[sdcc-devel] [ sdcc-Bugs-1838000 ] Bad code for a bitfield in triple operator
|
From: SourceForge.net <noreply@so...> - 2007-11-25 09:38:37
|
Bugs item #1838000, was opened at 2007-11-25 01:32 Message generated for change (Comment added) made by nobody You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 01:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
Thread view
|
From: SourceForge.net <noreply@so...> - 2007-11-25 09:32:05
|
Bugs item #1838000, was opened at 2007-11-25 01:32 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
|
From: SourceForge.net <noreply@so...> - 2007-11-25 09:38:37
|
Bugs item #1838000, was opened at 2007-11-25 01:32 Message generated for change (Comment added) made by nobody You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 01:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
|
From: SourceForge.net <noreply@so...> - 2008-02-26 20:09:28
|
Bugs item #1838000, was opened at 2007-11-25 10:32 Message generated for change (Comment added) made by rlar You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:09 Message: Logged In: YES user_id=1840151 Originator: NO for debugging purpose the offending code can be reduced to: struct { char b:1; } data * glbl; void func(char data *p) { *p++ = glbl->b ? 'A' : 'B'; *p = 'C'; } Robert Larice ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 10:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
|
From: SourceForge.net <noreply@so...> - 2008-02-26 20:37:37
|
Bugs item #1838000, was opened at 2007-11-25 10:32 Message generated for change (Comment added) made by rlar You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:37 Message: Logged In: YES user_id=1840151 Originator: NO Problem Analysis: within genPointerGet () sometimes a getFreePtr() is called (and emits a push) to free a register then sometimes genUnpackBits()/genIfxJump() is called and emits a conditional jump and after that freeAsmop() is called and emits a pop this results in something like: push .. ... conditional jump to label pop next-icode ... label: ... ------------------------------- call trace: gen51Code() processing the icode: bug-20-a.c(l13:s7:k8:d0:s0) iTemp5 [k10 lr7:8 so:0]{ ia0 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1}} = @[iTemp4 [k9 lr6:7 so:0]{ ia1 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1} near* }[r2 ]] calls: genPointerGet () calls genNearPointerGet() calls getFreePtr() emits a "push ..." then genPointerGet () calls genUnpackBits (operand * result, char *rname, int ptype, iCode *ifx) calls genIfxJump (iCode * ic, char *jval, operand *left, operand *right, operand *result) emits a conditional jump then genNearPointerGet() /* now some housekeeping stuff */ if (aop) /* we had to allocate for this iCode */ { ... freeAsmop (NULL, aop, ic, RESULTONSTACK (ic) ? FALSE : TRUE); and freeAsmop() emits a "pop" this results in "emit"-age of # Breakpoint 13, emitcode (inst=0x81249d9 "push", fmt=0x81249e2 "%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124a96 "%s,%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124bb0 "a,@%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81255bd "jb", fmt=0x812532d "%s,%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x8125289 "ljmp", fmt=0x81252ce "%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249bf "", fmt=0x81249b8 "%05d$:") at gen.c:158 # Breakpoint 4, freeAsmop (op=0x0, aaop=0x81e6600, ic=0x81dac10, pop=1) at gen.c:1076 # Breakpoint 13, emitcode (inst=0x81249de "pop", fmt=0x8124b1a "ar0") at gen.c:158 -------------------------------- Robert Larice ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:09 Message: Logged In: YES user_id=1840151 Originator: NO for debugging purpose the offending code can be reduced to: struct { char b:1; } data * glbl; void func(char data *p) { *p++ = glbl->b ? 'A' : 'B'; *p = 'C'; } Robert Larice ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 10:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
|
From: SourceForge.net <noreply@so...> - 2008-02-27 00:02:47
|
Bugs item #1838000, was opened at 2007-11-25 10:32 Message generated for change (Settings changed) made by maartenbrock You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) >Assigned to: Maarten Brock (maartenbrock) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:37 Message: Logged In: YES user_id=1840151 Originator: NO Problem Analysis: within genPointerGet () sometimes a getFreePtr() is called (and emits a push) to free a register then sometimes genUnpackBits()/genIfxJump() is called and emits a conditional jump and after that freeAsmop() is called and emits a pop this results in something like: push .. ... conditional jump to label pop next-icode ... label: ... ------------------------------- call trace: gen51Code() processing the icode: bug-20-a.c(l13:s7:k8:d0:s0) iTemp5 [k10 lr7:8 so:0]{ ia0 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1}} = @[iTemp4 [k9 lr6:7 so:0]{ ia1 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1} near* }[r2 ]] calls: genPointerGet () calls genNearPointerGet() calls getFreePtr() emits a "push ..." then genPointerGet () calls genUnpackBits (operand * result, char *rname, int ptype, iCode *ifx) calls genIfxJump (iCode * ic, char *jval, operand *left, operand *right, operand *result) emits a conditional jump then genNearPointerGet() /* now some housekeeping stuff */ if (aop) /* we had to allocate for this iCode */ { ... freeAsmop (NULL, aop, ic, RESULTONSTACK (ic) ? FALSE : TRUE); and freeAsmop() emits a "pop" this results in "emit"-age of # Breakpoint 13, emitcode (inst=0x81249d9 "push", fmt=0x81249e2 "%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124a96 "%s,%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124bb0 "a,@%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81255bd "jb", fmt=0x812532d "%s,%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x8125289 "ljmp", fmt=0x81252ce "%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249bf "", fmt=0x81249b8 "%05d$:") at gen.c:158 # Breakpoint 4, freeAsmop (op=0x0, aaop=0x81e6600, ic=0x81dac10, pop=1) at gen.c:1076 # Breakpoint 13, emitcode (inst=0x81249de "pop", fmt=0x8124b1a "ar0") at gen.c:158 -------------------------------- Robert Larice ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:09 Message: Logged In: YES user_id=1840151 Originator: NO for debugging purpose the offending code can be reduced to: struct { char b:1; } data * glbl; void func(char data *p) { *p++ = glbl->b ? 'A' : 'B'; *p = 'C'; } Robert Larice ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 10:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |
|
From: SourceForge.net <noreply@so...> - 2008-02-27 13:53:37
|
Bugs item #1838000, was opened at 2007-11-25 10:32 Message generated for change (Comment added) made by maartenbrock You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: msc51(8051) target >Group: fixed >Status: Closed >Resolution: Fixed Priority: 5 Private: No Submitted By: Nobody/Anonymous (nobody) Assigned to: Maarten Brock (maartenbrock) Summary: Bad code for a bitfield in triple operator Initial Comment: SDCC 2.7.0 produces erroneous code for a statement like *bp++ = sp->b ? 'A' : 'B'; The compiled code will crash when executed. More details in the enclosed file. Submitted by: j (dot) kangas (at) luukku (dot) com ---------------------------------------------------------------------- >Comment By: Maarten Brock (maartenbrock) Date: 2008-02-27 14:53 Message: Logged In: YES user_id=888171 Originator: NO Fixed in SDCC 2.7.5 #5054. Thanks to Robert Larice for the extensive pre-work. ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:37 Message: Logged In: YES user_id=1840151 Originator: NO Problem Analysis: within genPointerGet () sometimes a getFreePtr() is called (and emits a push) to free a register then sometimes genUnpackBits()/genIfxJump() is called and emits a conditional jump and after that freeAsmop() is called and emits a pop this results in something like: push .. ... conditional jump to label pop next-icode ... label: ... ------------------------------- call trace: gen51Code() processing the icode: bug-20-a.c(l13:s7:k8:d0:s0) iTemp5 [k10 lr7:8 so:0]{ ia0 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1}} = @[iTemp4 [k9 lr6:7 so:0]{ ia1 a2p0 re0 rm0 nos0 ru0 dp0}{unsigned-bitfield {0,1} near* }[r2 ]] calls: genPointerGet () calls genNearPointerGet() calls getFreePtr() emits a "push ..." then genPointerGet () calls genUnpackBits (operand * result, char *rname, int ptype, iCode *ifx) calls genIfxJump (iCode * ic, char *jval, operand *left, operand *right, operand *result) emits a conditional jump then genNearPointerGet() /* now some housekeeping stuff */ if (aop) /* we had to allocate for this iCode */ { ... freeAsmop (NULL, aop, ic, RESULTONSTACK (ic) ? FALSE : TRUE); and freeAsmop() emits a "pop" this results in "emit"-age of # Breakpoint 13, emitcode (inst=0x81249d9 "push", fmt=0x81249e2 "%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124a96 "%s,%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249d0 "mov", fmt=0x8124bb0 "a,@%s") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81255bd "jb", fmt=0x812532d "%s,%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x8125289 "ljmp", fmt=0x81252ce "%05d$") at gen.c:158 # Breakpoint 13, emitcode (inst=0x81249bf "", fmt=0x81249b8 "%05d$:") at gen.c:158 # Breakpoint 4, freeAsmop (op=0x0, aaop=0x81e6600, ic=0x81dac10, pop=1) at gen.c:1076 # Breakpoint 13, emitcode (inst=0x81249de "pop", fmt=0x8124b1a "ar0") at gen.c:158 -------------------------------- Robert Larice ---------------------------------------------------------------------- Comment By: rlar (rlar) Date: 2008-02-26 21:09 Message: Logged In: YES user_id=1840151 Originator: NO for debugging purpose the offending code can be reduced to: struct { char b:1; } data * glbl; void func(char data *p) { *p++ = glbl->b ? 'A' : 'B'; *p = 'C'; } Robert Larice ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2007-11-25 10:38 Message: Logged In: NO System sis not accept the enclosed file, I'll enclose it here: /* Demonstrates a problem in SDCC 2.7.0 This program will crash when run in an 8051. The error is on the line marked below. The compiler produces PUSH AR0, but the corresponding POP AR0 is misplaced and is not executed when s.b is 0. This leaves stack corrupted and executing RET in the end of func() crashes the program. The assembly for the marked statement: ; triple.c:33: *bp++ = sp->b ? 'A' : 'B'; mov ar1,r0 inc r0 push ar0 mov r0,#_s movx a,@r0 jnb acc.0,00103$ pop ar0 mov r2,#0x41 sjmp 00104$ 00103$: mov r2,#0x42 00104$: mov @r1,ar2 Compiler command line: sdcc --xram-size 256 triple.c Compiler version: SDCC : mcs51/gbz80/z80/avr/ds390/pic16/pic14/TININative/xa51/ds400/hc08 2.7.0 #4818 (May 31 2007) (MINGW32) Submitted by: j dot kangas at luukku dot com */ typedef struct st { unsigned char b:1; } ST; __pdata ST s; __idata unsigned char buf[5]; void func( void ) { __pdata ST *sp; __idata unsigned char *bp; sp = &s; bp = buf; *bp++ = sp->b ? 'A' : 'B'; // statement producing bad code *bp++ = 'C'; } void main( void ) { s.b = 0; func(); while (1); } ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=100599&aid=1838000&group_id=599 |