From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4556 invoked by alias); 25 Feb 2003 23:53:58 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 4547 invoked from network); 25 Feb 2003 23:53:58 -0000 Received: from unknown (HELO localhost.redhat.com) (172.16.49.200) by 172.16.49.205 with SMTP; 25 Feb 2003 23:53:58 -0000 Received: from redhat.com (localhost [127.0.0.1]) by localhost.redhat.com (Postfix) with ESMTP id 32CB52A9C for ; Tue, 25 Feb 2003 18:56:14 -0500 (EST) Message-ID: <3E5C029E.4070009@redhat.com> Date: Tue, 25 Feb 2003 23:53:00 -0000 From: Andrew Cagney User-Agent: Mozilla/5.0 (X11; U; NetBSD macppc; en-US; rv:1.0.2) Gecko/20030223 X-Accept-Language: en-us, en MIME-Version: 1.0 To: gdb-patches@sources.redhat.com Subject: Re: [patch/rfc] strengthn get_prev_frame() checks References: <3E5AE9C9.70801@mac.com> Content-Type: multipart/mixed; boundary="------------020308070403020700090405" X-SW-Source: 2003-02/txt/msg00677.txt.bz2 This is a multi-part message in MIME format. --------------020308070403020700090405 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-length: 338 > Hello, > > This improves the checks in get_prev_frame() that look for stuff like the top-of-stack or a corrupt stack. > > d10v (which uses this) showed no regressions, neither did i386. > > I'll commit `tomorrow'. > > Andrew The attached is what I actually checked in. It adds some debug messages to the previous patch. Andrew --------------020308070403020700090405 Content-Type: text/plain; name="diffs" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="diffs" Content-length: 3916 2003-02-25 Andrew Cagney * frame.c (get_prev_frame): Add comment on check for inside_entry_func. Only check for inside_entry_file when not a dummy and not a sentinel. Check that the new frame is not inner to the old frame. Index: frame.c =================================================================== RCS file: /cvs/src/src/gdb/frame.c,v retrieving revision 1.68 diff -u -r1.68 frame.c --- frame.c 25 Feb 2003 23:12:22 -0000 1.68 +++ frame.c 25 Feb 2003 23:50:52 -0000 @@ -1239,7 +1239,10 @@ return next_frame->prev; next_frame->prev_p = 1; - /* If we're inside the entry file, it isn't valid. */ + /* If we're inside the entry file, it isn't valid. Don't apply this + test to a dummy frame - dummy frame PC's typically land in the + entry file. Don't apply this test to the sentinel frame. + Sentinel frames should always be allowed to unwind. */ /* NOTE: drow/2002-12-25: should there be a way to disable this check? It assumes a single small entry file, and the way some debug readers (e.g. dbxread) figure out which object is the @@ -1247,7 +1250,8 @@ /* NOTE: cagney/2003-01-10: If there is a way of disabling this test then it should probably be moved to before the ->prev_p test, above. */ - if (inside_entry_file (get_frame_pc (next_frame))) + if (next_frame->type != DUMMY_FRAME && next_frame->level >= 0 + && inside_entry_file (get_frame_pc (next_frame))) { if (frame_debug) fprintf_unfiltered (gdb_stdlog, @@ -1255,6 +1259,23 @@ return NULL; } + /* If we're already inside the entry function for the main objfile, + then it isn't valid. Don't apply this test to a dummy frame - + dummy frame PC's typically land in the entry func. Don't apply + this test to the sentinel frame. Sentinel frames should always + be allowed to unwind. */ + /* NOTE: cagney/2003-02-25: Don't enable until someone has found + hard evidence that this is needed. */ + if (0 + && next_frame->type != DUMMY_FRAME && next_frame->level >= 0 + && inside_entry_func (get_frame_pc (next_frame))) + { + if (frame_debug) + fprintf_unfiltered (gdb_stdlog, + "Outermost frame - inside entry func\n"); + return NULL; + } + /* If any of the old frame initialization methods are around, use the legacy get_prev_frame method. Just don't try to unwind a sentinel frame using that method - it doesn't work. All sentinal @@ -1324,6 +1345,9 @@ /* FIXME: cagney/2002-12-18: Instead of this hack, should just save the frame ID directly. */ struct frame_id id = frame_id_unwind (next_frame); + /* Check that the unwound ID is valid. As of 2003-02-24 the + x86-64 was returning an invalid frame ID when trying to do an + unwind a sentinel frame that belonged to a frame dummy. */ if (!frame_id_p (id)) { if (frame_debug) @@ -1331,6 +1355,20 @@ "Outermost frame - unwound frame ID invalid\n"); return NULL; } + /* Check that the new frame isn't inner to (younger, below, next) + the old frame. If that happens the frame unwind is going + backwards. */ + /* FIXME: cagney/2003-02-25: Ignore the sentinel frame since that + doesn't have a valid frame ID. Should instead set the sentinel + frame's frame ID to a `sentinel'. Leave it until after the + switch to storing the frame ID, instead of the frame base, in + the frame object. */ + if (next_frame->level >= 0 + && frame_id_inner (id, get_frame_id (next_frame))) + error ("Unwound frame inner-to selected frame (corrupt stack?)"); + /* Note that, due to frameless functions, the stronger test of the + new frame being outer to the old frame can't be used - + frameless functions differ by only their PC value. */ prev_frame->frame = id.base; } --------------020308070403020700090405--