From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 77959 invoked by alias); 18 Mar 2015 09:02:42 -0000 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 Received: (qmail 77950 invoked by uid 89); 18 Mar 2015 09:02:41 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.8 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_LOW autolearn=ham version=3.3.2 X-HELO: smtp.eu.adacore.com Received: from mel.act-europe.fr (HELO smtp.eu.adacore.com) (194.98.77.210) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-GCM-SHA384 encrypted) ESMTPS; Wed, 18 Mar 2015 09:02:35 +0000 Received: from localhost (localhost [127.0.0.1]) by filtered-smtp.eu.adacore.com (Postfix) with ESMTP id C705F2D053D0 for ; Wed, 18 Mar 2015 10:02:32 +0100 (CET) Received: from smtp.eu.adacore.com ([127.0.0.1]) by localhost (smtp.eu.adacore.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id krV9bb7ZlqrQ for ; Wed, 18 Mar 2015 10:02:32 +0100 (CET) Received: from ulanbator.act-europe.fr (ulanbator.act-europe.fr [10.10.1.67]) (using TLSv1 with cipher ECDHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by smtp.eu.adacore.com (Postfix) with ESMTPSA id B7F5C2D0527E for ; Wed, 18 Mar 2015 10:02:32 +0100 (CET) Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2070.6\)) Subject: Ping: [Patch] Fix windows 64 unwinding issues From: Tristan Gingold In-Reply-To: <2D4C0DC1-2CE7-4C9B-9CA1-1BC13B0FC9E1@adacore.com> Date: Wed, 18 Mar 2015 09:02:00 -0000 Content-Transfer-Encoding: quoted-printable Message-Id: <5EE73165-87E2-467B-B165-9A187A9DA885@adacore.com> References: <2D4C0DC1-2CE7-4C9B-9CA1-1BC13B0FC9E1@adacore.com> To: " ml" X-IsSubscribed: yes X-SW-Source: 2015-03/txt/msg00535.txt.bz2 Apparently this patch should be reviewed by a global maintainer as there is no platform maintainer for Windows. Tristan. > On 05 Mar 2015, at 14:42, Tristan Gingold wrote: >=20 > Hello, >=20 > yet another patch to fix incorrect unwinding in system dlls. Was simply = manually tested. >=20 > Ok to commit ? >=20 > Tristan. >=20 > commit da3b5213dc072fca195451a04f35a2eb6342bb62 > Author: Tristan Gingold > Date: Thu Mar 5 14:36:32 2015 +0100 >=20 > Fix amd64 windows unwinding issues within MS dlls. >=20 > Unwind info in system dlls uses almost all possible codes, contrary to= unwind > info generated by gcc. A few issues have been discovered: incorrect h= andling > of SAVE_NONVOL opcodes and incorrect in prologue range checks. Furthe= rmore I > added comments not to forget what has been investigated. >=20 > gdb/ChangeLog: > * amd64-windows-tdep.c (amd64_windows_find_unwind_info): Move > redirection code to ... > (amd64_windows_frame_decode_insns): ... Here. Fix in prologue > checks. Fix SAVE_NONVOL operations. Add debug code and comments. >=20 > diff --git a/gdb/amd64-windows-tdep.c b/gdb/amd64-windows-tdep.c > index 2aa10a1..9278a26 100644 > --- a/gdb/amd64-windows-tdep.c > +++ b/gdb/amd64-windows-tdep.c > @@ -621,9 +621,47 @@ amd64_windows_frame_decode_insns (struct frame_info = *this_frame, > CORE_ADDR cur_sp =3D cache->sp; > struct gdbarch *gdbarch =3D get_frame_arch (this_frame); > enum bfd_endian byte_order =3D gdbarch_byte_order (gdbarch); > - int j; > + int first =3D 1; > + > + /* There are at least 3 possibilities to share an unwind info entry: > + 1. Two different runtime_function entries (in .pdata) can point to = the > + same unwind info entry. There is no such indication while unwin= ding, > + so we don't really care about that case. We suppose this scheme is > + used to save memory when the unwind entries are exactly the same. > + 2. Chained unwind_info entries, with no unwind codes (no prologue). > + There is a major difference with the previous case: the pc range= for > + the function is different (in case 1, the pc range comes from the > + runtime_function entry; in case 2, the pc range for the chained entry > + comes from the first unwind entry). Case 1 cannot be used instead as > + the pc is not in the prologue. This case is officially documented. > + (There might be unwind code in the first unwind entry to handle > + additionnal unwinding). GCC (at least until gcc 5.0) doesn't chain > + entries. > + 3. Undocumented unwind info redirection. Hard to know the exact pu= rpose, > + so it is considered as a memory optimization of case 2. > + */ >=20 > - for (j =3D 0; ; j++) > + if (unwind_info & 1) > + { > + /* Unofficially documented unwind info redirection, when UNWIND_IN= FO > + address is odd (http://www.codemachine.com/article_x64deepdive.html). > + */ > + struct external_pex64_runtime_function d; > + CORE_ADDR sa, ea; > + > + if (target_read_memory (cache->image_base + (unwind_info & ~1), > + (gdb_byte *) &d, sizeof (d)) !=3D 0) > + return; > + > + cache->start_rva =3D > + extract_unsigned_integer (d.rva_BeginAddress, 4, byte_order); > + cache->end_rva =3D > + extract_unsigned_integer (d.rva_EndAddress, 4, byte_order); > + unwind_info =3D > + extract_unsigned_integer (d.rva_UnwindData, 4, byte_order); > + } > + > + while (1) > { > struct external_pex64_unwind_info ex_ui; > /* There are at most 256 16-bit unwind insns. */ > @@ -633,6 +671,7 @@ amd64_windows_frame_decode_insns (struct frame_info *= this_frame, > unsigned char codes_count; > unsigned char frame_reg; > unsigned char frame_off; > + CORE_ADDR start; >=20 > /* Read and decode header. */ > if (target_read_memory (cache->image_base + unwind_info, > @@ -653,12 +692,13 @@ amd64_windows_frame_decode_insns (struct frame_info= *this_frame, > && PEX64_UWI_VERSION (ex_ui.Version_Flags) !=3D 2) > return; >=20 > - if (j =3D=3D 0 > - && (cache->pc >=3D > - cache->image_base + cache->start_rva + ex_ui.SizeOfPrologue)) > + start =3D cache->image_base + cache->start_rva; > + if (first > + && !(cache->pc >=3D start && cache->pc < start + ex_ui.SizeOfPrologue= )) > { > - /* Not in the prologue. We want to detect if the PC points to an > - epilogue. If so, the epilogue detection+decoding function is > + /* We want to detect if the PC points to an epilogue. This needs > + to be checked only once, and an epilogue can be anywhere but in > + the prologue. If so, the epilogue detection+decoding function is > sufficient. Otherwise, the unwinder will consider that the PC > is in the body of the function and will need to decode unwind > info. */ > @@ -711,19 +751,24 @@ amd64_windows_frame_decode_insns (struct frame_info= *this_frame, > { > int reg; >=20 > - if (frame_debug) > - fprintf_unfiltered > - (gdb_stdlog, " op #%u: off=3D0x%02x, insn=3D0x%02x\n", > - (unsigned) (p - insns), p[0], p[1]); > - > - /* Virtually execute the operation. */ > - if (cache->pc >=3D cache->image_base + cache->start_rva + p[0]) > + /* Virtually execute the operation if the pc is after the > + corresponding instruction (that does matter in case of break > + within the prologue). Note that for chained info (!first), the > + prologue has been fully executed. */ > + if (cache->pc >=3D start + p[0] || cache->pc < start) > { > + if (frame_debug) > + fprintf_unfiltered > + (gdb_stdlog, " op #%u: off=3D0x%02x, insn=3D0x%02x\n", > + (unsigned) (p - insns), p[0], p[1]); > + > /* If there is no frame registers defined, the current value of > rsp is used instead. */ > if (frame_reg =3D=3D 0) > save_addr =3D cur_sp; >=20 > + reg =3D -1; > + > switch (PEX64_UNWCODE_CODE (p[1])) > { > case UWOP_PUSH_NONVOL: > @@ -751,12 +796,12 @@ amd64_windows_frame_decode_insns (struct frame_info= *this_frame, > case UWOP_SAVE_NONVOL: > reg =3D amd64_windows_w2gdb_regnum[PEX64_UNWCODE_INFO (p[1])]; > cache->prev_reg_addr[reg] =3D save_addr > - - 8 * extract_unsigned_integer (p + 2, 2, byte_order); > + + 8 * extract_unsigned_integer (p + 2, 2, byte_order); > break; > case UWOP_SAVE_NONVOL_FAR: > reg =3D amd64_windows_w2gdb_regnum[PEX64_UNWCODE_INFO (p[1])]; > cache->prev_reg_addr[reg] =3D save_addr > - - 8 * extract_unsigned_integer (p + 2, 4, byte_order); > + + 8 * extract_unsigned_integer (p + 2, 4, byte_order); > break; > case UWOP_SAVE_XMM128: > cache->prev_xmm_addr[PEX64_UNWCODE_INFO (p[1])] =3D > @@ -787,6 +832,13 @@ amd64_windows_frame_decode_insns (struct frame_info = *this_frame, > default: > return; > } > + > + /* Display address where the register was saved. */ > + if (frame_debug && reg >=3D 0) > + fprintf_unfiltered > + (gdb_stdlog, " [reg %s at %s]\n", > + gdbarch_register_name (gdbarch, reg), > + paddress (gdbarch, cache->prev_reg_addr[reg])); > } >=20 > /* Adjust with the length of the opcode. */ > @@ -818,19 +870,29 @@ amd64_windows_frame_decode_insns (struct frame_info= *this_frame, > } > } > if (PEX64_UWI_FLAGS (ex_ui.Version_Flags) !=3D UNW_FLAG_CHAININFO) > - break; > + { > + /* End of unwind info. */ > + break; > + } > else > { > /* Read the chained unwind info. */ > struct external_pex64_runtime_function d; > CORE_ADDR chain_vma; >=20 > + /* Not anymore the first entry. */ > + first =3D 0; > + > + /* Stay aligned on word boundary. */ > chain_vma =3D cache->image_base + unwind_info > + sizeof (ex_ui) + ((codes_count + 1) & ~1) * 2; >=20 > if (target_read_memory (chain_vma, (gdb_byte *) &d, sizeof (d)) !=3D 0) > return; >=20 > + /* Decode begin/end. This may be different from .pdata index, as > + an unwind info may be shared by several functions (in particular > + if many functions have the same prolog and handler. */ > cache->start_rva =3D > extract_unsigned_integer (d.rva_BeginAddress, 4, byte_order); > cache->end_rva =3D > @@ -940,25 +1002,6 @@ amd64_windows_find_unwind_info (struct gdbarch *gdb= arch, CORE_ADDR pc, > "amd64_windows_find_unwind_data: image_base=3D%s, unwind_data=3D%= s\n", > paddress (gdbarch, base), paddress (gdbarch, *unwind_info)); >=20 > - if (*unwind_info & 1) > - { > - /* Unofficially documented unwind info redirection, when UNWIND_IN= FO > - address is odd (http://www.codemachine.com/article_x64deepdive.html). > - */ > - struct external_pex64_runtime_function d; > - CORE_ADDR sa, ea; > - > - if (target_read_memory (base + (*unwind_info & ~1), > - (gdb_byte *) &d, sizeof (d)) !=3D 0) > - return -1; > - > - *start_rva =3D > - extract_unsigned_integer (d.rva_BeginAddress, 4, byte_order); > - *end_rva =3D extract_unsigned_integer (d.rva_EndAddress, 4, byte_o= rder); > - *unwind_info =3D > - extract_unsigned_integer (d.rva_UnwindData, 4, byte_order); > - > - } > return 0; > } >=20 >=20