From patchwork Mon May 15 19:27:37 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69395 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 7B5263857010 for ; Mon, 15 May 2023 19:28:20 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 7B5263857010 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178900; bh=BL3zQbqTiqBqWMYiIwAhYXq+f4hw9Mce9MnChoQq6Jw=; 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=NZeT9L+hi5+nvhlmY9gspSqD+oBSNyigX9IgAZ3uDWFSOkjlc5h45m+Ln13/W9ynp hvc7PgxPWt9IGNRGUZPy5Srgr2hSDqGjkvVLFKZ1jU1EUe32Gz0HfTpJarM8Mm78gJ a2As4MCY4RjiPylvnRafUmg96go26bjQshrep7kk= 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 79E0A3858D37 for ; Mon, 15 May 2023 19:27:54 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 79E0A3858D37 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-290-e5s4Jj6SMdWP7HcR6Qu5xQ-1; Mon, 15 May 2023 15:27:52 -0400 X-MC-Unique: e5s4Jj6SMdWP7HcR6Qu5xQ-1 Received: by mail-wr1-f71.google.com with SMTP id ffacd0b85a97d-306362a1607so8207610f8f.1 for ; Mon, 15 May 2023 12:27:52 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178871; x=1686770871; 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=BL3zQbqTiqBqWMYiIwAhYXq+f4hw9Mce9MnChoQq6Jw=; b=Df4+F0MfAl7SQxq2Y+WSnEmb2iWTGzm6aejf0p1Lnt9rUobiscPjKqB/bb3St35F6u HqYsRyUtegovc6IVNv7AZi8mHHECr4FCtA8HCSgDOsi/NIwswX1oOgwD9+nFE4ZEg9rJ bM/XKIoYBUL0T7hcczUmPq9DQeUFkCQDbtVzm3262i1DTrPsp7hGofOmFaIjNbZ5s8EN //5isElO0817RyCAtkvJrrnxnNaXhJey2AdcnMHwELuomG20bf2OlTDkRmISpcdsIZII sw6y1LoAVcyo0zUsfp5MDX2fvSm1HczcLoH5Wnkm4tTmVKYIu65Li1OmVCH3ne+wZDO7 xdNQ== X-Gm-Message-State: AC+VfDwghTEd98C3hop+ByOf1R3Z025XVmQyrrR6abT02CufG7biV3io NySc+YWstjEVrC5zv4vHWvAglHC26JUl/gpCwBrRcz13jpV3K3POGv4B7lfgE0iJyz6xulQZgl9 6Pbs6zk5ZimyJXZbHpEPsYZ+v8Ls/THf2dFh9boqKtg0keIB1bLOC5ZpoGS/wHb0QAmc3kiCDou EUGiZUqg== X-Received: by 2002:a5d:510a:0:b0:304:79c1:725d with SMTP id s10-20020a5d510a000000b0030479c1725dmr24252726wrt.45.1684178871467; Mon, 15 May 2023 12:27:51 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ7ayNwtPzv7IN83aBxmpLrtSZW59dnmRh6hdfzdG9W/sT05etLR6X5b7RXFaZC1BQ4aHnb6XQ== X-Received: by 2002:a5d:510a:0:b0:304:79c1:725d with SMTP id s10-20020a5d510a000000b0030479c1725dmr24252713wrt.45.1684178871106; Mon, 15 May 2023 12:27:51 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id a3-20020adfeec3000000b0030649242b72sm45921wrp.113.2023.05.15.12.27.50 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:27:50 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 1/9] gdb: create_breakpoint: assert for a valid thread-id Date: Mon, 15 May 2023 20:27:37 +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=-11.6 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, RCVD_IN_MSPIKE_H2, 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" Add an assert into create_breakpoint (breakpoint.c) that the thread argument is a valid thread-id; either a value greater than zero, or -1 to indicate any thread. The thread is ignored if parse_extra is true though, so take that into account too. There should be no user visible changes after this commit. --- gdb/breakpoint.c | 4 ++++ gdb/breakpoint.h | 6 ++++++ 2 files changed, 10 insertions(+) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index fdb184ae81f..31448e0a895 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -9027,6 +9027,10 @@ create_breakpoint (struct gdbarch *gdbarch, gdb_assert (ops != NULL); + /* If PARSE_EXTRA is true then the thread-id will be parsed from + EXTRA_STRING, otherwise, ensure THREAD is valid. */ + gdb_assert (parse_extra || thread == -1 || thread > 0); + /* If extra_string isn't useful, set it to NULL. */ if (extra_string != NULL && *extra_string == '\0') extra_string = NULL; diff --git a/gdb/breakpoint.h b/gdb/breakpoint.h index 7c5cf3f2bef..e57dc2523fb 100644 --- a/gdb/breakpoint.h +++ b/gdb/breakpoint.h @@ -1530,6 +1530,12 @@ enum breakpoint_create_flags the FORCE_CONDITION parameter is ignored and the corresponding argument is parsed from EXTRA_STRING. + The THREAD should be a global thread number, the created breakpoint will + only apply for that thread. If the breakpoint should apply for all + threads then pass -1. However, if PARSE_EXTRA is non-zero then the + THREAD parameter is ignored and an optional thread number will be parsed + from EXTRA_STRING. + If INTERNAL is non-zero, the breakpoint number will be allocated from the internal breakpoint count. From patchwork Mon May 15 19:27:38 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69399 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 E0AD1385354C for ; Mon, 15 May 2023 19:28:54 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org E0AD1385354C DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178934; bh=nZanVJWAcbm0KzOzjv9s73VHPfwSg3O8veu+2Io28Gw=; 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=YYiEP66DBHubHF8wSH3LrpEdoDrxvDNStDP2eL/zfQfo4YdXHgaYT9EzxyTQDmSl1 PV2fLx37hbDfgicxrUTR433JFUow4vRR/iOaAOLD2QuSicqV8XNjfEGMhlDUjaK2Ke saU3ckR8jzWwXI9F6Ci/8kuzwsisVXymKjlOgPic= 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 04C16385843A for ; Mon, 15 May 2023 19:27:55 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 04C16385843A 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-552-JOdmiqVxM1iv8h4FtmqfjA-1; Mon, 15 May 2023 15:27:54 -0400 X-MC-Unique: JOdmiqVxM1iv8h4FtmqfjA-1 Received: by mail-wm1-f69.google.com with SMTP id 5b1f17b1804b1-3f509037a45so280725e9.1 for ; Mon, 15 May 2023 12:27:54 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178873; x=1686770873; 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=nZanVJWAcbm0KzOzjv9s73VHPfwSg3O8veu+2Io28Gw=; b=EG50SL/BJqpJTc8nlZCsULSiBFC56velAMBeksEVlarzD12qKDIoY5sXgDsIeSTsvk iQqSA24AE7zC6UGnLyo/jIMflTHLbBQKmlgIu2ofMl6LTv2geA+TITlNUFoH9zvB8UVK HHTmHtd1i5pV/fWc0H7oEwNpNSd7NQhcSpT/zH64cPLFXZmKO3FSB1ZvQF/Ovex28yTD LLWqCx2L2SWpYOr6qkRuZxFllByJ/3ej8ZaKdF64GIDx5HJh+Xn+o335KxXE8Ov+mJU3 ll32Tf/0DcXf/YjrmNn7P5XEX9SDRMZOsCuvornp+G0fzCyjX5pUtqnhoym3ZpX1CJYV ydmA== X-Gm-Message-State: AC+VfDxcNgNKOllV9kWjpD8GM2xzvXSkZruzVJkzdHq1ZWtlEPuG6CRn AZDbDCicqbnbF1cpzsgECfkJNu92qkEHHxzPv8ogRDfeprfADoGReHMblQ/bFi1d+ZO2G7Uzha+ GF6eUJR7Rai9l3Lwdb2TPB0NuEjme5LScMrY2SPJJVmufaD56QNLywCMonT4YL3VuEi0ajeCT82 UlRpCrKg== X-Received: by 2002:a5d:4ac1:0:b0:307:869c:99ce with SMTP id y1-20020a5d4ac1000000b00307869c99cemr22648110wrs.21.1684178872931; Mon, 15 May 2023 12:27:52 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ7XfRan+TIWWGowh41SiD1uxcAQ8jPIkvFuEB8bE47feTp6O8EcCxIxfMAMKAs+cpMDlDkehw== X-Received: by 2002:a5d:4ac1:0:b0:307:869c:99ce with SMTP id y1-20020a5d4ac1000000b00307869c99cemr22648104wrs.21.1684178872636; Mon, 15 May 2023 12:27:52 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id i8-20020a5d5228000000b003078a3f3a24sm54997wra.114.2023.05.15.12.27.51 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:27:52 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 2/9] gdb: create_breakpoint: asserts relating to extra_string/parse_extra Date: Mon, 15 May 2023 20:27:38 +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=-11.6 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, RCVD_IN_MSPIKE_H2, 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" The goal of this commit is to better define the API for create_breakpoint especially around the use of extra_string and parse_extra. This will be useful in the next commit when I plan to make some changes to create_breakpoint. This commit makes one possibly breaking change: until this commit it was possible to create thread-specific dprintf breakpoint like this: (gdb) dprintf call_me, thread 1 "%s", "hello" Dprintf 2 at 0x401152: file /tmp/hello.c, line 8. (gdb) info breakpoints Num Type Disp Enb Address What 2 dprintf keep y 0x0000000000401152 in call_me at /tmp/hello.c:8 thread 1 stop only in thread 1 printf "%s", "hello" (gdb) this feature of dprintf was not documented, was not tested, and is slightly different in syntax to how we create thread specific breakpoints and/or watchpoints -- the thread condition appears after the first ','. I believe that this worked at all was simply by luck. We happen to pass the parse_extra flag as true from dprintf_command to create_breakpoint. So in this commit I made the choice change this. We now pass parse_extra as false from dprintf_command to create_breakpoint. With this done it is assumed that the only thing in the extra_string is the dprintf format and arguments. Beyond this change I've updated the comment on create_breakpoint in breakpoint.h, and I've then added some asserts into create_breakpoint as well as moving around some of the error handling. - We now assert on the incoming argument values, - I've moved an error check to sit after the call to find_condition_and_thread_for_sals, this ensures the extra_string was parsed correctly, In dprintf_command: - We now throw an error if there is no format string after the dprintf location. This error was already being thrown, but was being caught later in the process. With this change we catch the missing string earlier, - And, as mentioned earlier, we pass parse_extra as false when calling create_breakpoint, In create_tracepoint_from_upload: - We now throw an error if the parsed location doesn't completely consume the addr_str variable. This error has now effectively moved out of create_breakpoint. --- gdb/breakpoint.c | 42 +++++++++++++++++++++++++++++------------- gdb/breakpoint.h | 38 ++++++++++++++++++++++++-------------- 2 files changed, 53 insertions(+), 27 deletions(-) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index 31448e0a895..4d182ac0143 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -9035,6 +9035,17 @@ create_breakpoint (struct gdbarch *gdbarch, if (extra_string != NULL && *extra_string == '\0') extra_string = NULL; + /* A bp_dprintf must always have an accompanying EXTRA_STRING containing + the dprintf format and arguments -- PARSE_EXTRA should always be false + in this case. + + For all other breakpoint types, EXTRA_STRING should be nullptr unless + PARSE_EXTRA is true. */ + gdb_assert ((type_wanted == bp_dprintf + && extra_string != nullptr && !parse_extra) + || (type_wanted != bp_dprintf + && (extra_string == nullptr || parse_extra))); + try { ops->create_sals_from_location_spec (locspec, &canonical); @@ -9098,6 +9109,8 @@ create_breakpoint (struct gdbarch *gdbarch, if (parse_extra) { + gdb_assert (type_wanted != bp_dprintf); + gdb::unique_xmalloc_ptr rest; gdb::unique_xmalloc_ptr cond; @@ -9105,15 +9118,15 @@ create_breakpoint (struct gdbarch *gdbarch, find_condition_and_thread_for_sals (lsal.sals, extra_string, &cond, &thread, &task, &rest); + + if (rest.get () != nullptr && *(rest.get ()) != '\0') + error (_("Garbage '%s' at end of command"), rest.get ()); + cond_string_copy = std::move (cond); extra_string_copy = std::move (rest); } else { - if (type_wanted != bp_dprintf - && extra_string != NULL && *extra_string != '\0') - error (_("Garbage '%s' at end of location"), extra_string); - /* Check the validity of the condition. We should error out if the condition is invalid at all of the locations and if it is not forced. In the PARSE_EXTRA case above, this @@ -9322,20 +9335,17 @@ dprintf_command (const char *arg, int from_tty) /* If non-NULL, ARG should have been advanced past the location; the next character must be ','. */ - if (arg != NULL) + if (arg == nullptr || arg[0] != ',' || arg[1] == '\0') + error (_("Format string required")); + else { - if (arg[0] != ',' || arg[1] == '\0') - error (_("Format string required")); - else - { - /* Skip the comma. */ - ++arg; - } + /* Skip the comma. */ + ++arg; } create_breakpoint (get_current_arch (), locspec.get (), - NULL, 0, arg, false, 1 /* parse arg */, + NULL, -1, arg, false, 0 /* parse arg */, 0, bp_dprintf, 0 /* Ignore count */, pending_break_support, @@ -13910,6 +13920,12 @@ create_tracepoint_from_upload (struct uploaded_tp *utp) location_spec_up locspec = string_to_location_spec (&addr_str, current_language); + + + gdb_assert (addr_str != nullptr); + if (*addr_str != '\0') + error (_("Garbage '%s' at end of location"), addr_str); + if (!create_breakpoint (get_current_arch (), locspec.get (), utp->cond_string.get (), -1, addr_str, diff --git a/gdb/breakpoint.h b/gdb/breakpoint.h index e57dc2523fb..2adcc9fa338 100644 --- a/gdb/breakpoint.h +++ b/gdb/breakpoint.h @@ -1515,26 +1515,36 @@ enum breakpoint_create_flags functions for setting a breakpoint at LOCSPEC. This function has two major modes of operations, selected by the - PARSE_EXTRA parameter. + PARSE_EXTRA and WANTED_TYPE parameters. - If PARSE_EXTRA is zero, LOCSPEC is just the breakpoint's location - spec, with condition, thread, and extra string specified by the - COND_STRING, THREAD, and EXTRA_STRING parameters. + When WANTED_TYPE is not bp_dprintf the following rules apply: - If PARSE_EXTRA is non-zero, this function will attempt to extract - the condition, thread, and extra string from EXTRA_STRING, ignoring - the similarly named parameters. + If PARSE_EXTRA is zero, LOCSPEC is just the breakpoint's location + spec, with condition, thread, and extra string specified by the + COND_STRING, THREAD, and EXTRA_STRING parameters. - If FORCE_CONDITION is true, the condition is accepted even when it is - invalid at all of the locations. However, if PARSE_EXTRA is non-zero, - the FORCE_CONDITION parameter is ignored and the corresponding argument - is parsed from EXTRA_STRING. + If PARSE_EXTRA is non-zero, this function will attempt to extract the + condition, thread, and extra string from EXTRA_STRING, ignoring the + similarly named parameters. + + When WANTED_TYPE is bp_dprintf the following rules apply: + + PARSE_EXTRA must always be zero, LOCSPEC is just the breakpoint's + location spec, with condition, thread, and extra string (which + contains the dprintf format and arguments) specified by the + COND_STRING, THREAD, and EXTRA_STRING parameters. + + If FORCE_CONDITION is true, the condition (in COND_STRING) is accepted + even when it is invalid at all of the locations. However, if + PARSE_EXTRA is non-zero and WANTED_TYPE is not bp_dprintf, the + FORCE_CONDITION parameter is ignored and the corresponding argument is + parsed from EXTRA_STRING. The THREAD should be a global thread number, the created breakpoint will only apply for that thread. If the breakpoint should apply for all - threads then pass -1. However, if PARSE_EXTRA is non-zero then the - THREAD parameter is ignored and an optional thread number will be parsed - from EXTRA_STRING. + threads then pass -1. However, if PARSE_EXTRA is non-zero and + WANTED_TYPE is not bp_dprintf, then the THREAD parameter is ignored and + an optional thread number will be parsed from EXTRA_STRING. If INTERNAL is non-zero, the breakpoint number will be allocated from the internal breakpoint count. From patchwork Mon May 15 19:27:39 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69398 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 91A863853823 for ; Mon, 15 May 2023 19:28:52 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 91A863853823 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178932; bh=DNJRsc2MGRFsbyKfRBnUxD/ubLoWKKujy2nErAu/wug=; 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=jUBRvcBZB7JKnYX2HORyCI+ahZuLwgAEqH9cWL7JFlRNu9imXIJC3oQyujBBcF+X5 mMSZ9y17dY4UNI5jgYXjTapOPSFP0yv1exsYXeGi2AFbFurUuv45z0EhldLyJtCKnQ UpuI47zJ4NhY9WDjzVWi6CK4XS+ezYo6/P6z5Adw= 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 83154385770D for ; Mon, 15 May 2023 19:27:57 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 83154385770D Received: from mail-wm1-f72.google.com (mail-wm1-f72.google.com [209.85.128.72]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-413-HM5RRzipMvyhojlkdpIxpQ-1; Mon, 15 May 2023 15:27:56 -0400 X-MC-Unique: HM5RRzipMvyhojlkdpIxpQ-1 Received: by mail-wm1-f72.google.com with SMTP id 5b1f17b1804b1-3f426ffdbc6so53331445e9.3 for ; Mon, 15 May 2023 12:27:55 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178874; x=1686770874; 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=DNJRsc2MGRFsbyKfRBnUxD/ubLoWKKujy2nErAu/wug=; b=fQifMazNVux/8Dx/nenVwzBS2+QRFf+ARnapUzBUAVrCoHoYb3jsFCF0zsQjtDaQXK YLrOsIPTmiryCcfTSvqBAQ2IOqvWM5rSzhORyJeA4+Z0W78/djBhknb8utwka/s8VCIj N5PKywabu+9LL8G4mqwryygwTSzbPsEqvZH/SzR8kF041ybwd1ur02TiqBW35JyShj+Q JpjXzyrB+E4wzbKayptqZZdduvreambiulR2smiO4gSfwBzoRhx6XbzYd+e96x93IqCh pKvHd7jtnvvpFFf9TUdb0MYRSvmIVXzS4i/+oq4VdXeU1TnnicT1aZoZfMMxCrYfn073 /Dbw== X-Gm-Message-State: AC+VfDydUlUMDSUWhJf66/qA8OYEVQOSEN1DLNUChUsDjlIhIWPMPjU8 MipSfTk3M3UKAOpY0yDOWtCzTgi2hggdaJPcT3HnYuBWG+J2HKDgs45fJis2rdwbHIEOWTYqQWd kRxtu9diWYxZPBj9mbjEm21kFO76RnDw5Jba5mvZN5FbIxBPrxXoUmMOHxWJEMGpPE2XCuQs8ok w2WkYtRw== X-Received: by 2002:a7b:ca51:0:b0:3f4:2174:b29d with SMTP id m17-20020a7bca51000000b003f42174b29dmr22086785wml.1.1684178874644; Mon, 15 May 2023 12:27:54 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ4LFjVQ97uwdV6dP2JYe4qV1CP0EkI/ObLzYW4DKnp6IcDvwJrE2m4L5w8uAtDmsrYCETsQUw== X-Received: by 2002:a7b:ca51:0:b0:3f4:2174:b29d with SMTP id m17-20020a7bca51000000b003f42174b29dmr22086770wml.1.1684178874241; Mon, 15 May 2023 12:27:54 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id 12-20020a05600c020c00b003f50e29bce3sm257114wmi.48.2023.05.15.12.27.53 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:27:53 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 3/9] gdb: change 'if' to gdb_assert in update_dprintf_command_list Date: Mon, 15 May 2023 20:27:39 +0100 Message-Id: <9ab5e8275f42891b7800f5bff9122b313b1658f0.1684178710.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=-11.6 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" I noticed in update_dprintf_command_list that we handle the case where the bp_dprintf style breakpoint doesn't have a format and args string. However, I don't believe such a situation is possible. The obvious approach certainly already catches this case: (gdb) dprintf main Format string required If it is possible to create a dprintf breakpoint without a format and args string then I think we should be catching this case and handling it at creation time, rather than having GDB just ignore the situation later on. And so, I propose that we change the 'if' that ignores the case where the format/args string is empty, and instead assert that we do always have a format/args string. The original code, that handled an empty format/args string has existed since commit e7e0cddfb0d4, which is when dprintf support was added to GDB. If I'm correct and this situation can't ever happen then there should be no user visible changes after this commit. --- gdb/breakpoint.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index 4d182ac0143..0827b587d5f 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -8383,8 +8383,9 @@ update_dprintf_command_list (struct breakpoint *b) const char *dprintf_args = b->extra_string.get (); gdb::unique_xmalloc_ptr printf_line = nullptr; - if (!dprintf_args) - return; + /* Trying to create a dprintf breakpoint without a format and args + string should be detected at creation time. */ + gdb_assert (dprintf_args != nullptr); dprintf_args = skip_spaces (dprintf_args); From patchwork Mon May 15 19:27:40 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69396 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 88FA13854142 for ; Mon, 15 May 2023 19:28:30 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 88FA13854142 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178910; bh=+QTPgKqm3qDcTYnvIUX1xDLdxtWpSpqooPEAUApNYFI=; 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=vdQyJP/eQ5+AVg1bp1LQK3/CN2drzJyyKuQDTa0QsdKSrLcwI0SZMNxVAs3TxsJ8O NtFwK+ZQ8EE4+Jdcq4HijNIS96dmhpxAkm7sJ0Pl/zJlCJCANr68nkny3gjDoCAJhf lqwl7yPAL+xuE1XdgdXO0OYC8T8EMTeI5uvVvnz0= 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 167DA3857720 for ; Mon, 15 May 2023 19:28:00 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 167DA3857720 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-640-K9FKz4SxP-q2fxS1KcU1AA-1; Mon, 15 May 2023 15:27:57 -0400 X-MC-Unique: K9FKz4SxP-q2fxS1KcU1AA-1 Received: by mail-wm1-f70.google.com with SMTP id 5b1f17b1804b1-3f426ffdbc6so53331485e9.3 for ; Mon, 15 May 2023 12:27:57 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178876; x=1686770876; 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=+QTPgKqm3qDcTYnvIUX1xDLdxtWpSpqooPEAUApNYFI=; b=fYygntTMW9rSKgEFTHX8zK89Y9GgbHqna7SrcO77SR532sTGIL/WyBoabvQNfJ+06s 1Qks6YP/2LKH86t8tvWYOJdRmsnw84BU0AS/Pz7k1X/UKUF1oQMgBxvLp8L8f7WRjdCk IMqCDipYS1mAyuhFoo31lRcwHr8YDePe1Y0L9I3TU/dw+GW2xIxpMet7mzXRxcSErQeW fHWK4Osc3LPozSYK+8H9GwhJDqHyWON0IMiKiFCUtqPgkOLJvyU5GrqDMeH+OehCOXyQ 4n3dQgua4nfO8CM4c7aAbJ3lWBZZvP8jyDI0fmgdiLk8L7Ps/APxzdjYDC7b88PQKUZ8 0uMQ== X-Gm-Message-State: AC+VfDwnxGzlEUWVLNg0bzgIM3qEaVm9/1fVFUmiVj041q5XfASfJ67n jN96WY5phFqCeXTL2X3EfALNi19Uo57AhzC/WSgE4CDshMN14xbz6Z2LTHdOqrBFF746s6uV8Qs 3ePzgDDzk3oV9T53OnomSHjO/yVa3FiwwONxaRdn0G4NWz+KzHuDyz/qzD9iAW+oPlAUdB7go21 PyeyzksQ== X-Received: by 2002:a1c:7712:0:b0:3f4:1dd9:e9e0 with SMTP id t18-20020a1c7712000000b003f41dd9e9e0mr22032866wmi.7.1684178876525; Mon, 15 May 2023 12:27:56 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ5lTv8TI0ghpjeFzdlrIiEV8s+QVGW+es2cqbDG2XIcOTF2Y8eZjloL7MXRxgPMJ7nrpr5ICQ== X-Received: by 2002:a1c:7712:0:b0:3f4:1dd9:e9e0 with SMTP id t18-20020a1c7712000000b003f41dd9e9e0mr22032856wmi.7.1684178876248; Mon, 15 May 2023 12:27:56 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id s3-20020a7bc383000000b003f4bef65a65sm309839wmj.28.2023.05.15.12.27.55 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:27:55 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 4/9] gdb: build dprintf commands just once in code_breakpoint constructor Date: Mon, 15 May 2023 20:27:40 +0100 Message-Id: <10b2c00a69eec9fe610e00f1ae10e1efad991737.1684178710.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=-11.7 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, RCVD_IN_MSPIKE_H2, 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" I noticed in code_breakpoint::code_breakpoint that we are calling update_dprintf_command_list once for each breakpoint location, when we really only need to call this once per breakpoint -- the data updated by this function, the breakpoint command list -- is per breakpoint, not per breakpoint location. Calling update_dprintf_command_list multiple times is just wasted effort, there's no per location error checking, we don't even pass the current location to the function. This commit moves the update_dprintf_command_list call outside of the per-location loop. I have also changes the 'if' that handles the case where the extra_string (which holds the format/args) is empty. I don't believe that this situation can ever arise -- and if it does we should be catching it earlier and throwing an error at that point. There should be no user visible changes after this commit. --- gdb/breakpoint.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index 0827b587d5f..d03c55587be 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -8544,19 +8544,17 @@ code_breakpoint::code_breakpoint (struct gdbarch *gdbarch_, /* Do not set breakpoint locations conditions yet. As locations are inserted, they get sorted based on their addresses. Let the list stabilize to have reliable location numbers. */ + } - /* Dynamic printf requires and uses additional arguments on the - command line, otherwise it's an error. */ - if (type == bp_dprintf) - { - if (extra_string != nullptr) - update_dprintf_command_list (this); - else - error (_("Format string required")); - } - else if (extra_string != nullptr) - error (_("Garbage '%s' at end of command"), extra_string.get ()); + /* Dynamic printf requires and uses additional arguments on the + command line, otherwise it's an error. */ + if (type == bp_dprintf) + { + gdb_assert (extra_string != nullptr); + update_dprintf_command_list (this); } + else if (extra_string != nullptr) + error (_("Garbage '%s' at end of command"), extra_string.get ()); /* The order of the locations is now stable. Set the location condition using the location's number. */ From patchwork Mon May 15 19:27:41 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69401 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 B44DE38555A7 for ; Mon, 15 May 2023 19:29:25 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B44DE38555A7 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178965; bh=U5uwxB0PtTniEt8wXUeFtzeCiCa0Ohe2WV+Ily+kHGc=; 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=EzCovOiOv8JkWd6d3vGyZpHHhbv+gDf9cusy5vVf9CKj0xYJ1x0QsYvlpDp+s1JoM keocggH8cJnKuX3eShL+Y2J8rctepUkVmsJADv698g11WrgIOl7q28k76Llxb+sFpP 1PZnro7NC+4hzT8IHbBtgCtLXA4kw8rCxA2yg9MU= 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 7283D385772F for ; Mon, 15 May 2023 19:28:02 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 7283D385772F 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-38-qHF4xwrmOvCjvoUoo8f1IA-1; Mon, 15 May 2023 15:28:01 -0400 X-MC-Unique: qHF4xwrmOvCjvoUoo8f1IA-1 Received: by mail-wm1-f71.google.com with SMTP id 5b1f17b1804b1-3f4245ffbb4so51376915e9.3 for ; Mon, 15 May 2023 12:28:00 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178880; x=1686770880; 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=U5uwxB0PtTniEt8wXUeFtzeCiCa0Ohe2WV+Ily+kHGc=; b=cmMJ4tMdGbqjQSeqKw7rINSRI/IbQyvPYqPOi0nQ96R098Abz2tjSuRcerjqOyEmT7 2YTPxNF0TPxFuzo/H2lC35Qc5w8JTdybNCEJ8eOtzWbh9IGGGzybuMZC16Zt4fOBRR7n O2Iw2kQlwxt4p77vKvtijqpbbaUtxrbhHbhxhrzSVxapxFHPCoJ4IIQB/OkN3rTtfLTc 5Xy/1wMez1cgMhl1CULAyNWwLJj7wwCJbhpOvPyOiKblRpN+gd/I7aO0JVTs8nWwGEzO cZx+lZjo3VpilpZWzViOETTKd9GteQKVer08F2A00JDNgsuht1iuMYWXbwdoFO16tMx4 u+ZA== X-Gm-Message-State: AC+VfDw5rzE8F+cFZVnRAzefG/NVikJizMZL+ELhZfsZtC1Jf6b6uw6y 8hVMu8STSm7NbOe3xGgMAr5iJRqGxMZxSpGVFo29Uot6fQxzTvbFNg4JjhFesHXTGGUMvx7p4aI uGdhGuz8t/Ex8vSjzHR+gGDYgNJX21/REq7Z3gAuoX87qmrVAS2RUyHpnR6NXzkuKGGg34A0rib Zho4l0xg== X-Received: by 2002:a05:600c:2191:b0:3f4:1a97:ab1d with SMTP id e17-20020a05600c219100b003f41a97ab1dmr25543065wme.1.1684178879663; Mon, 15 May 2023 12:27:59 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ6ku48GQd9t3z3TkYUpAWQdfwuXOnP0b0lM6TM0WzGoHtO+QxN7jVeH9LGakpyEgtlyrnd2Mg== X-Received: by 2002:a05:600c:2191:b0:3f4:1a97:ab1d with SMTP id e17-20020a05600c219100b003f41a97ab1dmr25543053wme.1.1684178879192; Mon, 15 May 2023 12:27:59 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id y13-20020a1c4b0d000000b003f19b3d89e9sm285723wma.33.2023.05.15.12.27.58 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:27:58 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 5/9] gdb: don't display inferior list for pending breakpoints Date: Mon, 15 May 2023 20:27:41 +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=-11.7 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, RCVD_IN_MSPIKE_H2, 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" I noticed that in the 'info breakpoints' output, GDB sometimes prints the inferior list for pending breakpoints, this doesn't seem right to me. A pending breakpoint has no locations (at least, as far as we display things in the 'info breakpoints' output), so including an inferior list seems odd. Here's what I see right now: (gdb) info breakpoint 5 Num Type Disp Enb Address What 5 breakpoint keep y foo inf 1 (gdb) It's the 'inf 1' at the end of the line that I'm objecting too. To trigger this behaviour we need to be in a multi-inferior debug session. The breakpoint must have been non-pending at some point in the past, and so have a location assigned to it. The breakpoint becomes pending again as a result of a shared library being unloaded. When this happens the location itself is marked pending (via bp_location::shlib_disabled). In print_one_breakpoint_location, in order to print the inferior list we check that the breakpoint has a location, and that we have multiple inferiors, but we don't check if the location itself is pending. This commit adds that check, which means the output is now: (gdb) info breakpoint 5 Num Type Disp Enb Address What 5 breakpoint keep y foo (gdb) Which I think makes more sense. --- gdb/breakpoint.c | 2 +- gdb/testsuite/gdb.multi/pending-bp-lib.c | 22 +++++ gdb/testsuite/gdb.multi/pending-bp.c | 66 +++++++++++++ gdb/testsuite/gdb.multi/pending-bp.exp | 115 +++++++++++++++++++++++ 4 files changed, 204 insertions(+), 1 deletion(-) create mode 100644 gdb/testsuite/gdb.multi/pending-bp-lib.c create mode 100644 gdb/testsuite/gdb.multi/pending-bp.c create mode 100644 gdb/testsuite/gdb.multi/pending-bp.exp diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index d03c55587be..f9020c49d90 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -6467,7 +6467,7 @@ print_one_breakpoint_location (struct breakpoint *b, } } - if (loc != NULL && !header_of_multiple) + if (loc != NULL && !header_of_multiple && !loc->shlib_disabled) { std::vector inf_nums; int mi_only = 1; diff --git a/gdb/testsuite/gdb.multi/pending-bp-lib.c b/gdb/testsuite/gdb.multi/pending-bp-lib.c new file mode 100644 index 00000000000..15d1b9833dd --- /dev/null +++ b/gdb/testsuite/gdb.multi/pending-bp-lib.c @@ -0,0 +1,22 @@ +/* Copyright 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 . */ + +int global_var = 0; + +void +foo (int arg) +{ + global_var = arg; +} diff --git a/gdb/testsuite/gdb.multi/pending-bp.c b/gdb/testsuite/gdb.multi/pending-bp.c new file mode 100644 index 00000000000..ca8a3c20b72 --- /dev/null +++ b/gdb/testsuite/gdb.multi/pending-bp.c @@ -0,0 +1,66 @@ +/* Copyright 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 + +void +breakpt () +{ + /* Nothing. */ +} + +volatile int global_counter = 0; + +volatile int call_count = 1; + +int +main (void) +{ + void *handle; + void (*func)(int); + + /* Some filler work so that we don't initially stop on the breakpt call + below. */ + ++global_counter; + + breakpt (); /* Break before open. */ + + /* Now load the shared library. */ + handle = dlopen (SHLIB_NAME, RTLD_LAZY); + if (handle == NULL) + abort (); + + breakpt (); /* Break after open. */ + + /* Find the function symbol. */ + func = (void (*)(int)) dlsym (handle, "foo"); + + for (; call_count > 0; --call_count) + { + /* Call the library function. */ + func (1); + } + + breakpt (); /* Break before close. */ + + /* Unload the shared library. */ + if (dlclose (handle) != 0) + abort (); + + breakpt (); /* Break after close. */ + + return 0; +} diff --git a/gdb/testsuite/gdb.multi/pending-bp.exp b/gdb/testsuite/gdb.multi/pending-bp.exp new file mode 100644 index 00000000000..4a65f79cfc4 --- /dev/null +++ b/gdb/testsuite/gdb.multi/pending-bp.exp @@ -0,0 +1,115 @@ +# Copyright 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 related to pending breakpoints in a multi-inferior environment. + +require allow_shlib_tests !use_gdb_stub + +standard_testfile + +set libname $testfile-lib +set srcfile_lib $srcdir/$subdir/$libname.c +set binfile_lib [standard_output_file $libname.so] + +if { [gdb_compile_shlib $srcfile_lib $binfile_lib {}] != "" } { + untested "failed to compile shared library 1" + return -1 +} + +set binfile_lib_target [gdb_download_shlib $binfile_lib] + +if { [build_executable "failed to prepare" $testfile $srcfile \ + [list debug \ + additional_flags=-DSHLIB_NAME=\"$binfile_lib_target\" \ + shlib_load pthreads]] } { + return -1 +} + +# Used to override delete_breakpoints when we don't want breakpoints +# deleted. +proc do_nothing {} {} + +# Start two inferiors, both running the same test binary. The arguments +# INF_1_STOP and INF_2_STOP are source code patterns that are passed to +# gdb_get_line_number to figure out where each inferior should be stopped. +# +# This proc does a clean_restart and leaves inferior 2 selected. Also the +# 'breakpoint pending' flag is enabled, so pending breakpoints can be created +# without GDB prompting the user. +proc do_test_setup { inf_1_stop inf_2_stop } { + clean_restart ${::binfile} + + gdb_locate_shlib $::binfile_lib + + if ![runto_main] { + return false + } + + gdb_breakpoint [gdb_get_line_number ${inf_1_stop}] temporary + gdb_continue_to_breakpoint "move inferior 1 into position" + + gdb_test "add-inferior -exec ${::binfile}" \ + "Added inferior 2.*" "add inferior 2" + gdb_test "inferior 2" "Switching to inferior 2 .*" "switch to inferior 2" + + with_override delete_breakpoints do_nothing { + if {![runto_main]} { + return false + } + } + + gdb_breakpoint [gdb_get_line_number ${inf_2_stop}] temporary + gdb_continue_to_breakpoint "move inferior 2 into position" + + gdb_test_no_output "set breakpoint pending on" + + return true +} + +# Check that when a breakpoint is in the pending state, but that breakpoint +# does have some locations (those locations themselves are pending), GDB +# doesn't display the inferior list in the 'info breakpoints' output. +proc_with_prefix test_no_inf_display {} { + do_test_setup "Break before close" "Break before open" + + # Create a breakpoint on 'foo'. This will find a location in inferior 1, + # but will not find a location in inferior 2 -- the shared library is not + # yet loaded in that inferior. + gdb_breakpoint "foo" + set bpnum [get_integer_valueof "\$bpnum" "*INVALID*" \ + "get foo breakpoint number"] + + # Check the 'info breakpoints' output. Notice we display the inferior + # list at the end of the breakpoint line. + gdb_test "info breakpoint $bpnum" \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+$::hex\\s+\]*>\\s+inf 1" \ + "check info breakpoints while breakpoint is inserted" + + # Ensure we are in inferior 1, move the inferior forward until the shared + # library has been unloaded. The breakpoint on 'foo' will return to the + # pending state. + gdb_test "inferior 1" "Switching to inferior 1 .*" + gdb_breakpoint [gdb_get_line_number "Break after close"] temporary + gdb_continue_to_breakpoint "after close library" + + # Check the 'info breakpoints' output, check there is no 'inf 1' at the + # end of the breakpoint line. + gdb_test "info breakpoint $bpnum" \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+\\s+foo" \ + "check info breakpoints while breakpoint is pending" +} + +# Run all the tests. +test_no_inf_display From patchwork Mon May 15 19:27:42 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69397 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 5577938555A5 for ; Mon, 15 May 2023 19:28:51 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 5577938555A5 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178931; bh=P8ajJcKyxoQuLkXsw29++sgxNk6DPHQywzbDphB6swE=; 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=jIKrfCU8eMt7X9yz6SLirCNnXBw6961NBVWi44L1WF93tWXrV1y82E4KDO1rn3a3u Bnld5Jx3To2VuSW+vhoKM9Xob35bMZZ8iWynptav0++5jp8KewqmbFy1TzEP+tr82U TV/RtQJGXA2XUnbUreubDkIMKGhXPVCrpwNmkpgE= 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 0E4D63858C36 for ; Mon, 15 May 2023 19:28:07 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 0E4D63858C36 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-120-UnsHfoJMO0Ou4llfwxhyTg-1; Mon, 15 May 2023 15:28:05 -0400 X-MC-Unique: UnsHfoJMO0Ou4llfwxhyTg-1 Received: by mail-wr1-f71.google.com with SMTP id ffacd0b85a97d-30793c16c78so6258874f8f.3 for ; Mon, 15 May 2023 12:28:05 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178884; x=1686770884; 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=P8ajJcKyxoQuLkXsw29++sgxNk6DPHQywzbDphB6swE=; b=aGRPU1R+c16cVniH0F5Ch97dq4LbDihW+3f874CcGqFdknF6odbhKmW5l9RhI2Gw0G V2Y0XINQxNH4VrbWS/4vQ29fQ+MhDkRi+k1DtHhq565r50qZXBd6bc1uH9rdBuPpuJh7 mYg3nZ9K3adriyuDrAOCLBpVJ6k4nu1rkThd5Le9F3RBTP8IYXKY1vvgeFiFGFGBssZT bUru19urC6hjm36GbSVtv+6kQmIJgUrBhOcxDkNkJNQk6PKDQjyA//+RpTN4LuoBWn4U yaqCycFn8iXdqIwiwh25CE6Ddw78jppGDbbXCXU+0fghcjZtiKtfZwE6MxCTmR3BM9Bf ESdA== X-Gm-Message-State: AC+VfDwHdUtB+5ZhmZihMmPwAC491YQVj15/L8kqQzNxOLeRucFl87vS 3j4G3u+n+N6nxJJ870h7vvDIIaLhaNXk3qL125VZ2Iu2MiMy5Pmyz5LU7s/jWuGnqRxxpCy0g7b oMDHIEeYNO52obVrjjeXttau51tcykC3coC50MghoGIQVYxsLcwgeyHyEcMyA7O0L8TRn25XNno hw0xf9VA== X-Received: by 2002:a5d:6309:0:b0:304:7eaa:b196 with SMTP id i9-20020a5d6309000000b003047eaab196mr25768851wru.24.1684178883072; Mon, 15 May 2023 12:28:03 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ75Sc/kKThAZipK6IRek0sGMykjWySWAbcu+iUpdyKOej4+JVaKef8ew5S2iThJZMu/E6/mTg== X-Received: by 2002:a5d:6309:0:b0:304:7eaa:b196 with SMTP id i9-20020a5d6309000000b003047eaab196mr25768811wru.24.1684178881148; Mon, 15 May 2023 12:28:01 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id b4-20020adfde04000000b0030632833e74sm119584wrm.11.2023.05.15.12.28.00 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:28:00 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess , Eli Zaretskii Subject: [PATCHv2 6/9] gdb: parse pending breakpoint thread/task immediately Date: Mon, 15 May 2023 20:27:42 +0100 Message-Id: <7c5086712bbda79773e6ff738b18a5055df82b11.1684178710.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=-11.7 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, RCVD_IN_MSPIKE_H2, 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" Eli approved the docs changes here: https://sourceware.org/pipermail/gdb-patches/2023-April/199221.html --- The initial motivation for this commit was to allow thread or (in the future) inferior specific breakpoints to only be inserted within the appropriate inferior's program-space. The benefit of this is that inferiors for which the breakpoint does not apply will no longer need to stop, and then resume, for such breakpoints. This commit does not make this change, but is a refactor to allow this change in a later commit. The problem we currently have is that when a thread-specific breakpoint is created, the thread number is only parsed by calling find_condition_and_thread_for_sals. This function is only called for non-pending breakpoints, and requires that we know the locations at which the breakpoint will be placed (for expression checking in case the breakpoint is also conditional). A consequence of this is that by the time we figure out the breakpoint is thread-specific we have already looked up locations in all program spaces. This feels wasteful -- if we knew the thread-id earlier then we could reduce the work GDB does by only looking up locations within the program space for which the breakpoint applies. Another consequence of how find_condition_and_thread_for_sals is called is that pending breakpoints don't currently know they are thread-specific, nor even that they are conditional! Additionally, by delaying parsing the thread-id, pending breakpoints can be created for non-existent threads, this is different to how non-pending breakpoints are handled. In this commit the create_breakpoint function is updated so that the extra_string, which contains the thread/task/condition information, is parsed immediately, even for pending breakpoints. Obviously, the condition still can't be validated until the breakpoint becomes non-pending, but the thread/task information can be pulled from the extra-string, and can be validated early on, even for pending breakpoints. There are a couple of benefits to doing this: 1. Printing of breakpoints is more consistent now. Consider creating a conditional breakpoint before this commit: (gdb) set breakpoint pending on (gdb) break pendingfunc if (0) Function "pendingfunc" not defined. Breakpoint 1 (pendingfunc if (0)) pending. (gdb) break main if (0) Breakpoint 2 at 0x401198: file /tmp/hello.c, line 18. (gdb) info breakpoints Num Type Disp Enb Address What 1 breakpoint keep y pendingfunc if (0) 2 breakpoint keep y 0x0000000000401198 in main at /tmp/hello.c:18 stop only if (0) (gdb) And after this commit: (gdb) set breakpoint pending on (gdb) break pendingfunc if (0) Function "pendingfunc" not defined. Breakpoint 1 (pendingfunc) pending. (gdb) break main if (0) Breakpoint 2 at 0x401198: file /home/andrew/tmp/hello.c, line 18. (gdb) info breakpoints Num Type Disp Enb Address What 1 breakpoint keep y pendingfunc stop only if (0) 2 breakpoint keep y 0x0000000000401198 in main at /home/andrew/tmp/hello.c:18 stop only if (0) (gdb) Notice that the display of the condition is now the same for the pending and non-pending breakpoints. The same is true for the thread information in thread-specific breakpoints, this information is displayed on its own line now rather than being part of the 'What' field. 2. We can check that the thread exists as soon as the pending breakpoint exits. Currently there is a weird different between pending and non-pending breakpoints when creating a thread-specific breakpoint. A pending thread-specific breakpoint only checks its thread when it becomes non-pending, at which point the thread the breakpoint was intended for might have exited. Here's the behaviour before this commit: (gdb) set breakpoint pending on (gdb) break foo thread 2 Function "foo" not defined. Breakpoint 2 (foo thread 2) pending. (gdb) c Continuing. [Thread 0x7ffff7c56700 (LWP 2948835) exited] Error in re-setting breakpoint 2: Unknown thread 2. [Inferior 1 (process 2948832) exited normally] (gdb) Notice the 'Error in re-setting breakpoint 2: Unknown thread 2.' line, this was triggered when the thread tried to become non-pending, and GDB discovers that the thread no longer exists. Compare that to the behaviour after this commit: (gdb) set breakpoint pending on (gdb) break foo thread 2 Function "foo" not defined. Breakpoint 2 (foo) pending. (gdb) c Continuing. [Thread 0x7ffff7c56700 (LWP 2949243) exited] Thread-specific breakpoint 2 deleted - thread 2 no longer in the thread list. [Inferior 1 (process 2949240) exited normally] (gdb) Now the behaviour for pending breakpoints is identical to non-pending breakpoints, the thread specific breakpoint is removed as soon as the thread the breakpoint is associated with exits. As a result of this commit the breakpoints 'extra_string' field is now only used by bp_dprintf type breakpoints to hold the printf format and arguments. This string should always be empty for other breakpoint types. This allows me to clean up some code in print_breakpoint_location. In code_breakpoint::code_breakpoint I'm able to change an error case into an assert. This is because this error is now handled earlier in create_breakpoint. As a result we know that by this point, the extra_string will always be nullptr for anything other than a bp_dprintf style breakpoint. The find_condition_and_thread_for_sals function is now no longer needed, this was previously doing the delayed splitting of the extra string into thread, task, and condition, but this is now all done in create_breakpoint, so find_condition_and_thread_for_sals can be deleted, and the code that calls this in code_breakpoint::location_spec_to_sals can be removed. With this update this code would only ever be reached for bp_dprintf style breakpoints, and in these cases the extra_string should not contain anything other than format and args. The most interesting changes are all in create_breakpoint and in the new file break-cond-parse.c. We have a new block of code early on in create_breakpoint that is responsible for splitting the extra_string into its component parts by calling create_breakpoint_parse_arg_string a function in the new break-cond-parse.c file. This means that some of the later code can be simplified a little. The new break-cond-parse.c file implements the splitting up the extra_string and finding all the parts, as well as some self-tests for the new function. Finally, now we know all the details, these can be stored within the breakpoint object if we end up creating a deferred breakpoint. Additionally, if we are creating a deferred bp_dprintf we can parse the extra_string to build the printf command. Reviewed-By: Eli Zaretskii --- gdb/Makefile.in | 2 + gdb/NEWS | 4 + gdb/break-cond-parse.c | 403 ++++++++++++++++++ gdb/break-cond-parse.h | 47 ++ gdb/breakpoint.c | 330 ++++---------- gdb/testsuite/gdb.base/condbreak.exp | 20 +- gdb/testsuite/gdb.base/pending.exp | 23 +- gdb/testsuite/gdb.linespec/explicit.exp | 4 +- gdb/testsuite/gdb.mi/mi-dprintf-pending.exp | 3 +- .../gdb.threads/del-pending-thread-bp-lib.c | 22 + .../gdb.threads/del-pending-thread-bp.c | 85 ++++ .../gdb.threads/del-pending-thread-bp.exp | 108 +++++ 12 files changed, 793 insertions(+), 258 deletions(-) create mode 100644 gdb/break-cond-parse.c create mode 100644 gdb/break-cond-parse.h create mode 100644 gdb/testsuite/gdb.threads/del-pending-thread-bp-lib.c create mode 100644 gdb/testsuite/gdb.threads/del-pending-thread-bp.c create mode 100644 gdb/testsuite/gdb.threads/del-pending-thread-bp.exp diff --git a/gdb/Makefile.in b/gdb/Makefile.in index 14b5dd0bad6..ad7014c8c48 100644 --- a/gdb/Makefile.in +++ b/gdb/Makefile.in @@ -1024,6 +1024,7 @@ COMMON_SFILES = \ break-catch-sig.c \ break-catch-syscall.c \ break-catch-throw.c \ + break-cond-parse.c \ breakpoint.c \ bt-utils.c \ btrace.c \ @@ -1291,6 +1292,7 @@ HFILES_NO_SRCDIR = \ bfd-target.h \ bfin-tdep.h \ block.h \ + break-cond-parse.h \ breakpoint.h \ bsd-kvm.h \ bsd-uthread.h \ diff --git a/gdb/NEWS b/gdb/NEWS index 6aa0d5171f2..376c4c223be 100644 --- a/gdb/NEWS +++ b/gdb/NEWS @@ -67,6 +67,10 @@ break foo thread 1 task 1 watch var thread 2 task 3 +* For breakpoints that are created in the 'pending' state, any + 'thread' or 'task' keywords are parsed at the time the breakpoint is + created, rather than at the time the breakpoint becomes non-pending. + * New commands maintenance print record-instruction [ N ] diff --git a/gdb/break-cond-parse.c b/gdb/break-cond-parse.c new file mode 100644 index 00000000000..e8da24d2643 --- /dev/null +++ b/gdb/break-cond-parse.c @@ -0,0 +1,403 @@ +/* Copyright (C) 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 "defs.h" +#include "gdbsupport/gdb_assert.h" +#include "gdbsupport/selftest.h" +#include "test-target.h" +#include "scoped-mock-context.h" +#include "break-cond-parse.h" +#include "tid-parse.h" +#include "ada-lang.h" + +/* When parsing tokens from a string, which direction are we parsing? + + Given the following string and pointer 'ptr': + + ABC DEF GHI JKL + ^ + ptr + + Parsing 'forward' will return the token 'GHI' and update 'ptr' to point + between GHI and JKL. Parsing 'backward' will return the token 'DEF' and + update 'ptr' to point between ABC and DEF. +*/ + +enum class parse_direction +{ + /* Parse the next token forwards. */ + forward, + + /* Parse the previous token backwards. */ + backward +}; + +struct token +{ + /* Create a new token. START points to the first character of the new + token, while END points at the last character of the new token. + + Neither START or END can be nullptr, and both START and END must point + into the same C style string (i.e. there should be no null character + between START and END). END must be equal to, or greater than START, + that is, it is not possible to create a zero length token. */ + + token (const char *start, const char *end) + : m_start (start), + m_end (end) + { + gdb_assert (m_end >= m_start); + gdb_assert (m_start + strlen (m_start) > m_end); + } + + /* Return a pointer to the first character of this token. */ + const char *start () const + { + return m_start; + } + + /* Return a pointer to the last character of this token. */ + const char *end () const + { + return m_end; + } + + /* Return the length of the token. */ + size_t length () const + { + return m_end - m_start + 1; + } + + /* Return true if this token matches STR. If this token is shorter than + STR then only a partial match is performed and true will be returned + if the token length sub-string matches. Otherwise, return false. */ + bool matches (const char *str) const + { + return strncmp (m_start, str, length ()) == 0; + } + +private: + /* The first character of this token. */ + const char *m_start; + + /* The last character of this token. */ + const char *m_end; +}; + +/* Find the next token in DIRECTION from *CURR. */ + +static token +find_next_token (const char **curr, parse_direction direction) +{ + const char *tok_start, *tok_end; + + if (direction == parse_direction::forward) + { + *curr = skip_spaces (*curr); + tok_start = *curr; + *curr = skip_to_space (*curr); + tok_end = *curr - 1; + } + else + { + gdb_assert (direction == parse_direction::backward); + + while (isspace (**curr)) + --(*curr); + + tok_end = *curr; + + while (!isspace (**curr)) + --(*curr); + + tok_start = (*curr) + 1; + } + + return token (tok_start, tok_end); +} + +/* See break-cond-parse.h. */ + +void +create_breakpoint_parse_arg_string + (const char *tok, gdb::unique_xmalloc_ptr *cond_string, + int *thread, int *task, gdb::unique_xmalloc_ptr *rest, bool *force) +{ + /* Set up the defaults. */ + cond_string->reset (); + rest->reset (); + *thread = -1; + *task = -1; + *force = false; + + if (tok == nullptr) + return; + + const char *cond_start = nullptr; + const char *cond_end = nullptr; + parse_direction direction = parse_direction::forward; + while (*tok != '\0') + { + /* Find the next token. If moving backward and this token starts at + the same location as the condition then we must have found the + other end of the condition string -- we're done. */ + token t = find_next_token (&tok, direction); + if (direction == parse_direction::backward && t.start () <= cond_start) + { + cond_end = t.end () + 1; + break; + } + + /* We only have a single flag option to check for. All the other + options take a value so require an additional token to be found. */ + if (t.length () > 0 && t.matches ("-force-condition")) + { + *force = true; + continue; + } + + /* Maybe the first token was the last token in the string. If this + is the case then we definitely can't try to extract a value + token. This also means that the token T is meaningless. Reset + TOK to point at the start of the unknown content and break out of + the loop. We'll record the unknown part of the string outside of + the scanning loop (below). */ + if (direction == parse_direction::forward && *tok == '\0') + { + tok = t.start (); + break; + } + + /* As before, find the next token and, if we are scanning backwards, + check that we have not reached the start of the condition string. */ + token v = find_next_token (&tok, direction); + if (direction == parse_direction::backward && v.start () <= cond_start) + { + /* Use token T here as that must also be part of the condition + string. */ + cond_end = t.end () + 1; + break; + } + + /* When moving backward we will first parse the value token then the + keyword token, so swap them now. */ + if (direction == parse_direction::backward) + std::swap (t, v); + + /* Check for valid option in token T. If we find a valid option then + parse the value from the token V. Except for 'if', that's handled + differently. + + For the 'if' token we need to capture the entire condition + string, so record the start of the condition string and then + start scanning backwards looking for the end of the condition + string. */ + if (direction == parse_direction::forward && t.length () > 0 + && t.matches ("if")) + { + cond_start = v.start (); + tok = tok + strlen (tok); + gdb_assert (*tok == '\0'); + --tok; + direction = parse_direction::backward; + continue; + } + else if (t.length () > 0 && t.matches ("thread")) + { + const char *tmptok; + struct thread_info *thr; + + if (*thread != -1) + error(_("You can specify only one thread.")); + + if (*task != -1) + error (_("You can specify only one of thread or task.")); + + thr = parse_thread_id (v.start (), &tmptok); + const char *expected_end = skip_spaces (v.end () + 1); + if (tmptok != expected_end) + error (_("Junk after thread keyword.")); + *thread = thr->global_num; + } + else if (t.length () > 0 && t.matches ("task")) + { + char *tmptok; + + if (*task != -1) + error(_("You can specify only one task.")); + + if (*thread != -1) + error (_("You can specify only one of thread or task.")); + + *task = strtol (v.start (), &tmptok, 0); + const char *expected_end = v.end () + 1; + if (tmptok != expected_end) + error (_("Junk after task keyword.")); + if (!valid_task_id (*task)) + error (_("Unknown task %d."), *task); + } + else + { + /* An unknown token. If we are scanning forward then reset TOK + to point at the start of the unknown content, we record this + outside of the scanning loop (below). + + If we are scanning backward then unknown content is assumed to + be the other end of the condition string, obviously, this is + just a heuristic, we could be looking at a mistyped command + line, but this will be spotted when the condition is + eventually evaluated. + + Either way, no more scanning is required after this. */ + if (direction == parse_direction::forward) + tok = t.start (); + else + { + gdb_assert (direction == parse_direction::backward); + cond_end = v.end () + 1; + } + break; + } + } + + if (cond_start != nullptr) + { + /* If we found the start of a condition string then we should have + switched to backward scan mode, and found the end of the condition + string. Capture the whole condition string into COND_STRING now. */ + gdb_assert (direction == parse_direction::backward); + gdb_assert (cond_end != nullptr); + cond_string->reset (savestring (cond_start, cond_end - cond_start)); + } + else if (*tok != '\0') + { + /* If we didn't have a condition start pointer then we should still + be in forward scanning mode. If we didn't reach the end of the + input string (TOK is not at the null character) then the rest of + the input string is garbage that we didn't understand. + + Record the unknown content into REST. The caller of this function + will report this as an error later on. We could report the error + here, but we prefer to allow the caller to run other checks, and + prioritise other errors before reporting this problem. */ + gdb_assert (direction == parse_direction::forward); + rest->reset (savestring (tok, strlen (tok))); + } +} + +#if GDB_SELF_TEST + +namespace selftests { + +/* Run a single test of the create_breakpoint_parse_arg_string function. + INPUT is passed to create_breakpoint_parse_arg_string while all other + arguments are the expected output from + create_breakpoint_parse_arg_string. */ + +static void +test (const char *input, const char *condition, int thread = -1, + int task = -1, bool force = false, const char *rest = nullptr) +{ + gdb::unique_xmalloc_ptr extracted_condition; + gdb::unique_xmalloc_ptr extracted_rest; + int extracted_thread, extracted_task; + bool extracted_force_condition; + std::string exception_msg; + + try + { + create_breakpoint_parse_arg_string (input, &extracted_condition, + &extracted_thread, + &extracted_task, &extracted_rest, + &extracted_force_condition); + } + catch (const gdb_exception_error &ex) + { + string_file buf; + + exception_print (&buf, ex); + exception_msg = buf.release (); + } + + if ((condition == nullptr) != (extracted_condition.get () == nullptr) + || (condition != nullptr + && strcmp (condition, extracted_condition.get ()) != 0) + || (rest == nullptr) != (extracted_rest.get () == nullptr) + || (rest != nullptr && strcmp (rest, extracted_rest.get ()) != 0) + || thread != extracted_thread + || task != extracted_task + || force != extracted_force_condition + || !exception_msg.empty ()) + { + if (run_verbose ()) + { + debug_printf ("input: '%s'\n", input); + debug_printf ("condition: '%s'\n", extracted_condition.get ()); + debug_printf ("rest: '%s'\n", extracted_rest.get ()); + debug_printf ("thread: %d\n", extracted_thread); + debug_printf ("task: %d\n", extracted_task); + debug_printf ("forced: %s\n", + extracted_force_condition ? "true" : "false"); + debug_printf ("exception: '%s'\n", exception_msg.c_str ()); + } + + /* Report the failure. */ + SELF_CHECK (false); + } +} + +/* Test the create_breakpoint_parse_arg_string function. Just wraps + multiple calls to the test function above. */ + +static void +create_breakpoint_parse_arg_string_tests (void) +{ + gdbarch *arch = current_inferior ()->gdbarch; + scoped_restore_current_pspace_and_thread restore; + scoped_mock_context mock_target (arch); + + int global_thread_num = mock_target.mock_thread.global_num; + + test (" if blah ", "blah"); + test (" if blah thread 1", "blah", global_thread_num); + test (" if blah thread 1 ", "blah", global_thread_num); + test ("thread 1 woof", nullptr, global_thread_num, -1, false, "woof"); + test ("thread 1 X", nullptr, global_thread_num, -1, false, "X"); + test (" if blah thread 1 -force-condition", "blah", global_thread_num, + -1, true); + test (" -force-condition if blah thread 1", "blah", global_thread_num, + -1, true); + test (" -force-condition if blah thread 1 ", "blah", global_thread_num, + -1, true); + test ("thread 1 -force-condition if blah", "blah", global_thread_num, + -1, true); + test ("if (A::outer::func ())", "(A::outer::func ())"); +} + +} // namespace selftests +#endif /* GDB_SELF_TEST */ + +void _initialize_break_cond_parse (); +void +_initialize_break_cond_parse () +{ +#if GDB_SELF_TEST + selftests::register_test + ("create_breakpoint_parse_arg_string", + selftests::create_breakpoint_parse_arg_string_tests); +#endif +} diff --git a/gdb/break-cond-parse.h b/gdb/break-cond-parse.h new file mode 100644 index 00000000000..ec62b88ea92 --- /dev/null +++ b/gdb/break-cond-parse.h @@ -0,0 +1,47 @@ +/* Copyright (C) 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 . */ + +#if !defined (BREAK_COND_PARSE_H) +#define BREAK_COND_PARSE_H 1 + +/* Given TOK, a string possibly containing a condition, thread, task and + force-condition flag, as accepted by the 'break' command, extract the + condition string, thread and task numbers, and the force_condition flag, + then set *COND_STRING, *THREAD, *TASK, and *FORCE. + + As TOK is parsed, if an unknown keyword is encountered before the 'if' + keyword then everything starting from the unknown keyword is placed into + *REST. + + Both *COND and *REST are initialized to nullptr. If no 'if' keyword is + found then *COND will be returned as nullptr. If no unknown content is + found then *REST is returned as nullptr. + + If no thread is found, *THREAD is set to -1. If no task is found, *TASK + is set to -1. If the -force-condition flag is not found then *FORCE is + set to false. + + Due to the free-form nature that the string TOK might take (a 'thread' + keyword can appear before or after an 'if' condition) then we end up + having to check for keywords from both the start of TOK and the end of + TOK. */ + +extern void create_breakpoint_parse_arg_string + (const char *tok, gdb::unique_xmalloc_ptr *cond_string, + int *thread, int *task, gdb::unique_xmalloc_ptr *rest, bool *force); + +#endif diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index f9020c49d90..c4accfb6d28 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -70,6 +70,7 @@ #include "cli/cli-style.h" #include "cli/cli-decode.h" #include +#include "break-cond-parse.h" /* readline include files */ #include "readline/tilde.h" @@ -6206,20 +6207,7 @@ print_breakpoint_location (const breakpoint *b, uiout->field_stream ("at", stb); } else - { - uiout->field_string ("pending", b->locspec->to_string ()); - /* If extra_string is available, it could be holding a condition - or dprintf arguments. In either case, make sure it is printed, - too, but only for non-MI streams. */ - if (!uiout->is_mi_like_p () && b->extra_string != NULL) - { - if (b->type == bp_dprintf) - uiout->text (","); - else - uiout->text (" "); - uiout->text (b->extra_string.get ()); - } - } + uiout->field_string ("pending", b->locspec->to_string ()); if (loc && is_breakpoint (b) && breakpoint_condition_evaluation_mode () == condition_evaluation_target @@ -8553,8 +8541,8 @@ code_breakpoint::code_breakpoint (struct gdbarch *gdbarch_, gdb_assert (extra_string != nullptr); update_dprintf_command_list (this); } - else if (extra_string != nullptr) - error (_("Garbage '%s' at end of command"), extra_string.get ()); + else + gdb_assert (extra_string == nullptr); /* The order of the locations is now stable. Set the location condition using the location's number. */ @@ -8781,160 +8769,6 @@ check_fast_tracepoint_sals (struct gdbarch *gdbarch, } } -/* Given TOK, a string specification of condition and thread, as - accepted by the 'break' command, extract the condition - string and thread number and set *COND_STRING and *THREAD. - PC identifies the context at which the condition should be parsed. - If no condition is found, *COND_STRING is set to NULL. - If no thread is found, *THREAD is set to -1. */ - -static void -find_condition_and_thread (const char *tok, CORE_ADDR pc, - gdb::unique_xmalloc_ptr *cond_string, - int *thread, int *task, - gdb::unique_xmalloc_ptr *rest) -{ - cond_string->reset (); - *thread = -1; - *task = -1; - rest->reset (); - bool force = false; - - while (tok && *tok) - { - const char *end_tok; - int toklen; - const char *cond_start = NULL; - const char *cond_end = NULL; - - tok = skip_spaces (tok); - - if ((*tok == '"' || *tok == ',') && rest) - { - rest->reset (savestring (tok, strlen (tok))); - return; - } - - end_tok = skip_to_space (tok); - - toklen = end_tok - tok; - - if (toklen >= 1 && strncmp (tok, "if", toklen) == 0) - { - tok = cond_start = end_tok + 1; - try - { - parse_exp_1 (&tok, pc, block_for_pc (pc), 0); - } - catch (const gdb_exception_error &) - { - if (!force) - throw; - else - tok = tok + strlen (tok); - } - cond_end = tok; - cond_string->reset (savestring (cond_start, cond_end - cond_start)); - } - else if (toklen >= 1 && strncmp (tok, "-force-condition", toklen) == 0) - { - tok = tok + toklen; - force = true; - } - else if (toklen >= 1 && strncmp (tok, "thread", toklen) == 0) - { - const char *tmptok; - struct thread_info *thr; - - if (*thread != -1) - error(_("You can specify only one thread.")); - - if (*task != -1) - error (_("You can specify only one of thread or task.")); - - tok = end_tok + 1; - thr = parse_thread_id (tok, &tmptok); - if (tok == tmptok) - error (_("Junk after thread keyword.")); - *thread = thr->global_num; - tok = tmptok; - } - else if (toklen >= 1 && strncmp (tok, "task", toklen) == 0) - { - char *tmptok; - - if (*task != -1) - error(_("You can specify only one task.")); - - if (*thread != -1) - error (_("You can specify only one of thread or task.")); - - tok = end_tok + 1; - *task = strtol (tok, &tmptok, 0); - if (tok == tmptok) - error (_("Junk after task keyword.")); - if (!valid_task_id (*task)) - error (_("Unknown task %d."), *task); - tok = tmptok; - } - else if (rest) - { - rest->reset (savestring (tok, strlen (tok))); - return; - } - else - error (_("Junk at end of arguments.")); - } -} - -/* Call 'find_condition_and_thread' for each sal in SALS until a parse - succeeds. The parsed values are written to COND_STRING, THREAD, - TASK, and REST. See the comment of 'find_condition_and_thread' - for the description of these parameters and INPUT. */ - -static void -find_condition_and_thread_for_sals (const std::vector &sals, - const char *input, - gdb::unique_xmalloc_ptr *cond_string, - int *thread, int *task, - gdb::unique_xmalloc_ptr *rest) -{ - int num_failures = 0; - for (auto &sal : sals) - { - gdb::unique_xmalloc_ptr cond; - int thread_id = -1; - int task_id = -1; - gdb::unique_xmalloc_ptr remaining; - - /* Here we want to parse 'arg' to separate condition from thread - number. But because parsing happens in a context and the - contexts of sals might be different, try each until there is - success. Finding one successful parse is sufficient for our - goal. When setting the breakpoint we'll re-parse the - condition in the context of each sal. */ - try - { - find_condition_and_thread (input, sal.pc, &cond, &thread_id, - &task_id, &remaining); - *cond_string = std::move (cond); - /* At most one of thread or task can be set. */ - gdb_assert (thread_id == -1 || task_id == -1); - *thread = thread_id; - *task = task_id; - *rest = std::move (remaining); - break; - } - catch (const gdb_exception_error &e) - { - num_failures++; - /* If no sal remains, do not continue. */ - if (num_failures == sals.size ()) - throw; - } - } -} - /* Decode a static tracepoint marker spec. */ static std::vector @@ -9045,6 +8879,44 @@ create_breakpoint (struct gdbarch *gdbarch, || (type_wanted != bp_dprintf && (extra_string == nullptr || parse_extra))); + /* Will hold either copies of the similarly named function argument, or + will hold a modified version of the function argument, depending on + the value of PARSE_EXTRA. */ + gdb::unique_xmalloc_ptr cond_string_copy; + gdb::unique_xmalloc_ptr extra_string_copy; + + if (parse_extra) + { + /* Parse EXTRA_STRING splitting the parts out. */ + create_breakpoint_parse_arg_string (extra_string, &cond_string_copy, + &thread, &task, &extra_string_copy, + &force_condition); + + /* We could check that EXTRA_STRING_COPY is empty at this point -- it + should be, as we only get here for things that are not bp_dprintf, + however, we prefer to give the location spec parser a chance to + run first, this means the user will get errors about invalid + location spec instead of an error about garbage at the end of the + command line. + + We still do the EXTRA_STRING_COPY is empty check, just later in + this function. */ + + gdb_assert (thread == -1 || thread > 0); + gdb_assert (task == -1 || task > 0); + } + else + { + if (cond_string != nullptr) + cond_string_copy.reset (xstrdup (cond_string)); + if (extra_string != nullptr) + extra_string_copy.reset (xstrdup (extra_string)); + } + + /* Clear these. Updated values are now held in the *_copy locals. */ + cond_string = nullptr; + extra_string = nullptr; + try { ops->create_sals_from_location_spec (locspec, &canonical); @@ -9080,6 +8952,13 @@ create_breakpoint (struct gdbarch *gdbarch, throw; } + /* The only bp_dprintf should have anything in EXTRA_STRING_COPY by this + point. For all other breakpoints this indicates an error. We could + place this check earlier in the function, but we prefer to see errors + from the location spec parser before we see this error message. */ + if (type_wanted != bp_dprintf && extra_string_copy.get () != nullptr) + error (_("Garbage '%s' at end of command"), extra_string_copy.get ()); + if (!pending && canonical.lsals.empty ()) return 0; @@ -9103,62 +8982,31 @@ create_breakpoint (struct gdbarch *gdbarch, breakpoint. */ if (!pending) { - gdb::unique_xmalloc_ptr cond_string_copy; - gdb::unique_xmalloc_ptr extra_string_copy; - - if (parse_extra) + /* Check the validity of the condition. We should error out if the + condition is invalid at all of the locations and if it is not + forced. In the PARSE_EXTRA case above, this check is done when + parsing the EXTRA_STRING. */ + if (cond_string_copy.get () != nullptr && !force_condition) { - gdb_assert (type_wanted != bp_dprintf); - - gdb::unique_xmalloc_ptr rest; - gdb::unique_xmalloc_ptr cond; - + int num_failures = 0; const linespec_sals &lsal = canonical.lsals[0]; - - find_condition_and_thread_for_sals (lsal.sals, extra_string, - &cond, &thread, &task, &rest); - - if (rest.get () != nullptr && *(rest.get ()) != '\0') - error (_("Garbage '%s' at end of command"), rest.get ()); - - cond_string_copy = std::move (cond); - extra_string_copy = std::move (rest); - } - else - { - /* Check the validity of the condition. We should error out - if the condition is invalid at all of the locations and - if it is not forced. In the PARSE_EXTRA case above, this - check is done when parsing the EXTRA_STRING. */ - if (cond_string != nullptr && !force_condition) + for (const auto &sal : lsal.sals) { - int num_failures = 0; - const linespec_sals &lsal = canonical.lsals[0]; - for (const auto &sal : lsal.sals) + const char *cond = cond_string_copy.get (); + try { - const char *cond = cond_string; - try - { - parse_exp_1 (&cond, sal.pc, block_for_pc (sal.pc), 0); - /* One success is sufficient to keep going. */ - break; - } - catch (const gdb_exception_error &) - { - num_failures++; - /* If this is the last sal, error out. */ - if (num_failures == lsal.sals.size ()) - throw; - } + parse_exp_1 (&cond, sal.pc, block_for_pc (sal.pc), 0); + /* One success is sufficient to keep going. */ + break; + } + catch (const gdb_exception_error &) + { + num_failures++; + /* If this is the last sal, error out. */ + if (num_failures == lsal.sals.size ()) + throw; } } - - /* Create a private copy of condition string. */ - if (cond_string) - cond_string_copy.reset (xstrdup (cond_string)); - /* Create a private copy of any extra string. */ - if (extra_string) - extra_string_copy.reset (xstrdup (extra_string)); } ops->create_breakpoints_sal (gdbarch, &canonical, @@ -9175,21 +9023,15 @@ create_breakpoint (struct gdbarch *gdbarch, type_wanted); b->locspec = locspec->clone (); - if (parse_extra) - b->cond_string = NULL; - else - { - /* Create a private copy of condition string. */ - b->cond_string.reset (cond_string != NULL - ? xstrdup (cond_string) - : NULL); - b->thread = thread; - } + /* Create a private copy of condition string. */ + b->cond_string = std::move (cond_string_copy); + + b->thread = thread; + b->task = task; /* Create a private copy of any extra string. */ - b->extra_string.reset (extra_string != NULL - ? xstrdup (extra_string) - : NULL); + b->extra_string = std::move (extra_string_copy); + b->ignore_count = ignore_count; b->disposition = tempflag ? disp_del : disp_donttouch; b->condition_not_parsed = 1; @@ -9198,9 +9040,12 @@ create_breakpoint (struct gdbarch *gdbarch, && type_wanted != bp_hardware_breakpoint) || thread != -1) b->pspace = current_program_space; + if (type_wanted == bp_dprintf) + update_dprintf_command_list (b.get ()); + install_breakpoint (internal, std::move (b), 0); } - + if (canonical.lsals.size () > 1) { warning (_("Multiple breakpoints were set.\nUse the " @@ -12916,23 +12761,6 @@ code_breakpoint::location_spec_to_sals (location_spec *locspec, { for (auto &sal : sals) resolve_sal_pc (&sal); - if (condition_not_parsed && extra_string != NULL) - { - gdb::unique_xmalloc_ptr local_cond, local_extra; - int local_thread, local_task; - - find_condition_and_thread_for_sals (sals, extra_string.get (), - &local_cond, &local_thread, - &local_task, &local_extra); - gdb_assert (cond_string == nullptr); - if (local_cond != nullptr) - cond_string = std::move (local_cond); - thread = local_thread; - task = local_task; - if (local_extra != nullptr) - extra_string = std::move (local_extra); - condition_not_parsed = 0; - } if (type == bp_static_tracepoint) sals[0] = update_static_tracepoint (this, sals[0]); diff --git a/gdb/testsuite/gdb.base/condbreak.exp b/gdb/testsuite/gdb.base/condbreak.exp index a5b2a28701b..50ae82c3f32 100644 --- a/gdb/testsuite/gdb.base/condbreak.exp +++ b/gdb/testsuite/gdb.base/condbreak.exp @@ -179,6 +179,10 @@ gdb_test "break -q main if (1==1) thread 999" \ "Unknown thread 999\\." gdb_test "break -q main thread 999 if (1==1)" \ "Unknown thread 999\\." +gdb_test "break -q main if (1==1) thread 999 -force-condition" \ + "Unknown thread 999\\." +gdb_test "break -q main thread 999 if (1==1) -force-condition" \ + "Unknown thread 999\\." # Verify that both if and thread can be distinguished from a breakpoint # address expression. @@ -186,20 +190,34 @@ gdb_test "break *main if (1==1) thread 999" \ "Unknown thread 999\\." gdb_test "break *main thread 999 if (1==1)" \ "Unknown thread 999\\." +gdb_test "break *main if (1==1) thread 999 -force-condition" \ + "Unknown thread 999\\." +gdb_test "break *main thread 999 if (1==1) -force-condition" \ + "Unknown thread 999\\." # Similarly for task. gdb_test "break *main if (1==1) task 999" \ "Unknown task 999\\." gdb_test "break *main task 999 if (1==1)" \ "Unknown task 999\\." +gdb_test "break *main if (1==1) task 999 -force-condition" \ + "Unknown task 999\\." +gdb_test "break *main task 999 if (1==1) -force-condition" \ + "Unknown task 999\\." -# GDB accepts abbreviations for "thread" and "task". +# GDB accepts abbreviations for "thread" and "task" and "-force-condition" gdb_test "break *main if (1==1) t 999" \ "Unknown thread 999\\." gdb_test "break *main if (1==1) th 999" \ "Unknown thread 999\\." gdb_test "break *main if (1==1) ta 999" \ "Unknown task 999\\." +gdb_test "break *main if (1==1) t 999 -force" \ + "Unknown thread 999\\." +gdb_test "break *main if (1==1) th 999 -force" \ + "Unknown thread 999\\." +gdb_test "break *main if (1==1) ta 999 -force" \ + "Unknown task 999\\." set test "run until breakpoint at marker3" gdb_test_multiple "continue" $test { diff --git a/gdb/testsuite/gdb.base/pending.exp b/gdb/testsuite/gdb.base/pending.exp index d7d3735000a..645f63bdb16 100644 --- a/gdb/testsuite/gdb.base/pending.exp +++ b/gdb/testsuite/gdb.base/pending.exp @@ -170,7 +170,8 @@ gdb_test "info break" \ \[\t \]+stop only if k == 1.* \[\t \]+print k.* \[0-9\]+\[\t \]+breakpoint keep y.* in main at .*$srcfile:$mainline.* -\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp2_loc if x > 3.*" \ +\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp2_loc.* +\\s+stop only if x > 3.*" \ "multiple pending breakpoints" @@ -195,8 +196,10 @@ gdb_test "info break" \ \[\t \]+stop only if k == 1.* \[\t \]+print k.* \[0-9\]+\[\t \]+breakpoint keep y.* in main at .*$srcfile:$mainline.* -\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp2_loc if x > 3.* -\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp3_loc.*ignore next 2 hits.*" \ +\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp2_loc.* +\\s+stop only if x > 3.* +\[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*pendshr.c:$bp3_loc.* +\\s+ignore next 2 hits.*" \ "multiple pending breakpoints 2" # @@ -267,3 +270,17 @@ gdb_test "info break" \ \[0-9\]+\[\t \]+breakpoint keep y.* in main at .*$srcfile:$mainline.* \[0-9\]+\[\t \]+breakpoint keep y.*PENDING.*imaginary.*" \ "verify pending breakpoint after restart" + +# Test GDB's parsing of pending breakpoint thread and condition. + +gdb_test_no_output "set breakpoint pending on" +gdb_test "break foo if (unknown_var && another_unknown_var) thread 1" \ + "Breakpoint $decimal \\(foo\\) pending\\." +set bpnum [get_integer_valueof "\$bpnum" "*INVALID" \ + "get number for foo breakpoint"] +gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+\\s+foo" \ + "\\s+stop only if \\(unknown_var && another_unknown_var\\)" \ + "\\s+stop only in thread 1"] \ + "check pending breakpoint on foo" diff --git a/gdb/testsuite/gdb.linespec/explicit.exp b/gdb/testsuite/gdb.linespec/explicit.exp index b08d65953d2..7c8b2475881 100644 --- a/gdb/testsuite/gdb.linespec/explicit.exp +++ b/gdb/testsuite/gdb.linespec/explicit.exp @@ -574,7 +574,7 @@ namespace eval $testfile { allow-pending]} { fail "set $tst" } else { - gdb_test "info break" ".*PENDING.*myfunction if foofoofoo == 1.*" $tst + gdb_test "info break" ".*PENDING.*myfunction\r\n\\s+stop only if foofoofoo == 1.*" $tst } gdb_exit @@ -585,7 +585,7 @@ namespace eval $testfile { allow-pending]} { fail "set $tst" } else { - gdb_test "info break" ".*PENDING.*myfunction if arg == 0" $tst + gdb_test "info break" ".*PENDING.*myfunction\r\n\\s+stop only if arg == 0" $tst gdb_load [standard_output_file $exefile] gdb_test "info break" \ diff --git a/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp b/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp index 28f52938aeb..bb9987a5572 100644 --- a/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp +++ b/gdb/testsuite/gdb.mi/mi-dprintf-pending.exp @@ -50,7 +50,8 @@ set bp_location1 [gdb_get_line_number "set breakpoint 1 here"] # Set pending dprintf via MI. set bp [mi_make_breakpoint_pending -number "1" -type "dprintf" \ -disp "keep" -enabled "y" -pending "pendfunc1" \ - -original-location "pendfunc1"] + -original-location "pendfunc1" \ + -script {\["printf \\\"hello\\\""\]}] mi_gdb_test "-dprintf-insert -f pendfunc1 \"hello\"" \ ".*\\^done,$bp" "mi set dprintf" diff --git a/gdb/testsuite/gdb.threads/del-pending-thread-bp-lib.c b/gdb/testsuite/gdb.threads/del-pending-thread-bp-lib.c new file mode 100644 index 00000000000..15d1b9833dd --- /dev/null +++ b/gdb/testsuite/gdb.threads/del-pending-thread-bp-lib.c @@ -0,0 +1,22 @@ +/* Copyright 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 . */ + +int global_var = 0; + +void +foo (int arg) +{ + global_var = arg; +} diff --git a/gdb/testsuite/gdb.threads/del-pending-thread-bp.c b/gdb/testsuite/gdb.threads/del-pending-thread-bp.c new file mode 100644 index 00000000000..938e05bc4d7 --- /dev/null +++ b/gdb/testsuite/gdb.threads/del-pending-thread-bp.c @@ -0,0 +1,85 @@ +/* Copyright 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 + +pthread_barrier_t barrier; + +static void +barrier_wait (pthread_barrier_t *b) +{ + int res = pthread_barrier_wait (b); + if (res != 0 && res != PTHREAD_BARRIER_SERIAL_THREAD) + abort (); +} + +static void * +thread_worker (void *arg) +{ + barrier_wait (&barrier); + return NULL; +} + +void +breakpt () +{ + /* Nothing. */ +} + +int +main (void) +{ + void *handle; + void (*func)(int); + pthread_t thread; + + if (pthread_barrier_init (&barrier, NULL, 2) != 0) + abort (); + + if (pthread_create (&thread, NULL, thread_worker, NULL) != 0) + abort (); + + breakpt (); + + /* Allow the worker thread to complete. */ + barrier_wait (&barrier); + + if (pthread_join (thread, NULL) != 0) + abort (); + + breakpt (); + + /* Now load the shared library. */ + handle = dlopen (SHLIB_NAME, RTLD_LAZY); + if (handle == NULL) + abort (); + + /* Find the function symbol. */ + func = (void (*)(int)) dlsym (handle, "foo"); + + /* Call the library function. */ + func (1); + + /* Unload the shared library. */ + if (dlclose (handle) != 0) + abort (); + + breakpt (); + + return 0; +} + diff --git a/gdb/testsuite/gdb.threads/del-pending-thread-bp.exp b/gdb/testsuite/gdb.threads/del-pending-thread-bp.exp new file mode 100644 index 00000000000..4ed05eaa42d --- /dev/null +++ b/gdb/testsuite/gdb.threads/del-pending-thread-bp.exp @@ -0,0 +1,108 @@ +# Copyright 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 checks that pending thread-specific breakpoints are +# correctly deleted when the thread the breakpoint is for goes out of +# scope. +# +# We also check that we can't create a pending thread-specific +# breakpoint for a non-existent thread. + +require allow_shlib_tests + +standard_testfile + +set libname $testfile-lib +set srcfile_lib $srcdir/$subdir/$libname.c +set binfile_lib [standard_output_file $libname.so] + +if { [gdb_compile_shlib $srcfile_lib $binfile_lib {}] != "" } { + untested "failed to compile shared library 1" + return -1 +} + +set binfile_lib_target [gdb_download_shlib $binfile_lib] + +if { [prepare_for_testing "failed to prepare" $testfile $srcfile \ + [list debug \ + additional_flags=-DSHLIB_NAME=\"$binfile_lib_target\" \ + shlib_load pthreads]] } { + return -1 +} + +gdb_locate_shlib $binfile_lib + +if ![runto_main] { + return 0 +} + +# Run until we have two threads. +gdb_breakpoint "breakpt" +gdb_continue_to_breakpoint "first breakpt call" + +# Confirm that we have a thread '2'. +gdb_test "info threads" "\r\n\\s+2\\s+\[^\r\n\]+" + +# Create a pending, thread-specific, breakpoint on 'foo'. +gdb_breakpoint "foo thread 2" allow-pending +set bpnum [get_integer_valueof "\$bpnum" "*INVALID*" \ + "get breakpoint number"] + +# Check we can't create a pending thread-specific breakpoint for a +# non-existent thread. +gdb_test "with breakpoint pending on -- break foo thread 99" \ + "Unknown thread 99\\." + +# Continue to 'breakpt' again. Don't use gdb_continue_to_breakpoint +# as we are looking for the thread exited and breakpoint deleted +# messages. +set output [list "Continuing\\."] + +if {!([target_info exists gdb_protocol] + && ([target_info gdb_protocol] == "remote" + || [target_info gdb_protocol] == "extended-remote"))} { + # Due to bug PR gdb/30129 we don't see the thread exited messages + # for remote targets. When this bit of this test can be cleaned + # up. + lappend output "\\\[Thread \[^\r\n\]+ exited\\\]" +} + +lappend output "Thread-specific breakpoint $bpnum deleted - thread 2 no longer in the thread list\\." \ + "" \ + "Thread 1 \[^\r\n\]+, breakpt \\(\\) at \[^\r\n\]+" \ + "$decimal\\s+\[^\r\n\]+" + +gdb_test "continue" \ + [multi_line {*}$output] \ + "second breakpt call" + +# Confirm breakpoint has been deleted. +gdb_test "info breakpoints $bpnum" \ + "No breakpoint or watchpoint matching '$bpnum'\\." + +# Continue again. This will pass through 'foo'. We should not stop +# in 'foo', the breakpoint has been deleted. We should next stop in +# breakpt again. +gdb_test "continue" \ + [multi_line \ + "Continuing\\." \ + "" \ + "Thread 1 \[^\r\n\]+ hit Breakpoint $decimal, breakpt \\(\\) at \[^\r\n\]+" \ + "$decimal\\s+\[^\r\n\]+"] \ + "third breakpt call" +gdb_test "bt 1" \ + [multi_line \ + "#0\\s+breakpt \\(\\) at \[^\r\n\]+" \ + "\\(More stack frames follow\\.\\.\\.\\)"] From patchwork Mon May 15 19:27:43 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69400 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 331D73856246 for ; Mon, 15 May 2023 19:29:05 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 331D73856246 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178945; bh=yE1vC2rESeNgkJeL8IZ100B6ToTIjQDT28pnh1sdagA=; 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=pgYpTayDiKEZLcf6wfNj217EtmZJwAqBZVAybv9v2grreM60ul/h0HEhcthm5XooZ fEoj+Dp8K8FOIRbCizluKUeJOzhTAiRiViuCzPle61t4Nx83aYPpYsst470nKq7YwW C+YF28lv8zvGXCdmGbN9Lm6h1SSh4Exi2zlHJGq8= 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 616CC3857701 for ; Mon, 15 May 2023 19:28:06 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 616CC3857701 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-517-zd-POR6ePLWx1y2clFMSYg-1; Mon, 15 May 2023 15:28:04 -0400 X-MC-Unique: zd-POR6ePLWx1y2clFMSYg-1 Received: by mail-wm1-f69.google.com with SMTP id 5b1f17b1804b1-3f433a2308bso582015e9.0 for ; Mon, 15 May 2023 12:28:04 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178883; x=1686770883; 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=yE1vC2rESeNgkJeL8IZ100B6ToTIjQDT28pnh1sdagA=; b=g8sqSnEHuUHxDd8TIyjeYHR5tNoG8hP+arS4EzTEGK9jd9cynY9srIzYsXojblFIDL l8HtoCTPuVKR17T6hUHfctDn+ALPqrWX142JTTWL2Jzk9dr89KxQjleYWdgqls9URS8q VvZ9xFKEkJ+6ZaNv/1xrB2tnr7IN+06aL0fLn9MypslEQN8xWY6ReVv0wcXELgMENBM0 +h2M0nCTV8/d/DpnubOGIrvPsyjKJQNhBpLMJnOJFiZaJvA0InxESZB+40e/sayRfdcR k1AMtOjigMjvKMKyROzBLpQP3UKryOPu/JgrWYJeaoPo5R5m4LCR15TAvRSBydedhcMg KZYA== X-Gm-Message-State: AC+VfDzmD/Nfmnc/EKXznDG89PQXEw7dPMpzIjI8dJwo5G08WZG86sp0 LNe4FJTr+GjsoquEXtlRviWBsYQ34FYb0tag7Xup1cBGf2iSFb+jOUzsM5wLqPkBApRApyIpt81 zE4b2x9wxsZabh8CWtcnagqbJz/n0diQ25jzEfIh8Zqj9IS6tNGW5VC9FJ3tYS3DpOOb/yEVc3w Xyee5Urg== X-Received: by 2002:a05:600c:19c6:b0:3f5:b15:87ee with SMTP id u6-20020a05600c19c600b003f50b1587eemr2841741wmq.18.1684178883245; Mon, 15 May 2023 12:28:03 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ56AXkJ094WND50AdlVMsudgkmndngs0hZXejltFnKcDzsBgdW0QkZYd4Gr3UsWbyNSL8o6Cw== X-Received: by 2002:a05:600c:19c6:b0:3f5:b15:87ee with SMTP id u6-20020a05600c19c600b003f50b1587eemr2841727wmq.18.1684178882917; Mon, 15 May 2023 12:28:02 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id t9-20020a7bc3c9000000b003f31d44f0cbsm306625wmj.29.2023.05.15.12.28.02 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:28:02 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 7/9] gdb: remove breakpoint_re_set_one Date: Mon, 15 May 2023 20:27:43 +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=-11.7 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" During a later patch I wanted to reset a single breakpoint, so I called breakpoint_re_set_one. However, this is not the right thing to do. If we look at breakpoint_re_set then we see that there's a whole bunch of state that needs to be preserved prior to calling breakpoint_re_set_one, and after calling breakpoint_re_set_one we still need to call update_global_location_list. I could just update the comment on breakpoint_re_set_one to make it clearer how the function should be used -- or more likely to warn that the function should only be used as a helper from breakpoint_re_set. However, breakpoint_re_set_one is only 3 lines long. So I figure it might actually be easier to just fold breakpoint_re_set_one into breakpoint_re_set, then there's no risk of accidentally calling breakpoint_re_set_one when we shouldn't. There should be no user visible changes after this commit. --- gdb/breakpoint.c | 26 ++++++++------------------ 1 file changed, 8 insertions(+), 18 deletions(-) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index c4accfb6d28..be035def12a 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -12812,17 +12812,6 @@ create_sals_from_location_spec_default (location_spec *locspec, parse_breakpoint_sals (locspec, canonical); } -/* Reset a breakpoint. */ - -static void -breakpoint_re_set_one (breakpoint *b) -{ - input_radix = b->input_radix; - set_language (b->language); - - b->re_set (); -} - /* Re-set breakpoint locations for the current program space. Locations bound to other program spaces are left untouched. */ @@ -12834,12 +12823,11 @@ breakpoint_re_set (void) scoped_restore save_input_radix = make_scoped_restore (&input_radix); scoped_restore_current_pspace_and_thread restore_pspace_thread; - /* breakpoint_re_set_one sets the current_language to the language - of the breakpoint it is resetting (see prepare_re_set_context) - before re-evaluating the breakpoint's location. This change can - unfortunately get undone by accident if the language_mode is set - to auto, and we either switch frames, or more likely in this context, - we select the current frame. + /* To ::re_set each breakpoint we set the current_language to the + language of the breakpoint before re-evaluating the breakpoint's + location. This change can unfortunately get undone by accident if + the language_mode is set to auto, and we either switch frames, or + more likely in this context, we select the current frame. We prevent this by temporarily turning the language_mode to language_mode_manual. We restore it once all breakpoints @@ -12856,7 +12844,9 @@ breakpoint_re_set (void) { try { - breakpoint_re_set_one (b); + input_radix = b->input_radix; + set_language (b->language); + b->re_set (); } catch (const gdb_exception &ex) { From patchwork Mon May 15 19:27:44 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69402 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 F03DF3857710 for ; Mon, 15 May 2023 19:29:35 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org F03DF3857710 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684178976; bh=4Uf0gbg5vvvtj8cq/FmaK8Epqc9CS7xHbx17vz0ynTA=; 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=DKz1DxNq7I+OqTvqp3uejdtSl+Stv6A0NgnxmKtKAY9jZ9JKq6uhR2ODs3CAVtHEd nzAnDCU0Ox/mPCA5BYe0UsZJMEuEwTkD5PGNEqBLGT5EjkmXuwhMa+wyKJjxXnhZBW o7/NCCeSEwtW6gCz0Vu8fu58d94StcIdnbavdgFs= 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 7BBE2385734D for ; Mon, 15 May 2023 19:28:12 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 7BBE2385734D 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-154-NEM2TmbDNGi4g0lqtftWtg-1; Mon, 15 May 2023 15:28:06 -0400 X-MC-Unique: NEM2TmbDNGi4g0lqtftWtg-1 Received: by mail-wr1-f71.google.com with SMTP id ffacd0b85a97d-3076fc5c278so7716880f8f.3 for ; Mon, 15 May 2023 12:28:06 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178885; x=1686770885; 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=4Uf0gbg5vvvtj8cq/FmaK8Epqc9CS7xHbx17vz0ynTA=; b=JKdQeGgBOLZILhgljectOFiPWxplhkMCsGNm+ief+1lUuuLLwxD+kSpq4zDAQrkWQ4 iqAj7zH8Rv56S/1i8qXleYzhFyJwwVe+k05trDOR0hNytiYxzZ2E84SHLFupGMNrH8in gnYBoX5W7FsmJ2o6vbMzpkxTBkTsBiC+f2xBDQSOvXqvxG3BJ7rJtlo25et2dJ/7NtlY 8E61/l/ENLywHbNnnfXq7Uw51qutDtMPWC3k2lkZU5siulBGR/2ezyD4drjGHCe91AiL 3hGBZgvjqKrojZACVgt9w/mwWBGJSWgGcSYdnCd9oBL3wd1Wkyi2fXSn99HY5geI8Fut Dayg== X-Gm-Message-State: AC+VfDx440Caotmef79WZ6uomODuxxnNhXXvgjbmkf9rArkVNpBpXr3J v+sZO1bvdHLM47tD30n8bDfxY1On8CSl9z8+pMZj6vWnxnLVwXpgaixancrFSYzRXo9P88zMdlU o38iwPtiE69EWBfxZ4slWWLig+WjvPpNCLf8YUExfcFNrWlILmIwrWDEnPfP3ceQGkd6DT7hPf5 xWK4GQlg== X-Received: by 2002:a5d:55cf:0:b0:304:b967:956f with SMTP id i15-20020a5d55cf000000b00304b967956fmr27708696wrw.8.1684178885051; Mon, 15 May 2023 12:28:05 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ4XWbD0vbrpDl1w1Y/CQjmWwi6GHiVHwwqK6AUDutOg3mi5nPzPuPig3xjJGKxW+2A+rgGhTA== X-Received: by 2002:a5d:55cf:0:b0:304:b967:956f with SMTP id i15-20020a5d55cf000000b00304b967956fmr27708684wrw.8.1684178884712; Mon, 15 May 2023 12:28:04 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id l11-20020a7bc34b000000b003f1751016desm296435wmj.28.2023.05.15.12.28.03 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:28:04 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess Subject: [PATCHv2 8/9] gdb: remove tracepoint_probe_create_sals_from_location_spec Date: Mon, 15 May 2023 20:27:44 +0100 Message-Id: <043a8df33eca35c82e68dca87e4e548f64872d0d.1684178710.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=-11.7 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, RCVD_IN_MSPIKE_H2, 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" The tracepoint_probe_create_sals_from_location_spec function just forwards all its arguments to bkpt_probe_create_sals_from_location_spec, and is only used in one place. Lets delete tracepoint_probe_create_sals_from_location_spec and replace it with bkpt_probe_create_sals_from_location_spec. There should be no user visible changes after this commit. --- gdb/breakpoint.c | 19 +++---------------- 1 file changed, 3 insertions(+), 16 deletions(-) diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index be035def12a..3964665d867 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -221,9 +221,6 @@ static bool strace_marker_p (struct breakpoint *b); static void bkpt_probe_create_sals_from_location_spec (location_spec *locspec, struct linespec_result *canonical); -static void tracepoint_probe_create_sals_from_location_spec - (location_spec *locspec, - struct linespec_result *canonical); const struct breakpoint_ops code_breakpoint_ops = { @@ -238,10 +235,11 @@ static const struct breakpoint_ops bkpt_probe_breakpoint_ops = create_breakpoints_sal, }; -/* Tracepoints set on probes. */ +/* Tracepoints set on probes. We use the same methods as for breakpoints + on probes. */ static const struct breakpoint_ops tracepoint_probe_breakpoint_ops = { - tracepoint_probe_create_sals_from_location_spec, + bkpt_probe_create_sals_from_location_spec, create_breakpoints_sal, }; @@ -12007,17 +12005,6 @@ tracepoint::print_recreate (struct ui_file *fp) const gdb_printf (fp, " passcount %d\n", pass_count); } -/* Virtual table for tracepoints on static probes. */ - -static void -tracepoint_probe_create_sals_from_location_spec - (location_spec *locspec, - struct linespec_result *canonical) -{ - /* We use the same method for breakpoint on probes. */ - bkpt_probe_create_sals_from_location_spec (locspec, canonical); -} - void dprintf_breakpoint::re_set () { From patchwork Mon May 15 19:27:45 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Andrew Burgess X-Patchwork-Id: 69403 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 41822385609A for ; Mon, 15 May 2023 19:30:05 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 41822385609A DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1684179005; bh=ONhdeklYT3m95jah97J8hqAops8OP+70ggbD8RgQTaA=; 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=R/oWSjTkfoZXtGKZWdEzjwEkPgw+WW+vlKcGz47NSuulXyXLMVrn9iK6hLoYajlEK 8nuBOcva9m+WmKDS1hc0jGSn1YYOLKbkRJHTFvyJ3OA8gGNSvCITUw4JOxsvCYoRhc UxSA+3zEtAvhX6Zx2TpPtQS1+Y7OVN3KwMfjPkz8= 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 951F6385771E for ; Mon, 15 May 2023 19:28:11 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 951F6385771E 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-396-M4kXMYhxPDCocWzVz_t6rw-1; Mon, 15 May 2023 15:28:10 -0400 X-MC-Unique: M4kXMYhxPDCocWzVz_t6rw-1 Received: by mail-wm1-f70.google.com with SMTP id 5b1f17b1804b1-3f429072212so32306705e9.1 for ; Mon, 15 May 2023 12:28:09 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1684178888; x=1686770888; 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=ONhdeklYT3m95jah97J8hqAops8OP+70ggbD8RgQTaA=; b=ZD8c6m32AbECHnar81slN2RWnu+VbuJfWBBPmw2/PJHlpb3FTJW8gzfYdWi1cGx/Re Zqv9nZVREkXxTCCX9wPlbHszcJwxZrsKLz4H0G8PjfySDkkut+ModuuxBWdbmO+gptUc xYul8WaBbvjMSb5vb/BOntTJyHiqzQRaEIwCwAdYRSr7CJ9JA2w9R2CCRkHkffX7lMLw 7UgdFA4VQvLTOyr3yaR6rAPzsqZv1nB2yaadt2Pz7p2o92EFyRZ78IOhGNsipwVQ8y6u J5/UfMkjJDA4gQDjbo1C7U1DCBn3H2uLI+wRI1weCSXDiJHr+fbOjaq2VBXZqIEk9jIE GqKw== X-Gm-Message-State: AC+VfDwTZJkvoy1nh2yDlMT7uKQUTp0dTMIzpTax4ps3D+HJoddhFrg6 Qd0IB/2ZobgspqUCSSnp9NwGX1dDNYtRPNg2FQldLdnM0oE8e2R0HICNm7TDXtb20QLVYRhJ/qG djpZ37Fgd5Y1QBgrFVF/Z848AqthUHWieowO3jZApy7MLYv7efcbFKX+T3brNsBjkNtJN86H33S iX2OC3xQ== X-Received: by 2002:a05:600c:c5:b0:3f4:2d85:53f6 with SMTP id u5-20020a05600c00c500b003f42d8553f6mr15293851wmm.34.1684178887968; Mon, 15 May 2023 12:28:07 -0700 (PDT) X-Google-Smtp-Source: ACHHUZ4NyxiH/g58SJPgRtL1zc/Vtlcf1hSzIfegJNCIhKNkG5mUZ1RWsf7f/1s3qdyx7VUzVVxHpQ== X-Received: by 2002:a05:600c:c5:b0:3f4:2d85:53f6 with SMTP id u5-20020a05600c00c500b003f42d8553f6mr15293828wmm.34.1684178887178; Mon, 15 May 2023 12:28:07 -0700 (PDT) Received: from localhost (11.72.115.87.dyn.plus.net. [87.115.72.11]) by smtp.gmail.com with ESMTPSA id n11-20020a5d4c4b000000b002c70ce264bfsm77214wrt.76.2023.05.15.12.28.05 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 15 May 2023 12:28:06 -0700 (PDT) To: gdb-patches@sourceware.org Cc: Andrew Burgess , Eli Zaretskii Subject: [PATCHv2 9/9] gdb: only insert thread-specific breakpoints in the relevant inferior Date: Mon, 15 May 2023 20:27:45 +0100 Message-Id: <628e17ca547f5b11096b036b1844ddb92da13da8.1684178710.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=-11.7 required=5.0 tests=BAYES_00, DKIM_INVALID, DKIM_SIGNED, GIT_PATCH_0, KAM_DMARC_NONE, KAM_DMARC_STATUS, RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H2, 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" Eli approved the docs changes here: https://sourceware.org/pipermail/gdb-patches/2023-April/199222.html --- This commit updates GDB so that thread-specific breakpoints are only inserted into the inferior that contains the thread we are interested in. Actually, as breakpoints are placed in program spaces, we insert the breakpoint in any inferior that shares a program space with the inferior containing the thread we are interested in. But as far as most users are concerned this really means the one inferior. In terms of implementation, getting this basically working is easy enough, now that a breakpoint's thread field is setup prior to GDB looking for locations, we can easily use the thread to find a suitable program_space and pass this to as a filter when creating the sals. Or we could if breakpoint_ops::create_sals_from_location_spec allowed us to pass in a filter program_space. So, this commit extends breakpoint_ops::create_sals_from_location_spec to take a program_space argument, and uses this to filter the set of returned sals. This accounts for about half the change in this patch. The second set of changes starts from breakpoint_set_thread, this is called when the thread for a thread-specific breakpoint changes, e.g. from the Python API. Previously this call would never result in the locations of a breakpoint changing, after all, locations were inserted in every program space, and we just use the thread field to decide when we should stop. Now though, changing a breakpoint's thread can mean we need to figure out a new set of breakpoint locations. To support this I've added a new breakpoint_re_set_one function, which is like breakpoint_re_set, but takes a single breakpoint, and just updates the locations for that one breakpoint. We only need to call this function if the program_space in which a breakpoint's thread is running actually changes. If the program_space does change then we call the new breakpoint_re_set_one function passing in the program_space which should be used to filter the locations (or nullptr to indicate we should set locations in all program spaces). This filter program_space needs to propagate down to all the re_set methods, this accounts for the remaining half of the changes in this patch. There were a couple of existing tests that created thread-specific breakpoints and then checked the 'info breakpoints' output, these needed updating. These were: gdb.mi/user-selected-context-sync.exp gdb.multi/bp-thread-specific.exp gdb.multi/multi-target-continue.exp gdb.multi/multi-target-ping-pong-next.exp gdb.multi/tids.exp I've also added some additional tests to: gdb.multi/pending-bp.exp I've updated the documentation and added a NEWS entry. Reviewed-By: Eli Zaretskii --- gdb/NEWS | 7 + gdb/ada-lang.c | 6 +- gdb/break-catch-throw.c | 6 +- gdb/breakpoint.c | 236 ++++++++++++++---- gdb/breakpoint.h | 26 +- .../gdb.mi/user-selected-context-sync.exp | 2 +- .../gdb.multi/bp-thread-specific.exp | 7 +- .../gdb.multi/multi-target-continue.exp | 2 +- .../gdb.multi/multi-target-ping-pong-next.exp | 4 +- gdb/testsuite/gdb.multi/pending-bp.exp | 206 +++++++++++++++ gdb/testsuite/gdb.multi/tids.exp | 6 +- 11 files changed, 427 insertions(+), 81 deletions(-) diff --git a/gdb/NEWS b/gdb/NEWS index 376c4c223be..8e0ed944fcf 100644 --- a/gdb/NEWS +++ b/gdb/NEWS @@ -71,6 +71,13 @@ 'thread' or 'task' keywords are parsed at the time the breakpoint is created, rather than at the time the breakpoint becomes non-pending. +* Thread-specific breakpoints are only inserted into the program space + in which the thread of interest is running. In most cases program + spaces are unique for each inferior, so this means that + thread-specific breakpoints will usually only be inserted for the + inferior containing the thread of interest. The breakpoint will + be hit no less than before. + * New commands maintenance print record-instruction [ N ] diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c index 21f3348a161..474601274ad 100644 --- a/gdb/ada-lang.c +++ b/gdb/ada-lang.c @@ -12176,7 +12176,7 @@ struct ada_catchpoint : public code_breakpoint } struct bp_location *allocate_location () override; - void re_set () override; + void re_set (program_space *pspace) override; void check_status (struct bpstat *bs) override; enum print_stop_action print_it (const bpstat *bs) const override; bool print_one (bp_location **) const override; @@ -12270,11 +12270,11 @@ ada_catchpoint::allocate_location () catchpoint kinds. */ void -ada_catchpoint::re_set () +ada_catchpoint::re_set (program_space *pspace) { /* Call the base class's method. This updates the catchpoint's locations. */ - this->code_breakpoint::re_set (); + this->code_breakpoint::re_set (pspace); /* Reparse the exception conditional expressions. One for each location. */ diff --git a/gdb/break-catch-throw.c b/gdb/break-catch-throw.c index 457446efbc6..854ca6a624c 100644 --- a/gdb/break-catch-throw.c +++ b/gdb/break-catch-throw.c @@ -82,10 +82,10 @@ struct exception_catchpoint : public code_breakpoint _("invalid type-matching regexp"))) { pspace = current_program_space; - re_set (); + re_set (pspace); } - void re_set () override; + void re_set (program_space *pspace) override; enum print_stop_action print_it (const bpstat *bs) const override; bool print_one (bp_location **) const override; void print_mention () const override; @@ -200,7 +200,7 @@ exception_catchpoint::check_status (struct bpstat *bs) /* Implement the 're_set' method. */ void -exception_catchpoint::re_set () +exception_catchpoint::re_set (program_space *pspace) { std::vector sals; struct program_space *filter_pspace = current_program_space; diff --git a/gdb/breakpoint.c b/gdb/breakpoint.c index 3964665d867..53c0c69a441 100644 --- a/gdb/breakpoint.c +++ b/gdb/breakpoint.c @@ -90,9 +90,12 @@ static void map_breakpoint_numbers (const char *, gdb::function_view); -static void - create_sals_from_location_spec_default (location_spec *locspec, - linespec_result *canonical); +static void parse_breakpoint_sals (location_spec *locspec, + linespec_result *canonical, + program_space *search_pspace); + +static void breakpoint_re_set_one (breakpoint *b, + program_space *filter_pspace); static void create_breakpoints_sal (struct gdbarch *, struct linespec_result *, @@ -220,11 +223,12 @@ static bool strace_marker_p (struct breakpoint *b); static void bkpt_probe_create_sals_from_location_spec (location_spec *locspec, - struct linespec_result *canonical); + struct linespec_result *canonical, + struct program_space *search_pspace); const struct breakpoint_ops code_breakpoint_ops = { - create_sals_from_location_spec_default, + parse_breakpoint_sals, create_breakpoints_sal, }; @@ -289,7 +293,7 @@ struct internal_breakpoint : public code_breakpoint disposition = disp_donttouch; } - void re_set () override; + void re_set (program_space *pspace) override; void check_status (struct bpstat *bs) override; enum print_stop_action print_it (const bpstat *bs) const override; void print_mention () const override; @@ -323,7 +327,7 @@ struct momentary_breakpoint : public code_breakpoint thread = thread_; } - void re_set () override; + void re_set (program_space *pspace) override; void check_status (struct bpstat *bs) override; enum print_stop_action print_it (const bpstat *bs) const override; void print_mention () const override; @@ -334,7 +338,7 @@ struct dprintf_breakpoint : public ordinary_breakpoint { using ordinary_breakpoint::ordinary_breakpoint; - void re_set () override; + void re_set (program_space *pspace) override; int breakpoint_hit (const struct bp_location *bl, const address_space *aspace, CORE_ADDR bp_addr, @@ -1460,7 +1464,36 @@ breakpoint_set_thread (struct breakpoint *b, int thread) b->thread = thread; if (old_thread != thread) - gdb::observers::breakpoint_modified.notify (b); + { + /* If THREAD is in a different program_space than OLD_THREAD, or the + breakpoint has switched to or from being thread-specific, then we + need to re-set the locations of this breakpoint. First, figure + out the program_space for the old and new threads, use a value of + nullptr to indicate the breakpoint is in all program spaces. */ + program_space *old_pspace = nullptr; + if (old_thread != -1) + { + struct thread_info *thr = find_thread_global_id (old_thread); + gdb_assert (thr != nullptr); + old_pspace = thr->inf->pspace; + } + + program_space *new_pspace = nullptr; + if (thread != -1) + { + struct thread_info *thr = find_thread_global_id (thread); + gdb_assert (thr != nullptr); + new_pspace = thr->inf->pspace; + } + + /* If the program space has changed for this breakpoint, then + re-evaluate it's locations. */ + if (old_pspace != new_pspace) + breakpoint_re_set_one (b, new_pspace); + + /* Let others know the breakpoint has changed. */ + gdb::observers::breakpoint_modified.notify (b); + } } /* See breakpoint.h. */ @@ -8653,7 +8686,8 @@ create_breakpoints_sal (struct gdbarch *gdbarch, static void parse_breakpoint_sals (location_spec *locspec, - struct linespec_result *canonical) + struct linespec_result *canonical, + struct program_space *search_pspace) { struct symtab_and_line cursal; @@ -8717,7 +8751,7 @@ parse_breakpoint_sals (location_spec *locspec, && strchr ("+-", spec[0]) != NULL && spec[1] != '[')) { - decode_line_full (locspec, DECODE_LINE_FUNFIRSTLINE, NULL, + decode_line_full (locspec, DECODE_LINE_FUNFIRSTLINE, search_pspace, get_last_displayed_symtab (), get_last_displayed_line (), canonical, NULL, NULL); @@ -8725,7 +8759,7 @@ parse_breakpoint_sals (location_spec *locspec, } } - decode_line_full (locspec, DECODE_LINE_FUNFIRSTLINE, NULL, + decode_line_full (locspec, DECODE_LINE_FUNFIRSTLINE, search_pspace, cursal.symtab, cursal.line, canonical, NULL, NULL); } @@ -8824,6 +8858,24 @@ breakpoint_ops_for_location_spec_type (enum location_spec_type locspec_type, } } +/* Return the program space to use as a filter when searching for locations + of a breakpoint specific to THREAD. If THREAD is -1, meaning all + threads, then this function returns nullptr, indicating no program space + filtering should be performed. Otherwise, this function returns the + program space for the inferior that contains THREAD. */ + +static struct program_space * +find_program_space_for_global_thread_id (int thread) +{ + if (thread == -1) + return nullptr; + + struct thread_info *thr = find_thread_global_id (thread); + gdb_assert (thr != nullptr); + gdb_assert (thr->inf != nullptr); + return thr->inf->pspace; +} + /* See breakpoint.h. */ const struct breakpoint_ops * @@ -8917,7 +8969,10 @@ create_breakpoint (struct gdbarch *gdbarch, try { - ops->create_sals_from_location_spec (locspec, &canonical); + struct program_space *search_pspace + = find_program_space_for_global_thread_id (thread); + ops->create_sals_from_location_spec (locspec, &canonical, + search_pspace); } catch (const gdb_exception_error &e) { @@ -9393,7 +9448,7 @@ break_range_command (const char *arg, int from_tty) arg_start = arg; location_spec_up start_locspec = string_to_location_spec (&arg, current_language); - parse_breakpoint_sals (start_locspec.get (), &canonical_start); + parse_breakpoint_sals (start_locspec.get (), &canonical_start, nullptr); if (arg[0] != ',') error (_("Too few arguments.")); @@ -9480,7 +9535,7 @@ watchpoint_exp_is_const (const struct expression *exp) /* Implement the "re_set" method for watchpoints. */ void -watchpoint::re_set () +watchpoint::re_set (struct program_space *pspace) { /* Watchpoint can be either on expression using entirely global variables, or it can be on local variables. @@ -11561,7 +11616,7 @@ breakpoint::print_recreate (struct ui_file *fp) const /* Default breakpoint_ops methods. */ void -code_breakpoint::re_set () +code_breakpoint::re_set (struct program_space *pspace) { /* FIXME: is this still reachable? */ if (breakpoint_location_spec_empty_p (this)) @@ -11571,7 +11626,7 @@ code_breakpoint::re_set () return; } - re_set_default (); + re_set_default (pspace); } int @@ -11777,7 +11832,7 @@ code_breakpoint::decode_location_spec (location_spec *locspec, /* Virtual table for internal breakpoints. */ void -internal_breakpoint::re_set () +internal_breakpoint::re_set (struct program_space *pspace) { switch (type) { @@ -11870,7 +11925,7 @@ internal_breakpoint::print_mention () const /* Virtual table for momentary breakpoints */ void -momentary_breakpoint::re_set () +momentary_breakpoint::re_set (struct program_space *pspace) { /* Keep temporary breakpoints, which can be encountered when we step over a dlopen call and solib_add is resetting the breakpoints. @@ -11911,12 +11966,13 @@ longjmp_breakpoint::~longjmp_breakpoint () static void bkpt_probe_create_sals_from_location_spec (location_spec *locspec, - struct linespec_result *canonical) + struct linespec_result *canonical, + struct program_space *search_pspace) { struct linespec_sals lsal; - lsal.sals = parse_probes (locspec, NULL, canonical); + lsal.sals = parse_probes (locspec, search_pspace, canonical); lsal.canonical = xstrdup (canonical->locspec->to_string ()); canonical->lsals.push_back (std::move (lsal)); } @@ -12006,9 +12062,9 @@ tracepoint::print_recreate (struct ui_file *fp) const } void -dprintf_breakpoint::re_set () +dprintf_breakpoint::re_set (struct program_space *pspace) { - re_set_default (); + re_set_default (pspace); /* extra_string should never be non-NULL for dprintf. */ gdb_assert (extra_string != NULL); @@ -12066,8 +12122,10 @@ dprintf_breakpoint::after_condition_true (struct bpstat *bs) markers (`-m'). */ static void -strace_marker_create_sals_from_location_spec (location_spec *locspec, - struct linespec_result *canonical) +strace_marker_create_sals_from_location_spec + (location_spec *locspec, + struct linespec_result *canonical, + struct program_space *search_pspace) { struct linespec_sals lsal; const char *arg_start, *arg; @@ -12597,12 +12655,32 @@ update_breakpoint_locations (code_breakpoint *b, all locations are in the same shared library, that was unloaded. We'd like to retain the location, so that when the library is loaded again, we don't loose the enabled/disabled status of the - individual locations. */ + individual locations. + + Thread specific breakpoints will also trigger this case if the thread + is changed to a different program space, and all of the old locations + go out of scope. In this case we do (currently) discard the old + locations -- we assume the change in thread is permanent and the old + locations will never come back into scope. */ if (all_locations_are_pending (b, filter_pspace) && sals.empty ()) - return; + { + if (b->thread != -1) + b->loc = nullptr; + return; + } existing_locations = hoist_existing_locations (b, filter_pspace); + /* If this is a thread-specific breakpoint then any locations left on the + breakpoint are for a program space in which the thread of interest + does not operate. This can happen when the user changes the thread of + a thread-specific breakpoint. + + We assume that the change in thread is permanent, and that the old + locations will never be used again, so discard them now. */ + if (b->thread != -1) + b->loc = nullptr; + for (const auto &sal : sals) { struct bp_location *new_loc; @@ -12765,40 +12843,45 @@ code_breakpoint::location_spec_to_sals (location_spec *locspec, locations. */ void -code_breakpoint::re_set_default () +code_breakpoint::re_set_default (struct program_space *filter_pspace) { - struct program_space *filter_pspace = current_program_space; std::vector expanded, expanded_end; - int found; - std::vector sals = location_spec_to_sals (locspec.get (), - filter_pspace, - &found); - if (found) - expanded = std::move (sals); - - if (locspec_range_end != nullptr) - { - std::vector sals_end - = location_spec_to_sals (locspec_range_end.get (), - filter_pspace, &found); + /* If this breakpoint is thread-specific then find the program space in + which the specific thread exists. Otherwise, for breakpoints that are + not thread-specific THREAD_PSPACE will be nullptr. */ + program_space *thread_pspace + = find_program_space_for_global_thread_id (this->thread); + + /* If this is not a thread-specific breakpoint, or it is a + thread-specific breakpoint but we are looking for new locations in the + program space that the specific thread is running, then look for new + locations for this breakpoint. */ + if (thread_pspace == nullptr || filter_pspace == thread_pspace) + { + int found; + std::vector sals + = location_spec_to_sals (locspec.get (), filter_pspace, &found); if (found) - expanded_end = std::move (sals_end); + expanded = std::move (sals); + + if (locspec_range_end != nullptr) + { + std::vector sals_end + = location_spec_to_sals (locspec_range_end.get (), + filter_pspace, &found); + if (found) + expanded_end = std::move (sals_end); + } } + /* Update the locations for this breakpoint. For thread-specific + breakpoints this will remove any old locations that are for the wrong + program space -- this can happen if the user changes the thread of a + thread-specific breakpoint. */ update_breakpoint_locations (this, filter_pspace, expanded, expanded_end); } -/* Default method for creating SALs from an address string. It basically - calls parse_breakpoint_sals. Return 1 for success, zero for failure. */ - -static void -create_sals_from_location_spec_default (location_spec *locspec, - struct linespec_result *canonical) -{ - parse_breakpoint_sals (locspec, canonical); -} - /* Re-set breakpoint locations for the current program space. Locations bound to other program spaces are left untouched. */ @@ -12833,7 +12916,7 @@ breakpoint_re_set (void) { input_radix = b->input_radix; set_language (b->language); - b->re_set (); + b->re_set (current_program_space); } catch (const gdb_exception &ex) { @@ -12854,6 +12937,53 @@ breakpoint_re_set (void) /* Now we can insert. */ update_global_location_list (UGLL_MAY_INSERT); } + +/* Re-set locations for breakpoint B in FILTER_PSPACE. If FILTER_PSPACE is + nullptr then re-set locations for B in all program spaces. Locations + bound to program spaces other than FILTER_PSPACE are left untouched. */ + +static void +breakpoint_re_set_one (breakpoint *b, program_space *filter_pspace) +{ + { + scoped_restore_current_language save_language; + scoped_restore save_input_radix = make_scoped_restore (&input_radix); + scoped_restore_current_pspace_and_thread restore_pspace_thread; + + /* To ::re_set each breakpoint we set the current_language to the + language of the breakpoint before re-evaluating the breakpoint's + location. This change can unfortunately get undone by accident if + the language_mode is set to auto, and we either switch frames, or + more likely in this context, we select the current frame. + + We prevent this by temporarily turning the language_mode to + language_mode_manual. We restore it once all breakpoints + have been reset. */ + scoped_restore save_language_mode = make_scoped_restore (&language_mode); + language_mode = language_mode_manual; + + /* Note: we must not try to insert locations until after all + breakpoints have been re-set. Otherwise, e.g., when re-setting + breakpoint 1, we'd insert the locations of breakpoint 2, which + hadn't been re-set yet, and thus may have stale locations. */ + + try + { + input_radix = b->input_radix; + set_language (b->language); + b->re_set (filter_pspace); + } + catch (const gdb_exception &ex) + { + exception_fprintf (gdb_stderr, ex, + "Error in re-setting breakpoint %d: ", + b->number); + } + } + + /* Now we can insert. */ + update_global_location_list (UGLL_MAY_INSERT); +} /* Reset the thread number of this breakpoint: diff --git a/gdb/breakpoint.h b/gdb/breakpoint.h index 2adcc9fa338..8a643301264 100644 --- a/gdb/breakpoint.h +++ b/gdb/breakpoint.h @@ -564,15 +564,15 @@ enum print_stop_action struct breakpoint_ops { - /* Create SALs from location spec, storing the result in - linespec_result. - - For an explanation about the arguments, see the function - `create_sals_from_location_spec_default'. + /* Create SALs from LOCSPEC, storing the result in linespec_result + CANONICAL. If SEARCH_PSPACE is not nullptr then only results in the + corresponding program space are returned. If SEARCH_PSPACE is nullptr + then results for all program spaces are returned. This function is called inside `create_breakpoint'. */ void (*create_sals_from_location_spec) (location_spec *locspec, - struct linespec_result *canonical); + linespec_result *canonical, + program_space *search_pspace); /* This method will be responsible for creating a breakpoint given its SALs. Usually, it just calls `create_breakpoints_sal' (for ordinary @@ -635,8 +635,12 @@ struct breakpoint /* Reevaluate a breakpoint. This is necessary after symbols change (e.g., an executable or DSO was loaded, or the inferior just - started). */ - virtual void re_set () + started). + + FILTER_PSPACE is the program space in which symbols may have changed, + or can be nullptr to indicate that all program spaces may have + changed. */ + virtual void re_set (program_space *filter_pspace) { /* Nothing to re-set. */ } @@ -868,7 +872,7 @@ struct code_breakpoint : public breakpoint /* Add a location for SAL to this breakpoint. */ bp_location *add_location (const symtab_and_line &sal); - void re_set () override; + void re_set (program_space *pspace) override; int insert_location (struct bp_location *) override; int remove_location (struct bp_location *, enum remove_bp_reason reason) override; @@ -890,7 +894,7 @@ struct code_breakpoint : public breakpoint struct program_space *search_pspace); /* Helper method that does the basic work of re_set. */ - void re_set_default (); + void re_set_default (program_space *pspace); /* Find the SaL locations corresponding to the given LOCATION. On return, FOUND will be 1 if any SaL was found, zero otherwise. */ @@ -912,7 +916,7 @@ struct watchpoint : public breakpoint { using breakpoint::breakpoint; - void re_set () override; + void re_set (program_space *pspace) override; int insert_location (struct bp_location *) override; int remove_location (struct bp_location *, enum remove_bp_reason reason) override; diff --git a/gdb/testsuite/gdb.mi/user-selected-context-sync.exp b/gdb/testsuite/gdb.mi/user-selected-context-sync.exp index 4889c31aff3..8cc47537e68 100644 --- a/gdb/testsuite/gdb.mi/user-selected-context-sync.exp +++ b/gdb/testsuite/gdb.mi/user-selected-context-sync.exp @@ -370,7 +370,7 @@ proc test_continue_to_start { mode inf } { "thread $inf.2 stops MI" } else { mi_expect_stop "breakpoint-hit" "child_sub_function" \ - "" "$srcfile" "$decimal" {"" "disp=\"del\"" "locno=\"[0-9]+\""} \ + "" "$srcfile" "$decimal" {"" "disp=\"del\""} \ "thread $inf.2 stops MI" } } diff --git a/gdb/testsuite/gdb.multi/bp-thread-specific.exp b/gdb/testsuite/gdb.multi/bp-thread-specific.exp index 85c08f44a2c..68001e92044 100644 --- a/gdb/testsuite/gdb.multi/bp-thread-specific.exp +++ b/gdb/testsuite/gdb.multi/bp-thread-specific.exp @@ -50,7 +50,7 @@ gdb_test "info threads" \ # locations ('foo' in both inferiors) even though only one of those # locations will ever trigger ('foo' in inferior 2). gdb_test "break foo thread 2.1" \ - "Breakpoint $decimal at $hex: foo\\. \\(2 locations\\)" + "Breakpoint $decimal at $hex: file \[^\r\n\]+$srcfile, line $decimal\\." set bpnum [get_integer_valueof "\$bpnum" "INVALID"] @@ -58,10 +58,7 @@ set bpnum [get_integer_valueof "\$bpnum" "INVALID"] # earlier breakpoint. Check that the thread-id used when describing # the earlier breakpoints is correct. gdb_test "break foo thread 1.1" \ - [multi_line \ - "Note: breakpoint $bpnum \\(thread 2.1\\) also set at pc $hex\\." \ - "Note: breakpoint $bpnum \\(thread 2.1\\) also set at pc $hex\\." \ - "Breakpoint $decimal at $hex: foo\\. \\(2 locations\\)"] + "Breakpoint $decimal at $hex: file \[^\r\n\]+$srcfile, line $decimal\\." # Save the breakpoints into a file. if {[is_remote host]} { diff --git a/gdb/testsuite/gdb.multi/multi-target-continue.exp b/gdb/testsuite/gdb.multi/multi-target-continue.exp index bf19cc00968..a1bffb7b613 100644 --- a/gdb/testsuite/gdb.multi/multi-target-continue.exp +++ b/gdb/testsuite/gdb.multi/multi-target-continue.exp @@ -30,7 +30,7 @@ proc test_continue {non-stop} { proc set_break {inf} { gdb_test "break function${inf} thread ${inf}.1" \ - "Breakpoint .* function${inf}\\..*" + "Breakpoint ${::decimal} at ${::hex}: file .*, line ${::decimal}\\." } # Select inferior INF, and then run to a breakpoint on inferior diff --git a/gdb/testsuite/gdb.multi/multi-target-ping-pong-next.exp b/gdb/testsuite/gdb.multi/multi-target-ping-pong-next.exp index 64645b8ec8a..51fd70b018a 100644 --- a/gdb/testsuite/gdb.multi/multi-target-ping-pong-next.exp +++ b/gdb/testsuite/gdb.multi/multi-target-ping-pong-next.exp @@ -52,12 +52,12 @@ proc test_ping_pong_next {} { gdb_test "thread 1.1" "Switching to thread 1.1 .*" gdb_test "break $srcfile:$line1 thread 1.1" \ - "Breakpoint .*$srcfile:$line1\\..*" + "Breakpoint .*$srcfile, line $line1\\." gdb_test "continue" "hit Breakpoint .*" gdb_test "break $srcfile:$line2 thread 2.1" \ - "Breakpoint .*$srcfile:$line2\\..*" + "Breakpoint .*$srcfile, line $line2\\." # Now block inferior 1 and issue "next". We should stop at the # breakpoint for inferior 2, given schedlock off. diff --git a/gdb/testsuite/gdb.multi/pending-bp.exp b/gdb/testsuite/gdb.multi/pending-bp.exp index 4a65f79cfc4..8e1d1234948 100644 --- a/gdb/testsuite/gdb.multi/pending-bp.exp +++ b/gdb/testsuite/gdb.multi/pending-bp.exp @@ -78,6 +78,48 @@ proc do_test_setup { inf_1_stop inf_2_stop } { return true } +# Create a breakpoint on the function 'foo' in THREAD. It is expected +# that the breakpoint created will be pending, this is checked by +# running the 'info breakpoints' command. +# +# Returns the number for the newly created breakpoint. +proc do_create_pending_foo_breakpoint { {thread "1.1"} } { + gdb_test "break foo thread $thread" \ + [multi_line \ + "Function \"foo\" not defined\\." \ + "Breakpoint $::decimal \\(foo\\) pending\."] \ + "set pending thread-specific breakpoint" + set bpnum [get_integer_valueof "\$bpnum" "*INVALID*" \ + "get number for thread-specific breakpoint on foo"] + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+\\s+foo" \ + "\\s+stop only in thread [string_to_regexp $thread]"] \ + "check thread-specific breakpoint is initially pending" + + return $bpnum +} + +# Create a breakpoint on the function 'foo' in THREAD. It is expected +# that the breakpoint created will not be pending, this is checked by +# running the 'info breakpoints' command. +# +# Returns the number for the newly created breakpoint. +proc do_create_foo_breakpoint { {thread "1.1"} } { + gdb_test "break foo thread $thread" \ + "Breakpoint $::decimal at $::hex" \ + "set thread-specific breakpoint" + set bpnum [get_integer_valueof "\$bpnum" "*INVALID*" \ + "get number for thread-specific breakpoint on foo"] + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+$::hex\\s+\]*> inf $::decimal" \ + "\\s+stop only in thread [string_to_regexp $thread]"] \ + "check thread-specific breakpoint is initially pending" + + return $bpnum +} + # Check that when a breakpoint is in the pending state, but that breakpoint # does have some locations (those locations themselves are pending), GDB # doesn't display the inferior list in the 'info breakpoints' output. @@ -111,5 +153,169 @@ proc_with_prefix test_no_inf_display {} { "check info breakpoints while breakpoint is pending" } +# Setup two inferiors. In #1 the symbol 'foo' has not yet been +# loaded, while in #2 the symbol 'foo' has been loaded. +# +# Create a thread-specific breakpoint on 'foo' tied to a thread in +# inferior #1, the breakpoint should be pending -- 'foo' is not yet +# loaded in #1. +# +# Now move inferior #1 forward until 'foo' is loaded, check the +# breakpoint is no longer pending. +# +# Move inferior #1 forward more until 'foo' is unloaded, check that +# the breakpoint returns to the pending state. +proc_with_prefix test_pending_toggle { } { + + do_test_setup "Break before open" "Break before close" + + set bpnum [do_create_pending_foo_breakpoint] + + # Now return to inferior 1 and continue until the shared library is + # loaded, the breakpoint should become non-pending. + gdb_test "inferior 1" "Switching to inferior 1 .*" \ + "switch back to inferior 1" + gdb_continue_to_breakpoint "stop in foo in inferior 1" "foo \\(\\) .*" + + gdb_test "info breakpoint $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+$::hex \]*> inf 1" \ + "\\s+stop only in thread 1\\.1" \ + "\\s+breakpoint already hit 1 time"] \ + "check thread-specific breakpoint is no longer pending" + + gdb_breakpoint [gdb_get_line_number "Break after close"] temporary + gdb_continue_to_breakpoint "close library" + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+\\s+foo" \ + "\\s+stop only in thread 1\\.1" \ + "\\s+breakpoint already hit 1 time"] \ + "check thread-specific breakpoint is pending again" +} + +# Create a Python variable VAR and set it to the gdb.Breakpoint object +# corresponding to the breakpoint numbered BPNUM. If THREAD is not +# the empty string then THREAD should be an integer, check that +# gdb.Breakpoint.thread is set to the value of THREAD. Otherwise, if +# THREAD is the empty string, check that gdb.Breakpoint.thread is set +# to None. +proc py_find_breakpoint { var bpnum {thread ""} } { + gdb_test_no_output \ + "python ${var}=\[b for b in gdb.breakpoints() if b.number == $bpnum\]\[0\]" \ + "find Python gdb.Breakpoint object" + if { $thread ne "" } { + gdb_test_no_output "python assert(${var}.thread == ${thread})" \ + "check thread attribute is currently correct" + } else { + gdb_test_no_output "python assert(${var}.thread is None)" \ + "check thread attribute is currently correct" + } +} + +# Setup two inferiors. In #1 the symbol 'foo' has not yet been +# loaded, while in #2 the symbol 'foo' has been loaded. +# +# Create a thread-specific breakpoint on 'foo' tied to a thread in +# inferior #1, the breakpoint should be pending -- 'foo' is not yet +# loaded in #1. +# +# Use Python to change the thread of the thread-specific breakpoint to +# a thread in inferior #2, at this point the thread should gain a +# location and become non-pending. +# +# Set the thread back to a thread in inferior #1, the breakpoint +# should return to the pending state. +proc_with_prefix py_test_toggle_thread {} { + do_test_setup "Break before open" "Break after open" + + set bpnum [do_create_pending_foo_breakpoint] + + py_find_breakpoint "bp" $bpnum 1 + + gdb_test_no_output "python bp.thread = 2" \ + "change thread on thread-specific breakpoint" + gdb_test "info breakpoint $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+$::hex \]*> inf 2" \ + "\\s+stop only in thread 2\\.1"] \ + "check thread-specific breakpoint now has a location" + + gdb_test_no_output "set call_count = 2" "set call_count in inferior 2" + gdb_continue_to_breakpoint "stop at foo in inferior 2" "foo \\(\\) .*" + + gdb_test_no_output "python bp.thread = 1" \ + "restore thread on thread-specific breakpoint" + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "$bpnum\\s+breakpoint\\s+keep\\s+y\\s+\\s+foo" \ + "\\s+stop only in thread 1\\.1" \ + "\\s+breakpoint already hit 1 time"] \ + "check thread-specific breakpoint has returned to pending" + + gdb_breakpoint [gdb_get_line_number "Break after close"] temporary + gdb_continue_to_breakpoint "stop after close in inferior 2" \ + ".* Break after close\\. .*" + + gdb_test "inferior 1" "Switching to inferior 1 .*" \ + "switch to inferior 1" + gdb_continue_to_breakpoint "stop at foo in inferior 1" "foo \\(\\) .*" +} + +# Setup two inferiors. Both inferiors have the symbol 'foo' +# available. +# +# Create a thread-specific breakpoint on 'foo' tied to a thread in +# inferior #1, the breakpoint should not be pending, but will only +# have a single location, the location in inferior #1. +# +# Use Python to change the thread of the thread-specific breakpoint to +# None. At this point the breakpoint should gain a second location, a +# location in inferior #2. +proc_with_prefix py_test_clear_thread {} { + do_test_setup "Break after open" "Break after open" + + set bpnum [do_create_foo_breakpoint] + + py_find_breakpoint "bp" $bpnum 1 + + gdb_test_no_output "python bp.thread = None" \ + "clear thread on thread-specific breakpoint" + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "${bpnum}\\s+breakpoint\\s+keep y\\s+\\s*" \ + "${bpnum}\\.1\\s+y\\s+${::hex}\\s+\]*> inf $::decimal" \ + "${bpnum}\\.2\\s+y\\s+${::hex}\\s+\]*> inf $::decimal"] \ + "check for a location in both inferiors" + + gdb_continue_to_breakpoint "stop at foo in inferior 2" "foo \\(\\) .*" + gdb_test_no_output "set call_count = 2" "set call_count in inferior 2" + + gdb_test "inferior 1" "Switching to inferior 1 .*" \ + "switch to inferior 1" + gdb_continue_to_breakpoint "stop at foo in inferior 1" "foo \\(\\) .*" + gdb_test_no_output "set call_count = 2" "set call_count in inferior 1" + + gdb_test_no_output "python bp.thread = 2" + gdb_test "info breakpoints $bpnum" \ + [multi_line \ + "${bpnum}\\s+breakpoint\\s+keep y\\s+${::hex}\\s+\]*> inf 2" \ + "\\s+stop only in thread 2\\.1" \ + "\\s+breakpoint already hit 2 times"] \ + "check for a location only in inferior 2" + + gdb_breakpoint [gdb_get_line_number "Break after close"] temporary + gdb_continue_to_breakpoint "stop after close in inferior 1" \ + ".* Break after close\\. .*" + + gdb_test "inferior 2" "Switching to inferior 2 .*" \ + "switch back to inferior 2" + gdb_continue_to_breakpoint "stop at foo again in inferior 2" \ + "foo \\(\\) .*" +} + # Run all the tests. test_no_inf_display +test_pending_toggle +py_test_toggle_thread +py_test_clear_thread diff --git a/gdb/testsuite/gdb.multi/tids.exp b/gdb/testsuite/gdb.multi/tids.exp index 18fc4970fe7..34c270309de 100644 --- a/gdb/testsuite/gdb.multi/tids.exp +++ b/gdb/testsuite/gdb.multi/tids.exp @@ -433,11 +433,13 @@ if { [allow_python_tests] } { gdb_py_test_silent_cmd "python bp = gdb.breakpoints()\[0\]" \ "get python breakpoint" 0 - gdb_test "python bp.thread = 6" "thread = 6" \ + gdb_test_no_output "python bp.thread = 6" \ "make breakpoint thread-specific with python" # Check that the inferior-qualified ID is correct. gdb_test "info breakpoint" \ - "stop only in thread 1.3\r\n.*" \ + [multi_line \ + "$decimal\\s+\[^\r\n\]+ in thread_function1 at \[^\r\n\]+" \ + "\\s+stop only in thread 1\\.3"] \ "thread specific breakpoint right thread" } }