Date: Sun, 7 Feb 2016 13:32:36 -0800 From: Mark Millard <markmi@dsl-only.net> To: FreeBSD PowerPC ML <freebsd-ppc@freebsd.org>, FreeBSD Toolchain <freebsd-toolchain@freebsd.org> Cc: Roman Divacky <rdivacky@vlakno.cz>, Justin Hibbits <chmeeedalf@gmail.com>, Konstantin Belousov <kib@freebsd.org>, Nathan Whitehorn <nwhitehorn@freebsd.org> Subject: Re: powerpc (32-bit) clang 3.8.0 vs. gcc 4.2.1 routine preamble mismatches: contributions to SEGV's differences Message-ID: <29B25F29-17DC-4F85-8C5A-3F286B260E29@dsl-only.net> In-Reply-To: <D985966A-2997-4A5D-955E-222D5F29645B@dsl-only.net> References: <3D08EB58-7FEF-432E-8192-77F988A75621@dsl-only.net> <D985966A-2997-4A5D-955E-222D5F29645B@dsl-only.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2016-Feb-6, at 9:58 PM, Mark Millard <markmi at dsl-only.net> wrote: >=20 > I've submitted bug 206990 ( = https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D206990 ) with a = proof-of-concept patch for avoiding signal delivery mixed with clang = 3.8.0 generated code causing SEGV's --a patch that has passed my = personal testing so far. "make -j 6 buildworld" finished normally = instead of getting a SEGV in a few minutes on a = dual-processor/each-being-dual-core G5. >=20 > Now a "make -j 3 buildworld" on a dual processor G4 is in process, = booted from the same SSD. We will see. >=20 > The official TARGET_ARCH=3Dpowerpc sendsig code could tromp on the = frame pointer stored at "-4(r1)" (as seen in the clang3.8.0-generated = code) during the period in which the frame pointer is outside the range = identified by the stack pointer (r1) and where the stack started. The = change respects a Darwin-like/AIX-like "red-zone"/scratch area on the = smaller-address side of the stack. This should still be compatible with = gcc 4.2.1 style code, although it "wastes" more bytes temporarily in = that context. >=20 >=20 > =3D=3D=3D > Mark Millard > markmi at dsl-only.net Nathan Whitehorn had me submit to llvm.org. So there now is: Bug 26519 ( https://llvm.org/bugs/show_bug.cgi?id=3D26519 ) - Clang = 3.8.0's "Target: powerpc-unknown-freebsd11.0" code generation is = violating the ABI involved I left the "importance" at "normal" but if it still makes sense it may = be that "release blocker" would be appropriate. (Can 32-bit powerpc = block clang releases? Is it "too late now, already released"?) =3D=3D=3D Mark Millard markmi at dsl-only.net On 2016-Feb-5, at 1:59 AM, Mark Millard <markmi at dsl-only.net> wrote: >=20 > Clang 3.8.0 produced code uses r31 as a frame pointer in contexts = where gcc 4.2.1 produced code does not (ever?). This leaves clang's = produced code being more dependent on r31 handling, such as when = resuming after signal delivery. >=20 > The following is one of the routines in "make" where a clang 3.8.0 = based "make" sometimes gets a SEGV after resuming after a SIGCHLD = delivery, the SEGV being from having r31=3D0x0 in a Frame Pointer (r31) = based address calculation that is at some point dereferenced. (See = https://lists.freebsd.org/pipermail/freebsd-ppc/2016-February/008002.html = .) >=20 > But gcc 4.2.1 does not use r31 as a frame pointer in the Str_Match = that it produces and so does not see the problem. gcc 4.2.1's produced = code simply uses the stack pointer as needed. >=20 >=20 > clang 3.8.0 based Str_Match preamble (from make): >=20 > 0x181a4a8 <Str_Match>: mflr r0 > 0x181a4ac <Str_Match+4>: stw r31,-4(r1) # Clang's frame = pointer (r31)=20 > # saved before stack = pointer changed. > 0x181a4b0 <Str_Match+8>: stw r0,4(r1) # lr saved before = stack pointer changed. > 0x181a4b4 <Str_Match+12>: stwu r1,-32(r1) # Stack pointer = finally saved and > # changed. > 0x181a4b8 <Str_Match+16>: mr r31,r1 # r31 is the frame = pointer under clang. > 0x181a4bc <Str_Match+20>: stw r30,24(r31) >=20 > gcc 4.2.1 based Str_Match preamble: >=20 > 0x1819cb8 <Str_Match>: mflr r0 > 0x1819cbc <Str_Match+4>: stwu r1,-32(r1) # Stack pointer saved = and changed first. > 0x1819cc0 <Str_Match+8>: stw r31,28(r1) # r31 saved after = stack pointer changed. > 0x1819cc4 <Str_Match+12>: mr r31,r3 # gcc 4.2.1 does not = reserve > # r31 for use as a = frame pointer. > 0x1819cc8 <Str_Match+16>: stw r30,24(r1) > 0x1819ccc <Str_Match+20>: stw r0,36(r1) # lr saved after = stack pointer changed. >=20 >=20 > (Str_Match is a self contained routine, although it is recursive.) >=20 >=20 > Looking at some other gcc 4.2.1 preamble examples. . . >=20 > 0x1823b58 <VarSYSVMatch>: cmpwi cr7,r6,0 > 0x1823b5c <VarSYSVMatch+4>: stwu r1,-64(r1) # Stack pointer saved = and changed "first" > 0x1823b60 <VarSYSVMatch+8>: mflr r0 > 0x1823b64 <VarSYSVMatch+12>: lis r9,396 > 0x1823b68 <VarSYSVMatch+16>: stw r25,36(r1) > 0x1823b6c <VarSYSVMatch+20>: addi r25,r9,8944 > 0x1823b70 <VarSYSVMatch+24>: stw r26,40(r1) > 0x1823b74 <VarSYSVMatch+28>: mr r26,r3 > 0x1823b78 <VarSYSVMatch+32>: stw r27,44(r1) > 0x1823b7c <VarSYSVMatch+36>: mr r27,r4 > 0x1823b80 <VarSYSVMatch+40>: stw r28,48(r1) > 0x1823b84 <VarSYSVMatch+44>: mr r28,r8 > 0x1823b88 <VarSYSVMatch+48>: stw r29,52(r1) > 0x1823b8c <VarSYSVMatch+52>: mr r29,r5 > 0x1823b90 <VarSYSVMatch+56>: stw r31,60(r1) > 0x1823b94 <VarSYSVMatch+60>: mr r31,r7 # Again r31 is not a = frame pointer > 0x1823b98 <VarSYSVMatch+64>: stw r0,68(r1) > 0x1823b9c <VarSYSVMatch+68>: lwz r0,0(r25) > 0x1823ba0 <VarSYSVMatch+72>: stw r0,28(r1) > 0x1823ba4 <VarSYSVMatch+76>: li r0,0 > 0x1823ba8 <VarSYSVMatch+80>: stw r30,56(r1) > 0x1823bac <VarSYSVMatch+84>: beq- cr7,0x1823bbc <VarSYSVMatch+100> >=20 >=20 > 0x1819f30 <Str_SYSVMatch>: mflr r0 # Stack pointer saved = and changed first > 0x1819f34 <Str_SYSVMatch+4>: stwu r1,-32(r1) > 0x1819f38 <Str_SYSVMatch+8>: stw r28,16(r1) > 0x1819f3c <Str_SYSVMatch+12>: mr r28,r5 > 0x1819f40 <Str_SYSVMatch+16>: stw r30,24(r1) > 0x1819f44 <Str_SYSVMatch+20>: mr r30,r3 > 0x1819f48 <Str_SYSVMatch+24>: stw r31,28(r1) > 0x1819f4c <Str_SYSVMatch+28>: mr r31,r4 # Again r31 is not a = frame pointer > 0x1819f50 <Str_SYSVMatch+32>: stw r29,20(r1) > 0x1819f54 <Str_SYSVMatch+36>: stw r0,36(r1) > 0x1819f58 <Str_SYSVMatch+40>: lbz r29,0(r4) >=20 >=20 > 0x181fcac <VarMatch>: mflr r0 # Stack pointer saved = and changed first > 0x181fcb0 <VarMatch+4>: stwu r1,-48(r1) > 0x181fcb4 <VarMatch+8>: lis r9,396 > 0x181fcb8 <VarMatch+12>: stw r27,28(r1) > 0x181fcbc <VarMatch+16>: mr r27,r4 > 0x181fcc0 <VarMatch+20>: stw r0,52(r1) > 0x181fcc4 <VarMatch+24>: stw r28,32(r1) > 0x181fcc8 <VarMatch+28>: mr r28,r7 > 0x181fccc <VarMatch+32>: lwz r0,-1344(r9) > 0x181fcd0 <VarMatch+36>: stw r29,36(r1) > 0x181fcd4 <VarMatch+40>: mr r29,r5 > 0x181fcd8 <VarMatch+44>: andi. r9,r0,512 > 0x181fcdc <VarMatch+48>: stw r30,40(r1) > 0x181fce0 <VarMatch+52>: stw r31,44(r1) > 0x181fce4 <VarMatch+56>: mr r30,r8 > 0x181fce8 <VarMatch+60>: mr r31,r6 # Again r31 is not a = frame pointer >=20 >=20 > 0x1801d58 <Buf_AddBytes>: mflr r0 # Stack pointer saved = and changed first > 0x1801d5c <Buf_AddBytes+4>: stwu r1,-48(r1) > 0x1801d60 <Buf_AddBytes+8>: stw r28,32(r1) > 0x1801d64 <Buf_AddBytes+12>: stw r0,52(r1) > 0x1801d68 <Buf_AddBytes+16>: stw r30,40(r1) > 0x1801d6c <Buf_AddBytes+20>: mr r30,r4 > 0x1801d70 <Buf_AddBytes+24>: lwz r28,4(r3) > 0x1801d74 <Buf_AddBytes+28>: lwz r4,0(r3) > 0x1801d78 <Buf_AddBytes+32>: stw r29,36(r1) > 0x1801d7c <Buf_AddBytes+36>: add r29,r30,r28 > 0x1801d80 <Buf_AddBytes+40>: cmpw cr7,r29,r4 > 0x1801d84 <Buf_AddBytes+44>: stw r27,28(r1) > 0x1801d88 <Buf_AddBytes+48>: stw r31,44(r1) > 0x1801d8c <Buf_AddBytes+52>: mr r27,r5 > 0x1801d90 <Buf_AddBytes+56>: mr r31,r3 # Again r31 is not a = frame pointer >=20 >=20 > And so it goes for every intermittent SEGV related example (clang = 3.8.0 buildworld based) that I've examined: the matching gcc 4.2.1 code = would not try to use the the r31 values that clang does use. Instead gcc = 4.2.1 assigns an independent value to r31 before using it. >=20 >=20 > In effect gcc 4.2.1 and clang 3.8.0 are not following the exact-same = call standard. If clang 3.8.0's code generation is left as is then a = conversion to its call standard requirements will be required if clang = 3.8.0 is to be used for powerpc (32-bit). >=20 > "Works when gcc 4.2.1 is used" is not a great guide to "appropriate = for use with clang 3.8.0", at least in this area for powerpc (32-bit). >=20 > (These notes presume a context with sys/powerpc/powerpc/sigcode32.S = -r295186 in place so that signal delivery maintains the modulo 16 byte = stack/frame alignment status instead of changing the alignment. It = appears that, while necessary, this is not sufficient for a clang 3.8.0 = based buildworld to operate with signals reliably. See = https://lists.freebsd.org/pipermail/freebsd-ppc/2016-February/008002.html = .) >=20 > =3D=3D=3D Mark Millard markmi at dsl-only.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?29B25F29-17DC-4F85-8C5A-3F286B260E29>