { + gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + ".*" \ + "Thread ${::decimal} \"infcall-from-bp\" received signal SIGSEGV, Segmentation fault\\." \ + "\\\[Switching to Thread \[^\r\n\]+\\\]" \ + "${::hex} in do_segfault \\(\\) at \[^\r\n\]+:${::segfault_line}" \ + "${::decimal}\\s+\[^\r\n\]+Segfault here\[^\r\n\]+"] \ + "hit the segfault" + } +} + +foreach_with_prefix target_async {"on" "off" } { + + if { !$target_async } { + # GDB can't timeout while waiting for a thread if the target + # runs with async-mode turned off; once the target is running + # GDB is effectively blocked until the target stops for some + # reason. + continue + } + + foreach_with_prefix target_non_stop {"off" "on"} { + foreach_with_prefix other_thread_bp { true false } { + run_test $target_async $target_non_stop $other_thread_bp + } + } +} From patchwork Wed Jun 7 10:01:16 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70718 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 58A48385589D for ; Wed, 7 Jun 2023 10:02:17 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 58A48385589D DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132137; bh=qmALBBPNvNEjEL0bJXuk51LOsQ+0sC7VJTYKxzDgIWs=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=ZdVOeaL/okjN+gXVHAcsbLoeFUtnI9jYQEa5D+3I7kxS8XHIEBZburFv40Hqgd+Ga GIWMtkjT9gjDU6MPD5YLZmJ5qgsRZQuNMsSiA8fdXG/JQjs1iWc3dPF/EuXRgs3Sp9 SstDRXSRdUHp2aGtluWuO6UH1Rhzpb9byvLL2cHk= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id 66AAD3857B98 for ; Wed, 7 Jun 2023 10:01:32 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 66AAD3857B98 Received: from mail-wr1-f71.google.com (mail-wr1-f71.google.com [209.85.221.71]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-515-xOynNJyWOKCWrcHp2W0pgw-1; Wed, 07 Jun 2023 06:01:30 -0400 X-MC-Unique: xOynNJyWOKCWrcHp2W0pgw-1 Received: by mail-wr1-f71.google.com with SMTP id ffacd0b85a97d-30af779d6e6so3254809f8f.3 for ; Wed, 07 Jun 2023 03:01:30 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132089; x=1688724089; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=qmALBBPNvNEjEL0bJXuk51LOsQ+0sC7VJTYKxzDgIWs=; b=FHp/cqF0oiJqw2nbB4mcKLhbhxzh6yafLilXtOjSZTpVcD/Rgi0yHsWxW8uD+hJVuq eEILreUpdV2cJARRr+rYbqAbmZc1X1zcjZ+vNQPlaV36HN/u5nViqCcwOaPGCcRC0TOs ItJbwKRD8CQSpcDOMVd6lAAn8qhnL6ALidMSWRN4o7qui6A/broi3yxK+xR/lTTkFOqL BbJD/KQsuGb81YXOLvHujYLj5DJa8WpaPFVy4MDVu/CKQx7YtHuSCLh3d+8SDWJVwA+8 iZX0UMoDgy3/WNwkUEw+Ggq5MzvmI2wryjWLNmzIFReyJ4ED58a0A3bu980v4xW0em8c AQUA== X-Gm-Message-State: AC+VfDw7tA35jXOZ8t+yXXgy/qj+nI0kDkgY5LjFkxuKWa4V3mvJXJA5 yvDPEAkxbolI3FY+C6sa8O51p5p5TavV9atR/OiFv2YLW3WdNzNw3A5qpg9lSeQGXEZT6CpGC0V tLF1cHrUCAsSZaFRmE3SwDxXlnxpVpaQ9xtXXT562ORM605dBQC98Rko2XtsMWob3Rk1LBzm6SN crkdOIng== X-Received: by 2002:a5d:60d2:0:b0:30e:590f:78d1 with SMTP id x18-20020a5d60d2000000b0030e590f78d1mr1215987wrt.63.1686132089325; Wed, 07 Jun 2023 03:01:29 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ5FGM5CLHnlwY9laVRk5Kphd69VzA+AVuKhAAtIgL5ZmXWxJoFnQ9Lt9IN2fzwcaac/yzpTrg== X-Received: by 2002:a5d:60d2:0:b0:30e:590f:78d1 with SMTP id x18-20020a5d60d2000000b0030e590f78d1mr1215971wrt.63.1686132088960; Wed, 07 Jun 2023 03:01:28 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id s9-20020a5d5109000000b003063a1cdaf2sm15096269wrt.48.2023.06.07.03.01.28 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:28 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv8 4/6] gdb/remote: avoid SIGINT after calling remote_target::stop Date: Wed, 7 Jun 2023 11:01:16 +0100 Message-Id: <0d839e4d9288c80fb4cba0b7f02ea6fd52b20fc6.1686131880.git.aburgess@redhat.com> X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" Currently, if the remote target is not running in non-stop mode, then, when GDB calls remote_target::stop, we end up sending an interrupt packet \x03 to the remote target. If the user interrupts the inferior from the GDB prompt (e.g. by typing Ctrl-c), then GDB calls remote_target::interrupt, which also ends up sending the interrupt packet. The problem here is that both of these mechanisms end up sending the interrupt packet, which means, when the target stops with a SIGINT, and this is reported back to GDB, we have no choice but to report this to the user as a SIGINT stop event. Now maybe this is the correct thing to do, after all the target has been stopped with SIGINT. However, this leads to an unfortunate change in behaviour when comparing non-stop vs all-stop mode. When running in non-stop mode, and remote_target::stop is called, the target will be stopped with a vCont packet, and this stop is then reported back to GDB as GDB_SIGNAL_0, this will cause GDB to print a message like: Program stopped. Or: Thread NN "binary name" stopped. In contrast, when non-stop mode is off, we get messages like: Program received SIGINT, Segmentation fault. Or: Thread NN "binary name" received SIGINT, Segmentation fault. In this commit I propose making use of thread_info::stop_requested within remote.c to know if the stop was triggered by GDB (and the SIGINT should be hidden) or if the stop was a user interrupt, and the SIGINT should be printed. In remote_target::process_stop_reply if the inferior stopped with SIGINT and the thread_info::stop_requested flag is set, then we change the stop signal to GDB_SIGNAL_0. Two of the tests added in the previous commit exposed this issue. In the previous commit the tests looked for either of the above patterns. In this commit I've updated these tests to only look for the "stopped" based messages. This commit is the reason why the previous commit took care to set the thread_info::stop_requested flag in infcall.c. Due to the changes to how get_remote_thread_info is now called in remote.c, one of the overloads of get_remote_thread_info is no longer used and can be deleted. --- gdb/remote.c | 23 +++++++++---------- gdb/testsuite/gdb.base/infcall-timeout.exp | 11 +-------- .../infcall-from-bp-cond-timeout.exp | 11 +-------- 3 files changed, 13 insertions(+), 32 deletions(-) diff --git a/gdb/remote.c b/gdb/remote.c index 02ff3a12bdb..d50802bc4a0 100644 --- a/gdb/remote.c +++ b/gdb/remote.c @@ -2729,8 +2729,6 @@ remote_target::remote_add_inferior (bool fake_pid_p, int pid, int attached, } static remote_thread_info *get_remote_thread_info (thread_info *thread); -static remote_thread_info *get_remote_thread_info (remote_target *target, - ptid_t ptid); /* Add thread PTID to GDB's thread list. Tag it as executing/running according to EXECUTING and RUNNING respectively. If SILENT_P (or the @@ -2870,15 +2868,6 @@ get_remote_thread_info (thread_info *thread) return gdb::checked_static_cast (thread->priv.get ()); } -/* Return PTID's private thread data, creating it if necessary. */ - -static remote_thread_info * -get_remote_thread_info (remote_target *target, ptid_t ptid) -{ - thread_info *thr = target->find_thread (ptid); - return get_remote_thread_info (thr); -} - /* Call this function as a result of 1) A halt indication (T packet) containing a thread id 2) A direct query of currthread @@ -8263,7 +8252,8 @@ remote_target::process_stop_reply (struct stop_reply *stop_reply, } remote_notice_new_inferior (ptid, false); - remote_thread_info *remote_thr = get_remote_thread_info (this, ptid); + thread_info *tinfo = this->find_thread (ptid); + remote_thread_info *remote_thr = get_remote_thread_info (tinfo); remote_thr->core = stop_reply->core; remote_thr->stop_reason = stop_reply->stop_reason; remote_thr->watch_data_address = stop_reply->watch_data_address; @@ -8276,6 +8266,15 @@ remote_target::process_stop_reply (struct stop_reply *stop_reply, } else { + /* If this stop was actually requested by GDB then we can hide + the SIGINT from the user. */ + if (status->kind () == TARGET_WAITKIND_STOPPED + && status->sig () == GDB_SIGNAL_INT) + { + if (tinfo->stop_requested) + status->set_stopped (GDB_SIGNAL_0); + } + /* If the target works in all-stop mode, a stop-reply indicates that all the target's threads stopped. */ for (thread_info *tp : all_non_exited_threads (this)) diff --git a/gdb/testsuite/gdb.base/infcall-timeout.exp b/gdb/testsuite/gdb.base/infcall-timeout.exp index 729d8d8734a..2af598e044d 100644 --- a/gdb/testsuite/gdb.base/infcall-timeout.exp +++ b/gdb/testsuite/gdb.base/infcall-timeout.exp @@ -44,18 +44,9 @@ proc_with_prefix run_test { target_async target_non_stop } { gdb_test_no_output "set direct-call-timeout 5" - # When non-stop mode is off we get slightly different output from GDB. - if { ([target_info gdb_protocol] == "remote" - || [target_info gdb_protocol] == "extended-remote") - && !$target_non_stop } { - set stopped_line_pattern "Program received signal SIGINT, Interrupt\\." - } else { - set stopped_line_pattern "Program stopped\\." - } - gdb_test "print function_that_never_returns ()" \ [multi_line \ - $stopped_line_pattern \ + "Program stopped\\." \ ".*" \ "The program being debugged timed out while in a function called from GDB\\." \ "GDB remains in the frame where the timeout occurred\\." \ diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp index 3e74958589c..a94c6a58192 100644 --- a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp @@ -91,18 +91,9 @@ proc run_test { target_async target_non_stop other_thread_bp } { "get number for segfault breakpoint"] } - # When non-stop mode is off we get slightly different output from GDB. - if { ([target_info gdb_protocol] == "remote" - || [target_info gdb_protocol] == "extended-remote") - && !$target_non_stop} { - set stopped_line_pattern "Thread ${::decimal} \"\[^\r\n\"\]+\" received signal SIGINT, Interrupt\\." - } else { - set stopped_line_pattern "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." - } - gdb_test "continue" \ [multi_line \ - $stopped_line_pattern \ + "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." \ ".*" \ "Error in testing condition for breakpoint ${bp_num}:" \ "The program being debugged timed out while in a function called from GDB\\." \ From patchwork Wed Jun 7 10:01:17 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70722 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id E0560385588D for ; Wed, 7 Jun 2023 10:03:51 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org E0560385588D DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132231; bh=5F+lji7+s4RrCo72kNnNfQdsLUhjBLykrQJZMOurtxs=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=E8KCS6lnov048/UW9mwVFOgdEHibZ3CAmZpP4rM6p2i1iwZnh2iWy+GBbTgdZAF7n KsTE3V5M6jz69anhwhxuWJheWO/lDwWKy678pk3dBj6KRPgOBOhYlHZNffm6LAgUhs 2PjZ1u1HM5dLVQG2iGpVjapDIjYyj4s6ptmr7ybw= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTPS id 840843857729 for ; Wed, 7 Jun 2023 10:01:36 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 840843857729 Received: from mail-wm1-f71.google.com (mail-wm1-f71.google.com [209.85.128.71]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-651-s11lZfNEPLqyPFlohDZOpw-1; Wed, 07 Jun 2023 06:01:32 -0400 X-MC-Unique: s11lZfNEPLqyPFlohDZOpw-1 Received: by mail-wm1-f71.google.com with SMTP id 5b1f17b1804b1-3f42bcef2acso34109655e9.2 for ; Wed, 07 Jun 2023 03:01:32 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132091; x=1688724091; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=5F+lji7+s4RrCo72kNnNfQdsLUhjBLykrQJZMOurtxs=; b=EQQA7iz4VhmSScwMNzTueDZHVBeWEZ5VhhlzpfQua7BRJgn6LMiWQsxK85/2MJI8sd 4IFWAbx5E4xk9VrpXxDWrGzVM1v09PlPzPlAN8tbRm6+KnMK5B4bmtPUeK+Bn9VKdhU2 9K5JS8xUV8/HIdywFY8rOglFCy+wZtQPkQckY5ebyuTRA6AE6HZaZ4ZdE0e+DUniJBNd PsQ88Kz9pEszbo1fpTQOFRbbEFZ3ms8Q+Y95UchQFuHPU9nPSitgkt1yUvGXV+v4JQVR D2/fo7eaN50HTLvajNdBpkIOj4HJ8YXCYPeD3mKQTa7BAF7Ox3zyyjUPsdqujLSS9YXL 780Q== X-Gm-Message-State: AC+VfDxsH+weA1ctCUFT4FVk8TKLGhQ1MkKeJsG6nn1NJCLj5keODQSq Lv1FDLzD+KMYPslOU9CPqrmyvXo56RRUsEwOavE7yOLFI9kXkl4TvZUAzoxap5ScbZKYEMgxYrk KUf3NrgrOBNmopLxYb1TVhiXcqi0J22TgUO3seSqPRhLVE0mt5idCRKHLfIrTqNdBI48HtPKIRt KJRZeF8g== X-Received: by 2002:a05:600c:290c:b0:3f5:d313:db5e with SMTP id i12-20020a05600c290c00b003f5d313db5emr3899530wmd.15.1686132090970; Wed, 07 Jun 2023 03:01:30 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ54tOnN1cy0RoShlRgybbrq2nGr8Ac2BuTRU7L/ccldCgRhNafFSOTEFWBuGcX6P7cV1yhZ/g== X-Received: by 2002:a05:600c:290c:b0:3f5:d313:db5e with SMTP id i12-20020a05600c290c00b003f5d313db5emr3899505wmd.15.1686132090463; Wed, 07 Jun 2023 03:01:30 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id n2-20020a05600c294200b003f7ba52eeccsm1633010wmd.7.2023.06.07.03.01.29 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:30 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv8 5/6] gdb: introduce unwind-on-timeout setting Date: Wed, 7 Jun 2023 11:01:17 +0100 Message-Id: <871a8d1e8da3df77b82da060b9cfbfc7848a1300.1686131880.git.aburgess@redhat.com> X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" Now that inferior function calls can timeout (see the recent introduction of direct-call-timeout and indirect-call-timeout), this commit adds a new setting unwind-on-timeout. This new setting is just like the existing unwindonsignal and unwind-on-terminating-exception, but the new setting will cause GDB to unwind the stack if an inferior function call times out. The existing inferior function call timeout tests have been updated to cover the new setting. --- gdb/NEWS | 9 +++ gdb/doc/gdb.texinfo | 25 ++++++-- gdb/infcall.c | 62 ++++++++++++++++--- gdb/testsuite/gdb.base/infcall-timeout.exp | 49 +++++++++++---- .../infcall-from-bp-cond-timeout.exp | 55 +++++++++++----- 5 files changed, 160 insertions(+), 40 deletions(-) diff --git a/gdb/NEWS b/gdb/NEWS index 4ddc0a047f0..81eb270c5ae 100644 --- a/gdb/NEWS +++ b/gdb/NEWS @@ -129,6 +129,15 @@ show indirect-call-timeout ignored, GDB will wait indefinitely for an inferior function to complete, unless interrupted by the user using Ctrl-C. +set unwind-on-timeout on|off +show unwind-on-timeout + These commands control whether GDB should unwind the stack when a + timeout occurs during an inferior function call. The default is + off, in which case the inferior will remain in the frame where the + timeout occurred. When on, GDB will unwind the stack removing the + dummy frame that was added for the inferior call, and restoring the + inferior state to how it was before the inferior call started. + * MI changes ** mi now reports 'no-history' as a stop reason when hitting the end of the diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo index f27ec719d11..4f2062ff7e2 100644 --- a/gdb/doc/gdb.texinfo +++ b/gdb/doc/gdb.texinfo @@ -20865,6 +20865,21 @@ Show the current setting of stack unwinding in the functions called by @value{GDBN}. +@item set unwind-on-timeout +@kindex set unwind-on-timeout +@cindex unwind stack in called functions when timing out +@cindex call dummy stack unwinding on timeout. +Set unwinding of the stack if a function called from @value{GDBN} +times out. If set to @code{off} (the default), @value{GDBN} stops in +the frame where the timeout occurred. If set to @code{on}, +@value{GDBN} unwinds the stack it created for the call and restores +the context to what it was before the call. + +@item show unwind-on-timeout +@kindex show unwind-on-timeout +Show whether @value{GDBN} will unwind the stack if a function called +from @value{GDBN} times out. + @item set may-call-functions @kindex set may-call-functions @cindex disabling calling functions in the program @@ -20896,11 +20911,11 @@ If a called function is interrupted for any reason, including hitting a breakpoint, or triggering a watchpoint, and the stack is not unwound -due to @code{set unwind-on-terminating-exception on} or @code{set -unwindonsignal on} (@pxref{stack unwind settings}), -then the dummy-frame, created by @value{GDBN} to facilitate the call -to the program function, will be visible in the backtrace, for example -frame @code{#3} in the following backtrace: +due to @code{set unwind-on-terminating-exception on}, @code{set +unwind-on-timeout on}, or @code{set unwindonsignal on} (@pxref{stack +unwind settings}), then the dummy-frame, created by @value{GDBN} to +facilitate the call to the program function, will be visible in the +backtrace, for example frame @code{#3} in the following backtrace: @smallexample (@value{GDBP}) backtrace diff --git a/gdb/infcall.c b/gdb/infcall.c index 162659ca8e6..526be837248 100644 --- a/gdb/infcall.c +++ b/gdb/infcall.c @@ -218,6 +218,27 @@ show_unwind_on_terminating_exception_p (struct ui_file *file, int from_tty, value); } +/* This boolean tells GDB what to do if an inferior function, called from + GDB, times out. If true, GDB unwinds the stack and restores the context + to what it was before the call. When false, GDB leaves the thread as it + is at the point of the timeout. + + The default is to stop in the frame where the timeout occurred. */ + +static bool unwind_on_timeout_p = false; + +/* Implement 'show unwind-on-timeout'. */ + +static void +show_unwind_on_timeout_p (struct ui_file *file, int from_tty, + struct cmd_list_element *c, const char *value) +{ + gdb_printf (file, + _("Unwinding of stack if a timeout occurs " + "while in a call dummy is %s.\n"), + value); +} + /* Perform the standard coercions that are specified for arguments to be passed to C, Ada or Fortran functions. @@ -1700,14 +1721,27 @@ When the function is done executing, GDB will silently stop."), /* A timeout results in a signal being sent to the inferior. */ gdb_assert (stopped_by_random_signal); - /* Indentation is weird here. A later patch is going to move the - following block into an if/else, so I'm leaving the indentation - here to minimise the later patch. + if (unwind_on_timeout_p) + { + /* The user wants the context restored. */ + + /* We must get back to the frame we were before the + dummy call. */ + dummy_frame_pop (dummy_id, call_thread.get ()); - Also, the error message used below refers to 'set - unwind-on-timeout' which doesn't exist yet. This will be added - in a later commit, I'm leaving this in for now to minimise the - churn caused by the commit that adds unwind-on-timeout. */ + /* We also need to restore inferior status to that before the + dummy call. */ + restore_infcall_control_state (inf_status.release ()); + + error (_("\ +The program being debugged timed out while in a function called from GDB.\n\ +GDB has restored the context to what it was before the call.\n\ +To change this behavior use \"set unwind-on-timeout off\".\n\ +Evaluation of the expression containing the function\n\ +(%s) will be abandoned."), + name.c_str ()); + } + else { /* The user wants to stay in the frame where we stopped (default). Discard inferior status, we're not at the same @@ -1833,6 +1867,20 @@ The default is to unwind the frame."), show_unwind_on_terminating_exception_p, &setlist, &showlist); + add_setshow_boolean_cmd ("unwind-on-timeout", no_class, + &unwind_on_timeout_p, _("\ +Set unwinding of stack if a timeout occurs while in a call dummy."), _("\ +Show unwinding of stack if a timeout occurs while in a call dummy."), + _("\ +The unwind on timeout flag lets the user determine what gdb should do if\n\ +gdb times out while in a function called from gdb. If set, gdb unwinds\n\ +the stack and restores the context to what it was before the call. If\n\ +unset, gdb leaves the inferior in the frame where the timeout occurred.\n\ +The default is to stop in the frame where the timeout occurred."), + NULL, + show_unwind_on_timeout_p, + &setlist, &showlist); + add_setshow_uinteger_cmd ("direct-call-timeout", no_class, &direct_call_timeout, _("\ Set the timeout, for direct calls to inferior function calls."), _("\ diff --git a/gdb/testsuite/gdb.base/infcall-timeout.exp b/gdb/testsuite/gdb.base/infcall-timeout.exp index 2af598e044d..1f47d143699 100644 --- a/gdb/testsuite/gdb.base/infcall-timeout.exp +++ b/gdb/testsuite/gdb.base/infcall-timeout.exp @@ -28,7 +28,11 @@ if { [build_executable "failed to prepare" ${binfile} "${srcfile}" \ # the direct-call-timeout, and make an inferior function call that # will never return. GDB should eventually timeout and stop the # inferior. -proc_with_prefix run_test { target_async target_non_stop } { +# +# When UNWIND is "off" the inferior wil be left in the frame where the +# timeout occurs, otherwise, when UNWIND is "on", GDB should unwind +# back to the frame where the inferior call was made. +proc_with_prefix run_test { target_async target_non_stop unwind } { save_vars { ::GDBFLAGS } { append ::GDBFLAGS \ " -ex \"maint set target-non-stop $target_non_stop\"" @@ -43,19 +47,36 @@ proc_with_prefix run_test { target_async target_non_stop } { } gdb_test_no_output "set direct-call-timeout 5" + gdb_test_no_output "set unwind-on-timeout $unwind" + + if { $unwind } { + gdb_test "print function_that_never_returns ()" \ + [multi_line \ + "Program stopped\\." \ + ".*" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB has restored the context to what it was before the call\\." \ + "To change this behavior use \"set unwind-on-timeout off\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(function_that_never_returns\\) will be abandoned\\."] - gdb_test "print function_that_never_returns ()" \ - [multi_line \ - "Program stopped\\." \ - ".*" \ - "The program being debugged timed out while in a function called from GDB\\." \ - "GDB remains in the frame where the timeout occurred\\." \ - "To change this behavior use \"set unwind-on-timeout on\"\\." \ - "Evaluation of the expression containing the function" \ - "\\(function_that_never_returns\\) will be abandoned\\." \ - "When the function is done executing, GDB will silently stop\\."] + gdb_test "bt" \ + "#0\\s+main \\(\\).*" + } else { + gdb_test "print function_that_never_returns ()" \ + [multi_line \ + "Program stopped\\." \ + ".*" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB remains in the frame where the timeout occurred\\." \ + "To change this behavior use \"set unwind-on-timeout on\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(function_that_never_returns\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] - gdb_test "bt" ".* function_that_never_returns .*.*" + gdb_test "bt" \ + ".* function_that_never_returns .*.*" + } } foreach_with_prefix target_async { "on" "off" } { @@ -69,6 +90,8 @@ foreach_with_prefix target_async { "on" "off" } { } foreach_with_prefix target_non_stop { "on" "off" } { - run_test $target_async $target_non_stop + foreach_with_prefix unwind { "on" "off" } { + run_test $target_async $target_non_stop $unwind + } } } diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp index a94c6a58192..fedcad3d6c0 100644 --- a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-timeout.exp @@ -41,7 +41,12 @@ set segfault_line [gdb_get_line_number "Segfault here"] # thread, on which the inferior call relies, either hits a breakpoint # (when OTHER_THREAD_BP is true), or crashes (when OTHER_THREAD_BP is # false). -proc run_test { target_async target_non_stop other_thread_bp } { +# +# When UNWIND is "on" GDB will unwind the thread which performed the +# inferior function call back to the state where the inferior call was +# made (when the inferior call times out). Otherwise, when UNWIND is +# "off", the inferior is left in the frame where the timeout occurred. +proc run_test { target_async target_non_stop other_thread_bp unwind } { save_vars { ::GDBFLAGS } { append ::GDBFLAGS " -ex \"maint set target-non-stop $target_non_stop\"" append ::GDBFLAGS " -ex \"maintenance set target-async ${target_async}\"" @@ -71,6 +76,7 @@ proc run_test { target_async target_non_stop other_thread_bp } { # for this timeout. For now though, we just hope 5 seconds is # enough. gdb_test_no_output "set indirect-call-timeout 5" + gdb_test_no_output "set unwind-on-timeout $unwind" gdb_breakpoint \ "${::srcfile}:${::cond_bp_line} if (condition_func ())" @@ -91,18 +97,35 @@ proc run_test { target_async target_non_stop other_thread_bp } { "get number for segfault breakpoint"] } - gdb_test "continue" \ - [multi_line \ - "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." \ - ".*" \ - "Error in testing condition for breakpoint ${bp_num}:" \ - "The program being debugged timed out while in a function called from GDB\\." \ - "GDB remains in the frame where the timeout occurred\\." \ - "To change this behavior use \"set unwind-on-timeout on\"\\." \ - "Evaluation of the expression containing the function" \ - "\\(condition_func\\) will be abandoned\\." \ - "When the function is done executing, GDB will silently stop\\."] \ - "expected timeout waiting for inferior call to complete" + if { $unwind } { + gdb_test "continue" \ + [multi_line \ + "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." \ + ".*" \ + "Error in testing condition for breakpoint ${bp_num}:" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB has restored the context to what it was before the call\\." \ + "To change this behavior use \"set unwind-on-timeout off\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(condition_func\\) will be abandoned\\." \ + "" \ + "Thread ${::decimal}\[^\r\n\]*hit Breakpoint ${bp_num}, \[^\r\n\]+" \ + "\[^\r\n\]+ Conditional breakpoint here\\. \[^\r\n\]+"] \ + "expected timeout waiting for inferior call to complete" + } else { + gdb_test "continue" \ + [multi_line \ + "Thread ${::decimal} \"\[^\r\n\"\]+\" stopped\\." \ + ".*" \ + "Error in testing condition for breakpoint ${bp_num}:" \ + "The program being debugged timed out while in a function called from GDB\\." \ + "GDB remains in the frame where the timeout occurred\\." \ + "To change this behavior use \"set unwind-on-timeout on\"\\." \ + "Evaluation of the expression containing the function" \ + "\\(condition_func\\) will be abandoned\\." \ + "When the function is done executing, GDB will silently stop\\."] \ + "expected timeout waiting for inferior call to complete" + } # Remember that other thread that either crashed (with a segfault) # or hit a breakpoint? Now that the inferior call has timed out, @@ -141,8 +164,10 @@ foreach_with_prefix target_async {"on" "off" } { } foreach_with_prefix target_non_stop {"off" "on"} { - foreach_with_prefix other_thread_bp { true false } { - run_test $target_async $target_non_stop $other_thread_bp + foreach_with_prefix unwind {"off" "on"} { + foreach_with_prefix other_thread_bp { true false } { + run_test $target_async $target_non_stop $other_thread_bp $unwind + } } } } From patchwork Wed Jun 7 10:01:18 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 70719 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 63FD03856DD0 for ; Wed, 7 Jun 2023 10:02:28 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 63FD03856DD0 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1686132148; bh=/cF18uKFYs69qOnEtVxbHe/kBB/ODQE5gFqZ+zEiDzA=; h=To:Cc:Subject:Date:In-Reply-To:References:List-Id: List-Unsubscribe:List-Archive:List-Post:List-Help:List-Subscribe: From:Reply-To:From; b=YWCQmd7T5j1g6A1CqvgkyXtnMDucshwmAqA6uNOuHw/ESRdKr+wLgieiYS91G9iC3 c2ztLRywkuWdwLqjpvb68NFSpMgNTkqSeV4C1FZyWHypBMmDTEyh1k5wQ/pxG2Ia7p 3OFQyCjXh99WPB4SxyNX76srhhQFskqsc2D3ogEU= X-Original-To: gdb-patches@sourceware.org Delivered-To: gdb-patches@sourceware.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.133.124]) by sourceware.org (Postfix) with ESMTPS id 924B03857724 for ; Wed, 7 Jun 2023 10:01:35 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 924B03857724 Received: from mail-wm1-f70.google.com (mail-wm1-f70.google.com [209.85.128.70]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-572-rCXUXbkoObmHDPnUpf7lZQ-1; Wed, 07 Jun 2023 06:01:34 -0400 X-MC-Unique: rCXUXbkoObmHDPnUpf7lZQ-1 Received: by mail-wm1-f70.google.com with SMTP id 5b1f17b1804b1-3f7f264c485so1646445e9.0 for ; Wed, 07 Jun 2023 03:01:33 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686132093; x=1688724093; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=/cF18uKFYs69qOnEtVxbHe/kBB/ODQE5gFqZ+zEiDzA=; b=jefgEFAk7BFVZkWZ0GH9XGdTTVWL+5OrwlDHwmDIeDfu77GkW5XwSQcFg4VwOZpBfZ +4JDbaZvqMWbSIB8AaYBSTgkwwRdS+HNqd1ASTSXq9N7jHVgZQPSC2KAlhkSJuhj27sh ue41eprSZ5SQ3T2mJa3jZwX3JBCgFz2ySMA9cRHiK4qQMbQHHt7d3STAh6Ki+4IQjW/L Zo172x78m7x3c0Di16fa3XNM1nwGZ6SC8BsnrwNgrnibVw30Vib2fxOd7nUn9YA8Qe7I sxI5sjZpR4NlnmLRMZ9aujNq+SnWr71y+NSJtmNtqXWywc0/FEozw8hTZ4kTM1le5zK9 WHMg== X-Gm-Message-State: AC+VfDyX73fdmX6QiWQVuuzLssD/BLGmPw4K4Qx0jmBNtNRxg327PGCB rTPvmj6HosZ+0TW3F0ByTVGhvQfYmpSysiMXj1K4VrQaG0NQko6SZpO9mz6DAROD7ebaDxp4bHK 8IRgN1NCV1TSui7mWfvltw5qA2UQbMHVXNIGcbzM6TW4zZA0RXS9quv+LOfeBYYtQSQuXX/kLQh oNzHD27g== X-Received: by 2002:a1c:7c05:0:b0:3f6:76e:6049 with SMTP id x5-20020a1c7c05000000b003f6076e6049mr4136317wmc.1.1686132092476; Wed, 07 Jun 2023 03:01:32 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ4TKQbU+xhdpy4EMi5cjuAamLSgJbPaDrjCU+WiosQ0ySBNj9tiTTTUDeTwjsRDltEF/hfyaw== X-Received: by 2002:a1c:7c05:0:b0:3f6:76e:6049 with SMTP id x5-20020a1c7c05000000b003f6076e6049mr4136288wmc.1.1686132091849; Wed, 07 Jun 2023 03:01:31 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id u26-20020a7bcb1a000000b003f7a562ff31sm1647457wmj.6.2023.06.07.03.01.31 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 07 Jun 2023 03:01:31 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv8 6/6] gdb: rename unwindonsignal to unwind-on-signal Date: Wed, 7 Jun 2023 11:01:18 +0100 Message-Id: X-Mailer: git-send-email 2.25.4 In-Reply-To: References: MIME-Version: 1.0 X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com X-Spam-Status: No, score=-11.8 required=5.0 tests=BAYES_00, DKIMWL_WL_HIGH, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_NONE, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: gdb-patches@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Gdb-patches mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Andrew Burgess via Gdb-patches From: Andrew Burgess Reply-To: Andrew Burgess Errors-To: gdb-patches-bounces+patchwork=sourceware.org@sourceware.org Sender: "Gdb-patches" We now have unwind-on-timeout and unwind-on-terminating-exception, and then the odd one out unwindonsignal. I'm not a great fan of these squashed together command names, so in this commit I propose renaming this to unwind-on-signal. Obviously I've added the hidden alias unwindonsignal so any existing GDB scripts will keep working. There's one test that I've extended to test the alias works, but in most of the other test scripts I've changed over to use the new name. The docs are updated to reference the new name. --- gdb/NEWS | 11 ++++++ gdb/doc/gdb.texinfo | 16 ++++++--- gdb/infcall.c | 21 ++++++----- gdb/testsuite/gdb.base/callfuncs.exp | 4 +-- gdb/testsuite/gdb.base/infcall-failure.exp | 4 +-- gdb/testsuite/gdb.base/unwindonsignal.exp | 36 +++++++++++++------ gdb/testsuite/gdb.compile/compile-cplus.exp | 2 +- gdb/testsuite/gdb.compile/compile.exp | 2 +- gdb/testsuite/gdb.cp/gdb2495.exp | 16 ++++----- gdb/testsuite/gdb.fortran/function-calls.exp | 2 +- gdb/testsuite/gdb.mi/mi-syn-frame.exp | 2 +- .../infcall-from-bp-cond-simple.exp | 2 +- .../gdb.threads/thread-unwindonsignal.exp | 8 ++--- 13 files changed, 83 insertions(+), 43 deletions(-) diff --git a/gdb/NEWS b/gdb/NEWS index 81eb270c5ae..c41ff0aad19 100644 --- a/gdb/NEWS +++ b/gdb/NEWS @@ -138,6 +138,17 @@ show unwind-on-timeout dummy frame that was added for the inferior call, and restoring the inferior state to how it was before the inferior call started. +set unwind-on-signal on|off +show unwind-on-signal + These new commands replaces the existing set/show unwindonsignal. The + old command is maintained as an alias. + +* Changed commands + +set unwindonsignal on|off +show unwindonsignal + These commands are now aliases for the new set/show unwind-on-signal. + * MI changes ** mi now reports 'no-history' as a stop reason when hitting the end of the diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo index 4f2062ff7e2..7e00d88a83c 100644 --- a/gdb/doc/gdb.texinfo +++ b/gdb/doc/gdb.texinfo @@ -20819,7 +20819,7 @@ It is possible for the function you call via the @code{print} or @code{call} command to generate a signal (e.g., if there's a bug in the function, or if you passed it incorrect arguments). What happens -in that case is controlled by the @code{set unwindonsignal} command. +in that case is controlled by the @code{set unwind-on-signal} command. Similarly, with a C@t{++} program it is possible for the function you call via the @code{print} or @code{call} command to generate an @@ -20833,7 +20833,8 @@ @anchor{stack unwind settings} @table @code -@item set unwindonsignal +@item set unwind-on-signal +@kindex set unwind-on-signal @kindex set unwindonsignal @cindex unwind stack in called functions @cindex call dummy stack unwinding @@ -20844,11 +20845,18 @@ default), @value{GDBN} stops in the frame where the signal was received. -@item show unwindonsignal +The command @code{set unwindonsignal} is an alias for this command, +and is maintained for backward compatibility. + +@item show unwind-on-signal +@kindex show unwind-on-signal @kindex show unwindonsignal Show the current setting of stack unwinding in the functions called by @value{GDBN}. +The command @code{show unwindonsignal} is an alias for this command, +and is maintained for backward compatibility. + @item set unwind-on-terminating-exception @kindex set unwind-on-terminating-exception @cindex unwind stack in called functions with unhandled exceptions @@ -20912,7 +20920,7 @@ If a called function is interrupted for any reason, including hitting a breakpoint, or triggering a watchpoint, and the stack is not unwound due to @code{set unwind-on-terminating-exception on}, @code{set -unwind-on-timeout on}, or @code{set unwindonsignal on} (@pxref{stack +unwind-on-timeout on}, or @code{set unwind-on-signal on} (@pxref{stack unwind settings}), then the dummy-frame, created by @value{GDBN} to facilitate the call to the program function, will be visible in the backtrace, for example frame @code{#3} in the following backtrace: diff --git a/gdb/infcall.c b/gdb/infcall.c index 526be837248..0a800994e4d 100644 --- a/gdb/infcall.c +++ b/gdb/infcall.c @@ -1690,7 +1690,7 @@ When the function is done executing, GDB will silently stop."), error (_("\ The program being debugged was signaled while in a function called from GDB.\n\ GDB has restored the context to what it was before the call.\n\ -To change this behavior use \"set unwindonsignal off\".\n\ +To change this behavior use \"set unwind-on-signal off\".\n\ Evaluation of the expression containing the function\n\ (%s) will be abandoned."), name.c_str ()); @@ -1708,7 +1708,7 @@ Evaluation of the expression containing the function\n\ error (_("\ The program being debugged was signaled while in a function called from GDB.\n\ GDB remains in the frame where the signal was received.\n\ -To change this behavior use \"set unwindonsignal on\".\n\ +To change this behavior use \"set unwind-on-signal on\".\n\ Evaluation of the expression containing the function\n\ (%s) will be abandoned.\n\ When the function is done executing, GDB will silently stop."), @@ -1840,17 +1840,22 @@ The default is to perform the conversion."), show_coerce_float_to_double_p, &setlist, &showlist); - add_setshow_boolean_cmd ("unwindonsignal", no_class, - &unwind_on_signal_p, _("\ + set_show_commands setshow_unwind_on_signal_cmds + = add_setshow_boolean_cmd ("unwind-on-signal", no_class, + &unwind_on_signal_p, _("\ Set unwinding of stack if a signal is received while in a call dummy."), _("\ Show unwinding of stack if a signal is received while in a call dummy."), _("\ -The unwindonsignal lets the user determine what gdb should do if a signal\n\ +The unwind-on-signal lets the user determine what gdb should do if a signal\n\ is received while in a function called from gdb (call dummy). If set, gdb\n\ unwinds the stack and restore the context to what as it was before the call.\n\ The default is to stop in the frame where the signal was received."), - NULL, - show_unwind_on_signal_p, - &setlist, &showlist); + NULL, + show_unwind_on_signal_p, + &setlist, &showlist); + add_alias_cmd ("unwindonsignal", setshow_unwind_on_signal_cmds.set, + no_class, 1, &setlist); + add_alias_cmd ("unwindonsignal", setshow_unwind_on_signal_cmds.show, + no_class, 1, &showlist); add_setshow_boolean_cmd ("unwind-on-terminating-exception", no_class, &unwind_on_terminating_exception_p, _("\ diff --git a/gdb/testsuite/gdb.base/callfuncs.exp b/gdb/testsuite/gdb.base/callfuncs.exp index 2c797a224a7..d9f63e944bf 100644 --- a/gdb/testsuite/gdb.base/callfuncs.exp +++ b/gdb/testsuite/gdb.base/callfuncs.exp @@ -46,7 +46,7 @@ proc do_function_calls {prototypes} { # If any of these calls segv we don't want to affect subsequent tests. # E.g., we want to ensure register values are restored. - gdb_test_no_output "set unwindonsignal on" + gdb_test_no_output "set unwind-on-signal on" gdb_test "p t_char_values(0,0)" " = 0" gdb_test "p t_char_values('a','b')" " = 1" @@ -237,7 +237,7 @@ proc do_function_calls {prototypes} { "call inferior func with struct - returns char *" # Restore default value. - gdb_test_no_output "set unwindonsignal off" + gdb_test_no_output "set unwind-on-signal off" } # Procedure to get current content of all registers. diff --git a/gdb/testsuite/gdb.base/infcall-failure.exp b/gdb/testsuite/gdb.base/infcall-failure.exp index 5ad179a089d..e2e71649dac 100644 --- a/gdb/testsuite/gdb.base/infcall-failure.exp +++ b/gdb/testsuite/gdb.base/infcall-failure.exp @@ -137,7 +137,7 @@ proc_with_prefix run_cond_hits_segfault_test { async_p non_stop_p } { "Error in testing condition for breakpoint ${bp_1_num}:" \ "The program being debugged was signaled while in a function called from GDB\\." \ "GDB remains in the frame where the signal was received\\." \ - "To change this behavior use \"set unwindonsignal on\"\\." \ + "To change this behavior use \"set unwind-on-signal on\"\\." \ "Evaluation of the expression containing the function" \ "\\(func_segfault\\) will be abandoned\\." \ "When the function is done executing, GDB will silently stop\\."] @@ -166,7 +166,7 @@ proc_with_prefix run_call_hits_segfault_test { async_p non_stop_p } { "${::decimal}\\s+\[^\r\n\]+Segfault here\[^\r\n\]+" \ "The program being debugged was signaled while in a function called from GDB\\." \ "GDB remains in the frame where the signal was received\\." \ - "To change this behavior use \"set unwindonsignal on\"\\." \ + "To change this behavior use \"set unwind-on-signal on\"\\." \ "Evaluation of the expression containing the function" \ "\\(func_segfault\\) will be abandoned\\." \ "When the function is done executing, GDB will silently stop\\."] diff --git a/gdb/testsuite/gdb.base/unwindonsignal.exp b/gdb/testsuite/gdb.base/unwindonsignal.exp index 625b0c4db12..9c1b36e95bc 100644 --- a/gdb/testsuite/gdb.base/unwindonsignal.exp +++ b/gdb/testsuite/gdb.base/unwindonsignal.exp @@ -34,34 +34,50 @@ gdb_test "break stop_here" "Breakpoint \[0-9\]* at .*" gdb_test "continue" "Continuing.*Breakpoint \[0-9\]*, stop_here.*" \ "continue to breakpoint at stop_here" -# Turn on unwindonsignal. -gdb_test_no_output "set unwindonsignal on" \ - "setting unwindonsignal" +# Turn on unwind-on-signal. +gdb_test_no_output "set unwind-on-signal on" \ + "setting unwind-on-signal" -gdb_test "show unwindonsignal" \ +gdb_test "show unwind-on-signal" \ "Unwinding of stack .* is on." \ - "showing unwindonsignal" + "showing unwind-on-signal" + +# For backward compatibility we maintain a 'unwindonsignal' alias for +# 'unwind-on-signal', check it now. +gdb_test "show unwindonsignal" \ + "Unwinding of stack .* is on\\." \ + "showing unwindonsignal alias" + +gdb_test_no_output "set unwindonsignal off" \ + "setting unwindonsignal alias to off" + +gdb_test "show unwind-on-signal" \ + "Unwinding of stack .* is off\\." \ + "showing unwind-on-signal after setting via alias" + +gdb_test_no_output "set unwindonsignal on" \ + "setting unwindonsignal alias to on" # Call function (causing the program to get a signal), and see if gdb handles # it properly. if {[gdb_test "call gen_signal ()" \ "\[\r\n\]*The program being debugged was signaled.*" \ - "unwindonsignal, inferior function call signaled"] != 0} { + "inferior function call signaled"] != 0} { return 0 } # Verify the stack got unwound. gdb_test "bt" \ "#0 *\[x0-9a-f in\]*stop_here \\(.*\\) at .*#1 *\[x0-9a-f in\]*main \\(.*\\) at .*" \ - "unwindonsignal, stack unwound" + "stack unwound" # Verify the dummy frame got removed from dummy_frame_stack. gdb_test_multiple "maint print dummy-frames" \ - "unwindonsignal, dummy frame removed" { + "unwind-on-signal, dummy frame removed" { -re "\[\r\n\]*.*stack=.*code=.*\[\r\n\]+$gdb_prompt $" { - fail "unwindonsignal, dummy frame removed" + fail $gdb_test_name } -re "\[\r\n\]+$gdb_prompt $" { - pass "unwindonsignal, dummy frame removed" + pass $gdb_test_name } } diff --git a/gdb/testsuite/gdb.compile/compile-cplus.exp b/gdb/testsuite/gdb.compile/compile-cplus.exp index 4e887daeb29..22eebb44953 100644 --- a/gdb/testsuite/gdb.compile/compile-cplus.exp +++ b/gdb/testsuite/gdb.compile/compile-cplus.exp @@ -129,7 +129,7 @@ gdb_test "return" "\r\n#0 main .*" "return" \ "Make _gdb_expr\\(__gdb_regs\\*\\) return now\\? \\(y or n\\) " "y" gdb_test "info sym $infcall_pc" "\r\nNo symbol matches .*" "info sym not found" -gdb_test_no_output "set unwindonsignal on" +gdb_test_no_output "set unwind-on-signal on" gdb_test "compile code *(volatile int *) 0 = 0;" \ "The program being debugged was signaled while in a function called from GDB\\.\r\nGDB has restored the context to what it was before the call\\.\r\n.*" \ "compile code segfault second" diff --git a/gdb/testsuite/gdb.compile/compile.exp b/gdb/testsuite/gdb.compile/compile.exp index f3d87cd2605..8822aab2e81 100644 --- a/gdb/testsuite/gdb.compile/compile.exp +++ b/gdb/testsuite/gdb.compile/compile.exp @@ -159,7 +159,7 @@ gdb_test "return" "\r\n#0 main .*" "return" \ "Make _gdb_expr return now\\? \\(y or n\\) " "y" gdb_test "info sym $infcall_pc" "\r\nNo symbol matches .*" "info sym not found" -gdb_test_no_output "set unwindonsignal on" +gdb_test_no_output "set unwind-on-signal on" gdb_test "compile code *(volatile int *) 0 = 0;" \ "The program being debugged was signaled while in a function called from GDB\\.\r\nGDB has restored the context to what it was before the call\\.\r\n.*" \ "compile code segfault second" diff --git a/gdb/testsuite/gdb.cp/gdb2495.exp b/gdb/testsuite/gdb.cp/gdb2495.exp index e3c0cca3175..d7985e5db48 100644 --- a/gdb/testsuite/gdb.cp/gdb2495.exp +++ b/gdb/testsuite/gdb.cp/gdb2495.exp @@ -98,29 +98,29 @@ if {![runto_main]} { # behaviour; it should not. Test both on and off states. # Turn on unwind on signal behaviour. -gdb_test_no_output "set unwindonsignal on" +gdb_test_no_output "set unwind-on-signal on" # Check that it is turned on. -gdb_test "show unwindonsignal" \ +gdb_test "show unwind-on-signal" \ "signal is received while in a call dummy is on.*" \ "turn on unwind on signal" # Check to see if new behaviour interferes with # normal signal handling in inferior function calls. gdb_test "p exceptions.raise_signal(1)" \ - "To change this behavior use \"set unwindonsignal off\".*" \ - "check for unwindonsignal off message" + "To change this behavior use \"set unwind-on-signal off\".*" \ + "check for unwind-on-signal off message" # And reverse - turn off again. -gdb_test_no_output "set unwindonsignal off" +gdb_test_no_output "set unwind-on-signal off" # Check that it is actually turned off. -gdb_test "show unwindonsignal" \ +gdb_test "show unwind-on-signal" \ "signal is received while in a call dummy is off.*" \ "turn off unwind on signal" # Check to see if new behaviour interferes with # normal signal handling in inferior function calls. gdb_test "p exceptions.raise_signal(1)" \ - "To change this behavior use \"set unwindonsignal on\".*" \ - "check for unwindonsignal on message" + "To change this behavior use \"set unwind-on-signal on\".*" \ + "check for unwind-on-signal on message" diff --git a/gdb/testsuite/gdb.fortran/function-calls.exp b/gdb/testsuite/gdb.fortran/function-calls.exp index f9a1efc241f..c4d1f232576 100644 --- a/gdb/testsuite/gdb.fortran/function-calls.exp +++ b/gdb/testsuite/gdb.fortran/function-calls.exp @@ -44,7 +44,7 @@ if {![runto [gdb_get_line_number "post_init"]]} { } # Use inspired by gdb.base/callfuncs.exp. -gdb_test_no_output "set unwindonsignal on" +gdb_test_no_output "set unwind-on-signal on" # Baseline: function and subroutine call with no arguments. gdb_test "p no_arg()" " = .TRUE." diff --git a/gdb/testsuite/gdb.mi/mi-syn-frame.exp b/gdb/testsuite/gdb.mi/mi-syn-frame.exp index 40df312a1cc..1324226412c 100644 --- a/gdb/testsuite/gdb.mi/mi-syn-frame.exp +++ b/gdb/testsuite/gdb.mi/mi-syn-frame.exp @@ -94,7 +94,7 @@ mi_gdb_test "409-stack-list-frames 0 0" \ # mi_gdb_test "410-data-evaluate-expression bar()" \ - ".*410\\^error,msg=\"The program being debugged was signaled while in a function called from GDB.\\\\nGDB remains in the frame where the signal was received.\\\\nTo change this behavior use \\\\\"set unwindonsignal on\\\\\".\\\\nEvaluation of the expression containing the function\\\\n\\(bar\\) will be abandoned.\\\\nWhen the function is done executing, GDB will silently stop.\"" \ + ".*410\\^error,msg=\"The program being debugged was signaled while in a function called from GDB.\\\\nGDB remains in the frame where the signal was received.\\\\nTo change this behavior use \\\\\"set unwind-on-signal on\\\\\".\\\\nEvaluation of the expression containing the function\\\\n\\(bar\\) will be abandoned.\\\\nWhen the function is done executing, GDB will silently stop.\"" \ "call inferior function which raises exception" mi_gdb_test "411-stack-list-frames" "411\\^done,stack=\\\[frame=\{level=\"0\",addr=\"$hex\",func=\"bar\",file=\".*mi-syn-frame.c\",fullname=\"${fullname_syntax}${srcfile}\",line=\"$decimal\",arch=\"$any\"},frame=\{level=\"1\",addr=\"$hex\",func=\"\"\},frame=\{level=\"2\",addr=\"$hex\",func=\"main\",file=\".*mi-syn-frame.c\",fullname=\"${fullname_syntax}${srcfile}\",line=\"$decimal\",arch=\"$any\"}.*\\\]" "backtrace from inferior function at exception" diff --git a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp index c46f8cdb981..26ce59d6335 100644 --- a/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp +++ b/gdb/testsuite/gdb.threads/infcall-from-bp-cond-simple.exp @@ -173,7 +173,7 @@ proc_with_prefix run_bp_cond_segfaults { target_async target_non_stop } { "Error in testing condition for breakpoint ${bp_1_num}:" \ "The program being debugged was signaled while in a function called from GDB\\." \ "GDB remains in the frame where the signal was received\\." \ - "To change this behavior use \"set unwindonsignal on\"\\." \ + "To change this behavior use \"set unwind-on-signal on\"\\." \ "Evaluation of the expression containing the function" \ "\\(function_that_segfaults\\) will be abandoned\\." \ "When the function is done executing, GDB will silently stop\\."] diff --git a/gdb/testsuite/gdb.threads/thread-unwindonsignal.exp b/gdb/testsuite/gdb.threads/thread-unwindonsignal.exp index 68a99f66f1c..a9cedc0c466 100644 --- a/gdb/testsuite/gdb.threads/thread-unwindonsignal.exp +++ b/gdb/testsuite/gdb.threads/thread-unwindonsignal.exp @@ -13,7 +13,7 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see . -# Test use of unwindonsignal when a hand function call that gets interrupted +# Test use of unwind-on-signal when a hand function call that gets interrupted # by a signal in another thread. set NR_THREADS 4 @@ -49,12 +49,12 @@ gdb_test "continue" \ # We want the main thread (hand_call_with_signal) and # thread 1 (sigabrt_handler) to both run. -# Do turn on unwindonsignal. +# Do turn on unwind-on-signal. # We want to test gdb handling of the current thread changing when # unwindonsignal is in effect. -gdb_test_no_output "set unwindonsignal on" \ +gdb_test_no_output "set unwind-on-signal on" \ "setting unwindonsignal" -gdb_test "show unwindonsignal" \ +gdb_test "show unwind-on-signal" \ "Unwinding of stack .* is on." \ "showing unwindonsignal"