[Bug c/11627] New: internal compiler error: in push_reload, at reload.c:1282

[Bug c/11627] New: internal compiler error: in push_reload, at reload.c:1282

[dean-gcc at arctic dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627

           Summary: internal compiler error: in push_reload, at
                    reload.c:1282
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dean-gcc at arctic dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu

i got this with top of tree (and also with 3.3).

the code being compiled is a fairly complex loop involving a mix of sse2 
intrinsics and integer operations.

-dean


$ /home/dean/gcc/bin/gcc -v -save-temps -std=c99 -O -g -Wall -march=pentium4 -
fomit-frame-pointer   -c -o sha1.o sha1.c
Reading specs from /home/dean/gcc/lib/gcc-lib/i686-pc-linux-gnu/3.4/specs
Configured with: ../gcc/configure --enable-languages=c --prefix=/home/dean/gcc
Thread model: posix
gcc version 3.4 20030722 (experimental)
 /home/dean/gcc/lib/gcc-lib/i686-pc-linux-gnu/3.4/cc1 -E -quiet -v sha1.c -
march=pentium4 -std=c99 -Wall -fomit-frame-pointer -O -o sha1.i
ignoring nonexistent directory "/home/dean/gcc/i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /usr/local/include
 /home/dean/gcc/include
 /home/dean/gcc/lib/gcc-lib/i686-pc-linux-gnu/3.4/include
 /usr/include
End of search list.
 /home/dean/gcc/lib/gcc-lib/i686-pc-linux-gnu/3.4/cc1 -fpreprocessed sha1.i -
quiet -dumpbase sha1.c -march=pentium4 -auxbase-strip sha1.o -g -O -Wall -
std=c99 -version -fomit-frame-pointer -o sha1.s
GNU C version 3.4 20030722 (experimental) (i686-pc-linux-gnu)
        compiled by GNU C version 3.2.3 (Debian).
GGC heuristics: --param ggc-min-expand=60 --param ggc-min-heapsize=56289
sha1.c: In function `sha1_step':
sha1.c:333: internal compiler error: in push_reload, at reload.c:1282
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.


begin 644 sha1.i.bz2
M0EIH.3%!629365GGNI<`'/9?@'\067___S_O__N_[___8##=Y'1C:&V^O7FZ
M^\KT\'MNGO;/>%\J]7?:Q]<N9H)],/?<^\=S='F]7M>2C8M&O,-VSW&X.G0=
M]V\E8\79FZ.X>2`9A5CQS`IBK85<C/=UY1[W-O=IEXG>[O>/O7+PZ[[N;L4`
MZ=<RK'T\;W>]-V5]==L0EB,)$@AH$!-,DVJ)Z3/3)1ZF@`,C3331IID&U/(@
M)!)-),IE0/R2'I!F4TT>H``Q``````E-J449$-'J:9&0!D`&FAH``````"3U
M4E(U/:IZ4:;(:C0````````!H`"%(D$:F0)JGXDGLDC]4]JGB@]1Z0#3:@`!
MZ@#U'J!2DDT$!,4V1J9"I^T3&I3]*,Q$>D!H#]4Q`T]0]0=]YK;-X:IK8C2I
M4J%EF820%/P/>__(L'K]GC;'9!"9@J&0A\IB<V(E77/6*,,13!,L4J+;9LTR
M9-@IDF99JI,E&1A:L`E5B36(-$H`32+KE7DYQ=S90SR#N!@P=L%1@,&F7#`,
MK+:C",M<W,:3:5+:QRDTLPTR4EKG-:KBJXT1J<KDMJ*S+<E7!:U,U:+3($)@
M$C$,0*,E&(+,43"9#(*"C(R"AL,HY@='IQ:DW399@%)AD1#7_`?UAK<9E(.I
M0+]H?KJ]N24(%MMMK&:BRK(VVZI58V.MS@````'DK.YSKG=W<*U953&[F9F.
MO(W"XWJNUC'J1R1MN&L(9<<F87,<ED6@S#,;M&-VVL:>J$K#(K0I52/,"XLC
MMR*)IK,;I1C&FFFL`4#*LDE"!H%`RN]!K3K+6CH"T%(!0*%!$-.VQHVJC&M%
MJD-MBVB-IAT=L;;:47YPA*W!T(9Y@Z,$@5LJ-\0+[/"53:@>3$)/'[>MDGF@
M>UZ2&K#U:TZ[ZO/SS.%):M[WN.XA]K:N-!>UANXU:];;7-(1_T$@R`*I-)>Q
MKM_X/EG*%6D@]C^#L<CDJAP"9],22IL=8_A!A@8$/^9)DD=399<`.H`.O6UE
MF69KB@#JK,H\Q>@7*,`CJ(EOJ;@`*"N]?M-$4$J):T]F8TWI]T86I36(PP0^
M58A_1S^@.,^2T=R:6C2:6C20B`A!SQZGK9I/0I#;A=LI2YNF4$#!I,'"$TPC
MCGWJY:R0,'+:LDBLCG%JL$D8W6.2%<K)9+P*&V$X!I"3&O%`!%NK4X]B$Y&V
MWSTTVVS7QL^.[N-R:YXD;<<<61=`"TRN#<@C=J@MMR$M5LWOI&W.MEDCAYK&
M[(VYW629I))>`T'F3OGW]';FKR@FNQQ!(KKFC/0E1Q5..K.-[S,[$EE;&F[(
M0@1MJ0=L;DMLK8[;;;2VQN1T?"#:`8<IL:2&C1A883+%EA5BR3&!R<W+'@K\
M%8&Y6*TK`TJQ6E8K0.BY;*X*Q7!5R'QQF5FP<`Q7`KDN&,RLZW7TT\,N(>GX
M.D[`:TM`T:FA<@[5W>[<.X,5T*Y$TVF]WR?$4H4;;[>?F:?#?QV98Y,U/U<3
MHSQA'K92<E([WJ18,&TEV&AX+%[0"`,_J(]>GH>U^YZ6VM6M:M99^?[@RG]4
MX[3<,C8+8."AH`WM>APQ-[B@H*Q!WP);Q^:+DHV#;AH4U-E&65I6IY<>Z<^?
MVMD5\[A?+^.E7:'(--&X(@@B""@AWSGG[K`U`U^@8:X\W/,"KITU`2/>VYZ2
MRE+32)9&TVGG.\[Q)+>%:XFX7T<*MO)AV':<OE4V#:IWU,KS#8-`0`"P"")=
MWIU545M!B&SO]4Y&WQQ45%14W87:O&59H:VTX%;W8X#4DD$9*T4M\<ZT3R%A
M=9<R5O)&VY8$[CD250)Q`&VE*`C*5'A%8RP#.1A:PG;L4MXH:WM#$7>\%V"&
MJJJJJD#0&AEAMNX>0/*1@.6>W@B,X(SH4>08PC!C",&RHP<<IC(@L*)E0:,I
M@F8(<*3"A#!2ZZUV\O?O#MR*[[P$8@BNU[.5I6>K8N<$O#=<-@#<D;<0'2H@
MJP*T(C&T0:$*M4:9(6:KERLGAV``&9)F[[:\'END&8:;=Y6JTZFW0L::C\!"
M!+>&^$Y)+LE&J#ELD:(V",BV8)(9[HQF<`8;.IZX7`:Z((JD4%6<W8W568P<
M8VZ@-T&MQM2F'FV@!FT^"IO6@H*%B8CH'ECN2X/5LN4FV-AK4JYA1)!F@&`,
MT8%F/Y9:P<*R&C<-6U3$RIELT$"&`-,`0XRM,?A\=*75,)SG,:2#SH#B7G#N
M?21CX-G3F?('=K6+DC2E*TC24A2T`516!F9W'\H)R@2\<!,@6V/+]I>?1+@Z
M+=@[A2@#9T!U#1*D)#,0+$H<7HPBI"[EPBI(UI331K2F\()3$`[MG+PNG<6^
M`!<`MPX!M1-JC:MJV!Z5@."JP&`_T3_)S2AWZS20SD'L"`(6%A86%A86%@#N
M['00S(0-02'0RD%T]41`$C5IM-ED%X51`H!2#I=1)!NG-)(7J@&PP]D`^CG.
M*H-_37^$4[QE38!\:_;:,^ON]*\(RJ7#Y?=\/$\#3]['[BS,94?EU$O].1V/
MX9F!2@XII25&2$2`B`(A:E2]3%=@?R!^&RM@V5:5I6BM$TK16@TK16I6E:5:
MC@/(-4<:ED?35+2-5/JT&@Q5D^=8J\S8T31HT31HU1HT:I\V27780Y^W24<,
M69$/"P)`$`4Q(@?R!^@^XVCN"1IE1@$8E%(6"A%A6("6E4JR4S1365E2FM%?
M,'W'!L0V-,@LB,(1"P%#-(,H5*S`20D@0A`(RJL2,*L2,*8L9@8(Q)A,#$C$
MF%?AB:3%DP*[0P,"ZP?,.Q6JFP,!P&ZL2,*>UY?0<"_AEU<36614PF8JJY8!
M_38\VJ>R>C<<*[@P:DE%^J79"42N%:R_\/E&!]1\"%I;9$*AMD!C:?DAB62+
MO`,%T;WHXY1I-P#PR-):*UB:%`H&&C,1X0B&USRQV%"E`G$-#J>W]E_C\WLR
M`,!+]WVUPK)^N/XU/X>CD=!C#^L4\P[E?UJY*_Q*]P:5I6E;AN&ZNJ+Z`\6E
MI7BTI@8V5N.(Y#F.@Z#<;C=;K=;KCSY]+)+;)+997)@!H`X$MB7(ET$NWV?'
MV4Q(OP5/HJ,M:C#1%)B:QPTF+H`-,(Q)%@9H,F4SR?;XKR!/B`R,5TV]EC(8
MQ'7U<[H'VE7%7$/4&$J[W[QUJ0.=*)7I!C[H0,&C@`<`#@5NKS5YJ\@]P>`=
MIV':;'Z"G`X'`X'0KHJZ`[%785V%:5I6E;JW#<.AT.AT.#CSS'26+I+:LI58
MU6%J*K"+CBXF.=DA,C$B*(B.R`;S2-SM76?1J%I<-\A>-FV9PGEZ<RFMA%F@
M))'AZ.YUK$G.\N(,(QP0>?0"H;$%6D0HA0ZM%7<>+.MSL<^F]`\ISX^';%Q>
M]Y*;2:-;#+$QQRG=!<[/O808$(OW5(:3;:382&0R#3"1$V"@TFAI]0"("A3L
M8!!XEJEHV3D'=!D`?3)MH[H[C0$,;5:T<@`FTAIB7EM=+J$7$JH-C]+KYYX:
M^3OWLWX@'B+P`0L`0L+"PL#%8&*QZMGU<`<`T&Q,-'`V#1HT;&5P#[E;AH,!
MN`RE#606".#"R9@I$-T/N*O:T!D.!62N(O[$@4.`&8&)(L5,GP\]MMV1IDF8
MF)BC)*0"4I'C*<DQ3@32#TW^IFM61FK"8<P<8`,J%!K%Y"81!$J$!"1(4C3D
M/,>>"^8?3*")L`O].7IV[B7M=)!)DI[_>[1MM#?``H)<0#`Y=DO<QQ9JPH]&
M-MZ^W8U5NM(N''0S*-;ZFZ<R1_M32T5R^ZVOOORNP.X.0I@8);JVS;DZ1=F?
MWO'^7PE]^_L].GX]PD=?O'?8((6U!L$)3!!$`*H!H1##@]F/!Y/7A+#H%:B1
M78<N/":KP`*^7'B>VWV]8.LX)OM?;S@],X)LVZ;92Y\^G24OL._(Y'XM#0F3
M!2EK.4NX`TV#8!J`0`0`>"3`#:`2$E`!,`W[^/&4L<<\Y2MA;&4OK_#]?E^T
M_89GXM0;\$'Z>^#K@Z*K^=DC(#X+_?SJO$*LH.1$)P.$@=1OT)(/%+4,QD,R
M**(TTHBBBB--,(BB(BTM(HHB(M+6ZH`T31E`5LT#L@U_9!M!W/NIBP9&*\H-
MPYYQAS:5V?=\[H&/CAH%\'Y;?T]WYU+==/EXN$<"YUJN>"7+3G[Q3BZ@P8&@
MP+6!OORV"<=`YNGKF?!L?5W(0P.<CICDS&9;(`^X5XQ)3$EW`ET2/@H'X^FD
MI*4^K7=X;-QDTO=?2^UQ3N[[OD8]V&E3S>?[.'Y>J-UE=)@:*J3$7)(?AXUF
M)+R$@LH$J\\0H)(6>?EIC*7EY6TE+$QUS$D9&B"0`:^8:X.!H1M0D'(!T2.1
MML_F>/7W*$X,%K(RJ4$WPPF6G6?9P[!?H5.J"ABH,*N`$P&=YV]WDT?#Y_AL
MV<M:T<^D4#5!5%+5+0U0T-(Z"5^7V\&N"^D?Q`3GU_B%5#?U.IU,CJ=YWG4[
MDMHLQB2V(Z>)$,3@B/<?B?K/F]I[_U??JD;#M!]SVZJ/)O27V)'Q#9<E&@\0
MV+G&]#,J.0;.(8:##*U@->_?Y:<`:;"^?L4R<NCM[-MGVZUWF2C;;42K+FZ^
MK777J=(IZ+FQ5H0>.*5?B?8B$A><]P]27S9*03M:3+*0,!(E..AX$7=V@]NO
M15J8R6P8/#?-\X!K^,'!)?EU*OOQ+,#IHU47#44^,471"DR@A2!,D0I\RK"&
MN!XYVY]Y7;W=P>U<6%?J<DN81@O#<V@+92Q<[^,'Q^J3)QX[YZSJ7!BYT]$H
M=77\'%S%.EFL=]+;H<'TGU<G)P%/7*M*,U<KZH.3O]DK)S]IT^L(\Q,D#`Q(
MR>B]=0<.O/4&"R!@(SW9#=5.[:D!,)WA"$@JWZ.<:MKM[VD192*+EFYR<RXP
MHP(80!Z+`&2(A@L(^;L7UY;0L@"XNY&06QGM%D'(2!:=E,Y;_&DX9,G(<EL(
M4M*6E0Z6%+*6DI2TI2.D*665PI:6E*66M2%EL<C<Q32SX.X>W94+@U.05-$=
M/3`>IC)-$,MW;T'OS7ZV;?0!%7\>$#<`?!6GC7S=<^PL>A@\)=.G/MN0O=/6
M?,/G'K/6=CZ1BM-FE#9BAL9P,T,P9<</REEWM1[Y!`<U`)(:/5"$>HNY<?@`
MKW=W=7J>_O#!@NOMUS_ON(HR9Q<0D$_*2/:4/=(.35+X0O\?YL44-R`%%"![
MX%^KD8(?"3<O$/TP?>01!$%(CB&2-\J#:O96E6@P#V!@?3652_U?S_]Z!^N@
M?TL57_84P2?\\M8H3!@JQD?W?;]6W[GSP""'X&&"JC\0_]#T'Z#_PY1?U!]I
M]?X$5=PCA@48;.QZ*'VJ"-%_[^7G,A0[!FF5N/N&9,R9D_1S1XEA^]^C6'D<
M#1_^.!1_8?P\-FS`\'1CS?O<(?1[&S]#FN$3P#X'XF#YHB2J%\Y@+,"YDQ@+
M$(V*/((0?0?F`["F?96A<("B1*VRI42'*VR)42+0H-JR""#P-3R/0,-Y!\@R
MI,94D,J2)J#XR910\C#7"I^`WN&RJCX."EI97"TL+:6DHX4M)24=*4LI*,<*
M-VJE=K*5['LZ%,XU).#,S))F>`=Z$'`"28@/B86/I&GM?WG$P^STS-SEA0B!
MYAS,#1!2BOH/@!\$#^F#WOSU>1$?\9'[S[P_V/L#WW_,93-!R)(AU@XP3-K4
M:$TK2"9"!DA2*&(9&+-90(F/-Q!:17&,E,5=TVJW4MK;K:(B43:BQ1JHVT:L
MDF3;:)K*<+IW65N+:AH#4&@R<")HT)!(:!3%D*1"@4_P!(E:4FMDEMW*,8QC
M&.G,`@`]SGN=<=<ZZX`.N==<=<ZZX`UEFR4@S!4)@U@@L52RL!B`````"1)`
M`P-&J9;):DDDDDM220``!,9C3*6T&U!K`:M#`&:OL^D\CCJ\LT5>1S2KOP,S
M5E45F#9-#%@T#2C1&I48RQ$RE;Z@:I&V0I4@0E!#&E602!1A$-1@B$!+;A,&
M#",D(E3;-`31U+2&FE1)4(4"1`E`.(Q1!@5*@H1""E`,''!8!9,JLF2658B,
MF$J%68$(#@"`P+<,0Q99(:>2AX*7P86,3#2M+*E5K2RE)K)2S;2RE*2DI92R
ME*:1$1$MRJKE<DS2III:3)4FDDP2:&OE7T?O\"/I[?G^O[/2\/Z4HUX?-C(2
MJ8C>4S"SM=9JUKL$D,`8!4`H`%/"M;_=O@OOP9*`#^],\\"^5]V6-:?96,7+
M9LRGO`O`W)$)"5`0A*$07MKT<`6`RD1,1`3&P%4#VH`:<#6-*22D&CU*A-56
MH`HHP6`K`6E4VU@T&+`18C1*]FG`VH-$;$6BL$4J8"HT21HA0-*5E4T!-BJ`
M;!@H";%4`#5E4P&H`M0J2`(JJ20$M#PI'"DC2DVRE9*D1!$41$1$!$1$48WK
M;O[=<\_CQY_1T1$-R/EV;-W#CQX2]=--LI9_VW7-ZN-9VK6=][>&[7714YNF
M&%)/&)1LW6MA9X+`PNM$[4M*5-=D5;)$!186<4NPKA:A=>[U*ZEUU+Z&%J\^
M$6;OOOPM=*LKJ2NNM:M-_K(B(B(BB\F[OT+Q\?'S]QSG#HX[]^[@]V[=AAN-
MGPGPSEOUY<+V\'%T[6+6PPN9*S!`&U!@P8,&U!@B044445^3MZG,QSS^?ILW
MO>]G+G%)03!L!M-\Q(@@8R%"EA9MNO&)1$2.-K<.%]][<$0W&4I88@"FP0T9
M`X8ATW4KP)2==Z`FDA%-3M8[C?:+8;S4G:SK,H4*:U+LGHD60D0#0A/MN+]F
M4HB)&_?OI1MZZ7RFW!&5G3NWO4>H[!Q*0^><(F(B8DZRI0,$P5PQ!-/<0;>R
M(^>NLKB*JX8S<))ZP@G&%%%%%$,`D&*(HFI)INMD[G@?!U_%0;&L-O$]:>A)
M7_`_:;NI0]B]Q5A[SUPQE98;\THE%#D<#$88#C$#IT'!`D/I\9.C(!R.5S29
M#V&V)\@J'"N_+NF8F.Y!]B@B^OA':=<&'L/04^DZ]>IIM.6?9(NNN+IEUUY?
M,E++<;8WB0!$NP0!-+/'`SRQSSH1+,P0+%N:26T&N\+PA?Z&A+30C]6P0A*"
M\$)]_B>=RVMUE\>1\)"5O]U>U`U;[^L'N@R#XG!\3T0!Y@\WSCWG)KZ[)#\:
MM5(:M?0:U2?KVZ;60][N7ZBQ&%86$]'4BR6UE,VIJ32JM[`````2``3T]^VO
M46_!7=55Z*]>^"O=.FX351Z'Q?%B=K&9DS,S,LP)(B3$;V_61Y/@]O3Y]N^^
M'M7HYNPQ5X=G;R=NWAV7J:\-AC`8``(RLC8HT7KWCJO,JW@UN3$-);Q`!.BI
MEF\<\\Y9A4+T9KW`T#`F,0C3Z_%*46#5,N!F$Q[,K;PN(J:]DPU8!U&*/MCZ
M$65\(.LIQ.7-]146',U2GAK-45,:Z^L>8DJC&4S4#2WI)"X`F,\T$"RG8(C$
MQV>[CAPTK?+'K,JX/2?#WH`+;X#23.L<&QDPF2'QJB#E[ZJO.Z3>.T/*0'B`
MU,&5)E4MWL(,</[=/N@P+J&&]]X+0-KO[8WLD;A"TC;I2%(YF;=+%O!BV'PG
M]1LZ`?B>I39RA8ZLF8R>^GO.B[1]N.-2NSNS.-"-DJ])'>ZV&E.W=5[J[!DM
M65995BU54]3EXB)B=0H(?2H'=!`9&HDL2RZ\=8_/L`[!Q:J#"PDR9O:K:5RE
M-!"DI2$DE9F^[S7$!LVUFU9,:-%;-6@`@`2%FQ]$GS,$?8X^7%U#'QH]LU[O
M!FNW>O>G6L?7BY.%<,3IW>D4\H[$'@ZWGTVW"2$C>T%2$']T`<<%.`X:3;A$
M00(AMHN6AX@A5LWG-([Q)!Y"_!R:!TI;2BFU*U*KR=4!9?(VI>R7/K=\'54<
MNO=G/;+.%)H\ZQM-J*>!P1'Z>.5)W[_+F5ELH9(.T4$0D`_;R.RM%E3J@$+%
M6`L"**F*(H.0$8VC+3':T!42PD;B<!@-(*BRR$'6G".0;;=D=M;L15N#+M&,
MB!=W&XR7&-EU(-IU765JI+"Z`!LVW5KG+G#(1506,A%59&H-56%U``LH-O#M
M30!((\E1."0QB*A29`F%H06&E()2!B$)[(`T4"9YH<J55,90`PE1B1B1C&)8
M<22L0?5+KHE9/OBO@O4=VEU)'+0Z52YZ7KAA)725Z[.!)6#N1<@E:78JBZYA
M335/V:FJ3JX/@;!F3/BTC3&89DI[JKS<:L```$G?57AJKWZ=XV+J+\',<!XX
M/@48,)Y0*=?17LS.?8MUO=57=P:U6'+UKO=553VR4`$@"\ZF:!,`2Y;&VX(B
M"\"4FP#EYKS>-O;I;MZM2HJE4`9C&><'N@\>JJ*^C:#RJKG8,S,S'/P4QB[5
M3Z8N1>>0Q+(-D3NWZCEQF\#4;','`YE#>MX(U;;@B($A%S*JJKT"#P>!SZ58
MMD5U&RM]E2AW^92M,3SPK6$,9B9@\F)F56BC3$+<R0VR(9=EK1[)@>]AP7*#
M3%6)BPU;Q9T#H$12X^/JC)HH,P:PS-C\B9+NQC(S%RLC3$=%5["1QL'5Z3MN
M)<UEEAC@N7T/8@S"I?(PH=Z:GME=(-]FW`8JE=BQ%O)@2N`&[3I8S&'&*87@
MF)M5%8GIL-YA1PUF9+15S&N*K*D;`9%=K&,9DKFAEO[.;W<(.Q.JJ65[9.[-
M9&494.\.N*VA3,]T'C"YHXQ32KW;URO>>SZ><GL<#59LU4G/*!"?`:[:CAOF
M==1U)QG=%U`=Q1[?)J#$3&1C*6&;-Q.8KP@T00DC'E@;'*]]\0W$2WYZVA]V
MRSFWZ%MRTZ/I#<-EYV2L8C@G]Q3VRKF."IX*9C!"\!D,J82LH3*@Y14+D@>6
MQ0D8-<*AB+`),!\GL]JS`!M)M``U?KK'9@=M*<7'9$4S?*(;B*4Y<ML\NP.6
M^=TI=CY1G$28\6DE@TPT7)2&2(0M((GHL9@F`B_)*F'I]3H3@-E@,*)`2`#,
M$UJ&I4'LL0B02S"6VT)3F8$;Q=+!T.2#A#QX2>_XR12LI$9)AN&S`W/K+#5E
MV-LS&/&=!A5HM-4Q9J'L!/>$P,!B5?X%:DM5.+)*T%E0P`P#@/`,.9T:I9:A
M)$.G1(7`8!C)WZFJ/>I7]NGI,=:I+)O5.=(X1)=:C>%/34XG494>\N?-U9HT
M@(HH#`0`)KVFN`HJ`@P&``)H%+%F9K11HU>6[_4\>!Z.3@%M#WZ/-\+IHRUM
MNDE)*JQVMVRQPVW5%513K1PX<.'!R#M,,6Y'FB6L1%>SI:_@XXY=DC9?67<=
MO!(V\OD4TUJ2M;#L]2'J/74G[-]LFO@64E*>I`.@GJ%[7C'C73*6:57+J==:
M-H(K2).G:@O@RD,`Q<$IS?0"!YL?E0^5ACA#O%(`('6?/%%%?.`>B&ENL6+O
M<]HV4P&%T>9['=YYF%.E9SZ1$2Z;>EIO&"^`08M8)C1W`7B2W`'Z7V5^V0"8
M0JD%1,#]4.,8PK@D*M`M.I5=0NH0-6K4@'Q!8)`L43E(`-PF,8)IC!;K:<=K
MG.<IV6]YOW94738*?'TU58=P<G)0\J\SUF^%`#"H@(LB(B(DB*HDJJ#$D,0Y
M&Q@X=!M315%P0D%@._;H'J5:8(T#C7/ILG,\):2TR"J:KJ#[:(0Y+VY\\S(`
MLS.8!=.ZE&YD#.C"!C&4(FW293G6I4;;I6K*'!O9F;[\>NB5O."L-%SW=)>(
MX-F#-^4$1>H(BX=A5XI4._KQ7'3OJ*H`X..*XXX:::*PF[T/2]$X1DVBWY#F
M'7IRB@=&B,-15*=Z,GD.>EU1)=<Z[TE[\+\V>OJTX9')+XW;XQ'BA\U3S[`[
M\V*FVJ-XX\_AF8SSZ9F9TD>D6'')7(7GYDOQI/C)*[?5D9F&2?2JZYI.\=!,
M6,#"$QX4DI]$6!BQ!+$OD``\?+7B*B/S:[57L\3YD#>Z+A[":1CROB@.^*([
MC&99B2?$PNZ'86!-)!J^4)*E:-@#8-]0!&?F[#OW9,;D4Z<-E*&Z$I)0BN%#
M"=<'6AAAA2EU1E,P4(G2]X.5V%\+"V$B=A$)<-MKV[.(B(B"220GDQNGLN$S
MM@_SGV8'(O>M&.2O.[:DWKW1EWW!U8ZZV*=^D]IS#@EYW!,BLR2+!!E09D1,
2KXY&L$6E__%W)%.%"06>>ZEP
`
end

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[bangerth at dealii dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


bangerth at dealii dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


------- Additional Comments From bangerth at dealii dot org  2003-07-22 18:20 -------
I can't reproduce this problem with a snapshot from this night -- can you
double check whether you can see the problem with the preprocessed
sources (and then attach them to the report form, rather than mailing
to the audit trail -- thanks) and send us the command line you used when
invoking the compiler on the preprocessed sources?

Here's for the reference what I get with your code after stripping the
#line markers:

tmp/gg> ../build-gcc/gcc-install/bin/gcc -std=c99 -O -g -Wall -march=pentium4
-fomit-frame-pointer   -c -o sha1.o sha1.i -msse
sha1.i: In function `_mm_cvtsi32_si64':
sha1.i:118: warning: implicit declaration of function `__builtin_ia32_loadd_si64'
sha1.i:118: error: can't convert between vector values of different size
sha1.i: In function `_mm_cvtsi64_si32':
sha1.i:130: warning: implicit declaration of function `__builtin_ia32_si64_stored'
sha1.i: At top level:
sha1.i:1953: error: redefinition of '__v4si'
sha1.i:947: error: '__v4si' previously declared here

W.

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627



------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-22 18:24 -------
The preprocessed source is from the mainline and I could reproduce it on the mainline 
(20030714), but I did not look to reduce it.

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[dean-gcc at arctic dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627



------- Additional Comments From dean-gcc at arctic dot org  2003-07-22 18:29 -------
Subject: Re:  internal compiler error: in push_reload, at
 reload.c:1282

On Tue, 22 Jul 2003, bangerth at dealii dot org wrote:

> Here's for the reference what I get with your code after stripping the
> #line markers:
>
> tmp/gg> ../build-gcc/gcc-install/bin/gcc -std=c99 -O -g -Wall -march=pentium4
> -fomit-frame-pointer   -c -o sha1.o sha1.i -msse
> sha1.i: In function `_mm_cvtsi32_si64':
> sha1.i:118: warning: implicit declaration of function `__builtin_ia32_loadd_si64'
> sha1.i:118: error: can't convert between vector values of different size
> sha1.i: In function `_mm_cvtsi64_si32':
> sha1.i:130: warning: implicit declaration of function `__builtin_ia32_si64_stored'
> sha1.i: At top level:
> sha1.i:1953: error: redefinition of '__v4si'
> sha1.i:947: error: '__v4si' previously declared here

ack shoot sorry -- i had a local patch in my tree which modifies both
those _mm_cvtsi32_si64 and _mm_cvtsi64_si32 ... i don't think either are
used by this particular code though.  anyhow i'll resubmit ASAP with that
cleared up.

-dean

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |INVALID


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-25 00:31 -------
Mark this as invalid so it can be ...

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|INVALID                     |


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-25 00:32 -------
unconfirmed.

[Bug c/11627] internal compiler error: in push_reload, at reload.c:1282

[dean-gcc at arctic dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627



------- Additional Comments From dean-gcc at arctic dot org  2003-07-25 00:42 -------
Subject: Re:  internal compiler error: in push_reload, at
 reload.c:1282

On Thu, 25 Jul 2003, pinskia at physics dot uc dot edu wrote:

> unconfirmed.

does that mean you've been unable to reproduce it with the newer
sha1.i.bz2 i attached?  (i'm a bugzilla newbie.)

-dean

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |optimization


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-25 00:46 -------
unconfirmed just means the bug is new and no one has looked into it yet.  I can reproduce it on 
the mainline (20030724) but since I have not reduced the testcase yet, this bug is not marked as 
NEW yet.

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4                         |---

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[dhazeghi at yahoo dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627



------- Additional Comments From dhazeghi at yahoo dot com  2003-08-24 23:19 -------
I'm reducing this one.

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[dhazeghi at yahoo dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


dhazeghi at yahoo dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #4451 is|0                           |1
           obsolete|                            |


------- Additional Comments From dhazeghi at yahoo dot com  2003-08-24 23:52 -------
Created an attachment (id=4648)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4648&action=view)
reduced testcase (100 lines)

Takes -O -march=pentium4 -std=c99 to reproduce the ICE with this testcase
(mainline).

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[dhazeghi at yahoo dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627


dhazeghi at yahoo dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |ice-on-valid-code
   Last reconfirmed|0000-00-00 00:00:00         |2003-08-24 23:53:40
               date|                            |


------- Additional Comments From dhazeghi at yahoo dot com  2003-08-24 23:53 -------
Does not ICE with -O0.

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[bbrand at nki dot nl]

------- Additional Comments From bbrand at nki dot nl  2003-12-02 10:17 -------
Hello,

The MinGW developers suggested that I should add the following bug report
to this existing bug since it is probably the same one but with a much
smaller source to trigger it. (Item #852092 on MinGW's bug tracker for
the original report.)

Platform: windows NT4 sp6a
> gcc --version
gcc.exe (GCC) 3.3.1 (mingw special 20030804-1)
> grep __MINGW32_VERSION \lang\mingw331
\include\_mingw.h
#define __MINGW32_VERSION 3.2

Compilation of sse2 instructions using -msse2,
-mcpu=pentium4 or -march=pentium4 together
with optimization -O, -O1 or higher generates an
internal compiler error.

Preprocessed code:

typedef int __v4si __attribute__ ((__mode__(__V4SI__)));
typedef int __v2di __attribute__ ((mode (V2DI)));
typedef int __v8hi __attribute__ ((mode (V8HI)));

static __inline __v2di
_mm_add_epi32 (__v2di __A, __v2di __B)
{
  return (__v2di)__builtin_ia32_paddd128 ((__v4si)__A, (__v4si)__B);
}

static __inline __v2di
_mm_madd_epi16 (__v2di __A, __v2di __B)
{
  return (__v2di)__builtin_ia32_pmaddwd128 ((__v8hi)__A, (__v8hi)__B);
}

static __inline __v2di
_mm_srai_epi32 (__v2di __A, int __B)
{
  return (__v2di)__builtin_ia32_psradi128 ((__v4si)__A, __B);
}

static __inline __v2di
_mm_mulhi_epu16 (__v2di __A, __v2di __B)
{
  return (__v2di)__builtin_ia32_pmulhuw128 ((__v8hi)__A, (__v8hi)__B);
}

void test(void)
{
  short *p;
  int *q;
  __v2di v, u, w, _xf, _yf, _yf1;

  for(;;)
  {
    int a = __builtin_ia32_pextrw128 ((__v8hi)v, 0);
    w = ((__v2di)__builtin_ia32_pinsrw128 ((__v8hi)w, p[a+1], 7));

    v = _mm_mulhi_epu16(_yf1, _xf);
    u = _mm_madd_epi16(u, v);

    v = _mm_mulhi_epu16(_yf, _xf);
    w = _mm_madd_epi16(w, v);
    u = _mm_add_epi32(u, w);

    u = _mm_srai_epi32(u, 15);
    *(__v2di *)(q) = _mm_add_epi32(*(__v2di *)(q), u);
  }
}

Compile with e.g.:
gcc -c -O2 -msse2 prob3.i
or (rename snippet to prob3.ii):
g++ -c -O -march=pentium4 prob3.ii

Compiler output:
prob3.c: In function `test':
prob3.c:50: internal compiler error: in push_reload, at 
reload.c:1261
Please submit a full bug report,
with preprocessed source if appropriate.

The code snipped is very much reduced. Removing
any further lines or removal of the loop removes the
internal compiler error.

Manually substituting the _mm_... functions with the
__builtin_ia32_... functions as defined in xmmintrinsic.h
(or as shown in the preprocessed snipped above)
changes the internal compiler error from one in
push_reload to:
internal compiler error: in ix86_expand_binop_builtin, 
at config/i386/i386.c:13113

Possibly a red herring: a slightly larger snipped did
compile when for(;;) was changed into while(1),
in the above snippet this does not matter.

Regards,

Bob Brand


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627

[Bug optimization/11627] internal compiler error: in push_reload, at reload.c:1282

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-12 06:30 -------
Fixed for 3.4.0.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
      Known to work|                            |3.4.0 3.5.0
         Resolution|                            |FIXED
   Target Milestone|---                         |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11627