From patchwork Thu Feb 2 16:46:23 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Pedro Alves X-Patchwork-Id: 19094 Received: (qmail 95409 invoked by alias); 2 Feb 2017 16:46:36 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Delivered-To: mailing list gdb-patches@sourceware.org Received: (qmail 95379 invoked by uid 89); 2 Feb 2017 16:46:36 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.9 required=5.0 tests=BAYES_00, RP_MATCHES_RCVD, SPF_HELO_PASS autolearn=ham version=3.3.2 spammy=UD:L, cris, 102, 12, 10212 X-HELO: mx1.redhat.com Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 02 Feb 2017 16:46:26 +0000 Received: from int-mx11.intmail.prod.int.phx2.redhat.com (int-mx11.intmail.prod.int.phx2.redhat.com [10.5.11.24]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id E8CEA3A76B9; Thu, 2 Feb 2017 16:46:25 +0000 (UTC) Received: from [127.0.0.1] (ovpn04.gateway.prod.ext.phx2.redhat.com [10.5.9.4]) by int-mx11.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id v12GkONv011873; Thu, 2 Feb 2017 11:46:25 -0500 Subject: Re: [PATCH 4/6] Disassembly unit test: disassemble one instruction To: Yao Qi References: <1484051178-16013-1-git-send-email-yao.qi@linaro.org> <1484560977-8693-1-git-send-email-yao.qi@linaro.org> <1484560977-8693-5-git-send-email-yao.qi@linaro.org> <620db0ef-5e3d-3b93-5596-33d24a78f6a9@redhat.com> <20170124152326.GR28060@E107787-LIN> Cc: gdb-patches@sourceware.org From: Pedro Alves Message-ID: <83019a6f-ad2e-0ecf-0c07-8fc479e46b7b@redhat.com> Date: Thu, 2 Feb 2017 16:46:23 +0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0 MIME-Version: 1.0 In-Reply-To: <20170124152326.GR28060@E107787-LIN> On 01/24/2017 03:23 PM, Yao Qi wrote: > On 17-01-20 00:03:48, Pedro Alves wrote: >>> + /* Test gdb_disassembler for a given gdbarch by reading data from a >>> + pre-allocated buffer. If you want to see the disassembled >>> + instruction printed to gdb_stdout, set DISASSEMBLER_TEST_VERBOSE >>> + to true. */ >>> + >>> + class gdb_disassembler_test : public gdb_disassembler >>> + { >>> + public: >>> + >>> + const bool DISASSEMBLER_TEST_VERBOSE = false; >> >> static. We give macros long unique names in order to >> avoid naming conflicts, but if this is no longer a macro, >> the name could be shortened, to e.g., just: >> >> static const bool verbose = false; >> > > gdb_disassembler_test is a local class, so it can't have a static data > member. The end result is broken then, unfortunately. I didn't realize it either until I saw a spurious fail in the gdb.base/unittest.exp test today. Running gdb -ex "maintenance selftest" manually I see: ~~~ Type "apropos word" to search for commands related to "word". warning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default ARC600 settings. .long 0x256f003fwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default ARC600 settings. .long 0x256f003fwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default A6 settings. .long 0x256f003fwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default ARC601 settings. warning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default ARC700 settings. brkwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default A7 settings. brkwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default ARCv2 settings. brkwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default EM settings. brkwarning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default HS settings. brkmov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0mov r0, #0breakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakbreakM3.L = 0xffff;/* ( -1) M3=0x0xffff(65535) */break 8break 8warning: A handler for the OS ABI "GNU/Linux" is not built into this configuration of GDB. Attempting to continue with the default cris:common_v10_v32 settings. ~~~ etc. (this is what shows up in the gdb.log too). Note the odd bits of instructions printed. They appear because here: class gdb_disassembler_test : public gdb_disassembler { public: const bool verbose = false; explicit gdb_disassembler_test (struct gdbarch *gdbarch, const gdb_byte *insn, size_t len) : gdb_disassembler (gdbarch, (verbose ? gdb_stdout : &null_stream), gdb_disassembler_test::read_memory), specifically in this line: (verbose ? gdb_stdout : &null_stream), "verbose" has not been initialized yet, because the order of initialization is always base classes first. I.e. "verbose" is only initialized after the base constructor is called. Since the gdb_disassembler_test object is created on the stack, "verbose" has garbage at that point, and thus we end up with the stream incorrectly pointing to gdb_stdout. I was surprised that GCC doesn't warn about this. I filed a GCC bug here: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79345 I'm not sure whether this was the reason for the unittest.exp random failure, but I suspect so. (I failed to save the log, and also forgot to check whether all selftests had passed.) The patchlet below fixes the spurious insn printing. OK? diff --git a/gdb/disasm-selftests.c b/gdb/disasm-selftests.c index 7d0b006..9eb80b4 100644 --- a/gdb/disasm-selftests.c +++ b/gdb/disasm-selftests.c @@ -102,13 +102,12 @@ print_one_insn_test (struct gdbarch *gdbarch) /* Test gdb_disassembler for a given gdbarch by reading data from a pre-allocated buffer. If you want to see the disassembled instruction printed to gdb_stdout, set verbose to true. */ + static const bool verbose = false; class gdb_disassembler_test : public gdb_disassembler { public: - const bool verbose = false; - explicit gdb_disassembler_test (struct gdbarch *gdbarch, const gdb_byte *insn, size_t len)