From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 27502 invoked by alias); 25 Oct 2012 17:42:24 -0000 Received: (qmail 27453 invoked by uid 22791); 25 Oct 2012 17:42:23 -0000 X-SWARE-Spam-Status: No, hits=-8.0 required=5.0 tests=AWL,BAYES_00,KHOP_RCVD_UNTRUST,KHOP_THREADED,RCVD_IN_DNSWL_HI,RCVD_IN_HOSTKARMA_W,RP_MATCHES_RCVD,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 25 Oct 2012 17:42:15 +0000 Received: from int-mx12.intmail.prod.int.phx2.redhat.com (int-mx12.intmail.prod.int.phx2.redhat.com [10.5.11.25]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id q9PHgEjr026940 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 25 Oct 2012 13:42:14 -0400 Received: from brno.lan (ovpn01.gateway.prod.ext.ams2.redhat.com [10.39.146.11]) by int-mx12.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id q9PHgDI9014097 for ; Thu, 25 Oct 2012 13:42:13 -0400 Subject: [PATCH 3/8] foll-vfork.exp: Expose more testing on non-HP/UX targets To: gdb-patches@sourceware.org From: Pedro Alves Date: Thu, 25 Oct 2012 17:42:00 -0000 Message-ID: <20121025174212.13324.41225.stgit@brno.lan> In-Reply-To: <20121025174002.13324.8164.stgit@brno.lan> References: <20121025174002.13324.8164.stgit@brno.lan> User-Agent: StGit/0.16 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2012-10/txt/msg00507.txt.bz2 Well, exposes more to Linux, since we have at the top: if {![istarget "hppa*-hp-hpux*"] && ![istarget "*-linux*"]} then { continue } The vfork_and_exec_child_follow_through_step procedure is about following the child over a "next". This is currently disabled on non HP/UX targets, but it's a useful test to run. AFAICS from the test, HP/UX can't debug the vfork child until it execs or exits. The Linux kernel can, so the test is adjusted to expect the ideal behavior out of non HP/UX targets (and we get that right on Linux). gdb/testsuite/ 2012-10-25 Pedro Alves * gdb.base/foll-vfork.exp (vfork_and_exec_child_follow_through_step): Don't skip on non-HP/UX targets. Expect the next to only step one line on non-HP/UX targets, rather than stopping only after the exec. --- gdb/testsuite/gdb.base/foll-vfork.exp | 47 +++++++++++++++++++++++---------- 1 file changed, 33 insertions(+), 14 deletions(-) diff --git a/gdb/testsuite/gdb.base/foll-vfork.exp b/gdb/testsuite/gdb.base/foll-vfork.exp index 67d077d..54f5971 100644 --- a/gdb/testsuite/gdb.base/foll-vfork.exp +++ b/gdb/testsuite/gdb.base/foll-vfork.exp @@ -183,15 +183,17 @@ proc vfork_and_exec_child_follow_through_step {} { global gdb_prompt global srcfile2 -# This test cannot be performed prior to HP-UX 10.30, because ptrace-based -# debugging of a vforking program basically doesn't allow the child to do -# things like hit a breakpoint between a vfork and exec. This means that -# saying "set follow-fork child; next" at a vfork() call won't work, because -# the implementation of "next" sets a "step resume" breakpoint at the -# return from the vfork(), which the child will hit on its way to exec'ing. -# - if { ![istarget "hppa*-*-hpux11.*"] } { - verbose "vfork child-following next test ignored for non-hppa or pre-HP/UX-10.30 targets." + if { [istarget "hppa*-*-hpux*"] && ![istarget "hppa*-*-hpux11.*"] } { + # This test cannot be performed prior to HP-UX 10.30, because + # ptrace-based debugging of a vforking program basically doesn't + # allow the child to do things like hit a breakpoint between a + # vfork and exec. This means that saying "set follow-fork + # child; next" at a vfork() call won't work, because the + # implementation of "next" sets a "step resume" breakpoint at + # the return from the vfork(), which the child will hit on its + # way to exec'ing. + # + verbose "vfork child-following next test ignored for pre-HP/UX-10.30 targets." return 0 } @@ -200,11 +202,28 @@ proc vfork_and_exec_child_follow_through_step {} { "set follow-fork child, vfork and exec through step" set test "vfork and exec child follow, through step" - set linenum [gdb_get_line_number "printf(\"Hello from vforked-prog" ${srcfile2}] - gdb_test_multiple "next" $test { - -re "Attaching after fork to.*Executing new program.*Breakpoint.*vforked-prog.c:${linenum}.*$gdb_prompt " { - pass "$test" - } + if { [istarget "hppa*-*-hpux*"]} { + # Since the child cannot be debugged until after it has exec'd, + # and since there's a bp on "main" in the parent, and since the + # bp's for the parent are recomputed in the exec'd child, the + # step through a vfork should land us in the "main" for the + # exec'd child, too. + # + set linenum [gdb_get_line_number "printf(\"Hello from vforked-prog" ${srcfile2}] + gdb_test_multiple "next" $test { + -re "Attaching after fork to.*Executing new program.*Breakpoint.*vforked-prog.c:${linenum}.*$gdb_prompt " { + pass "$test" + } + } + } else { + # The ideal support is to be able to debug the child even + # before it execs. Thus, "next" lands on the next line after + # the vfork. + gdb_test_multiple "next" $test { + -re "Attaching after .* vfork to child.*if \\(pid == 0\\).*$gdb_prompt " { + pass "$test" + } + } } # The parent has been detached; allow time for any output it might # generate to arrive, so that output doesn't get confused with