Note that the indentation in these (and most other) files is 4 spaced - not tabs. So you can fix that with a followup push (don't force-push).
Also, next time, try to follow the common convention of commit messages: <subject>: <short summary (up to 50-ish chars)> <empty line> <description, preferably wrap lines at 72 chars, except long URLs etc>. You can check prior commits for the commit message format, or elsewhere, like described here or at many other places: https://ec.europa.eu/component-library/v1.15.0/eu/docs/conventions/git/ avih On Thursday, December 26, 2024 at 06:36:03 AM GMT+2, kbkp...@sina.com <kbkp...@sina.com> wrote: I commit a patch here: https://repo.or.cz/tinycc.git/commit/af1cfd9e821fde20ad5bc486ca401f3fcd640c9f In `void gfunc_call(int nb_args)`, it is needed fetch cpu flag before generating any code. This is true for other gen, but not for x86_64/i386. So I commit this patch. ________________________________ kbkp...@sina.com > > From: kbkp...@sina.com > Date: 2024-12-20 11:36 > To: tinycc-devel > Subject: Re: [Tinycc-devel] [tcc-win32] compiler generate wrong instruction > sequence > > A more simple testcase : > > #include <stdio.h> > > struct string { > char *str; > int len; > }; > > void dummy(struct string fpath, int dump_arg) { > } > > int main() { > int a = 1; > struct string x; > x.str = "gg.v"; > x.len = 4; > dummy(x, a == 0); > printf("done\n"); > return 0; > } > > run this will cause invalid memory also: > > bug_tcc.c:16: at main: RUNTIME ERROR: invalid memory access > > > > ________________________________ > > > kbkp...@sina.com > >> >> From: kbkp...@sina.com >> Date: 2024-12-20 10:45 >> To: tinycc-devel >> Subject: Re: [Tinycc-devel] [tcc-win32] compiler generate wrong instruction >> sequence >> >> By a quick search the commits, I found this bug was introduced by commit >> 8227db3a23fd3cf11840eaa25eab5f3f5f813ac7: >> >> >> >> commit 8227db3a23fd3cf11840eaa25eab5f3f5f813ac7 >> Author: grischka <grischka> >> Date: Sat Jun 22 11:45:35 2019 +0200 >> >> jump optimizations >> >> This unifies VT_CMP with VT_JMP(i) by using mostly VT_CMP >> with both a positive and a negative jump target list. >> >> Such we can delay putting the non-inverted or inverted jump >> until we can see which one is nore suitable (in most cases). >> >> example: >> if (a && b || c && d) >> e = 0; >> >> before this patch: >> a: 8b 45 fc mov 0xfffffffc(%ebp),%eax >> d: 83 f8 00 cmp $0x0,%eax >> 10: 0f 84 11 00 00 00 je 27 <main+0x27> >> 16: 8b 45 f8 mov 0xfffffff8(%ebp),%eax >> 19: 83 f8 00 cmp $0x0,%eax >> 1c: 0f 84 05 00 00 00 je 27 <main+0x27> >> 22: e9 22 00 00 00 jmp 49 <main+0x49> >> 27: 8b 45 f4 mov 0xfffffff4(%ebp),%eax >> 2a: 83 f8 00 cmp $0x0,%eax >> 2d: 0f 84 11 00 00 00 je 44 <main+0x44> >> 33: 8b 45 f0 mov 0xfffffff0(%ebp),%eax >> 36: 83 f8 00 cmp $0x0,%eax >> 39: 0f 84 05 00 00 00 je 44 <main+0x44> >> 3f: e9 05 00 00 00 jmp 49 <main+0x49> >> 44: e9 08 00 00 00 jmp 51 <main+0x51> >> 49: b8 00 00 00 00 mov $0x0,%eax >> 4e: 89 45 ec mov %eax,0xffffffec(%ebp) >> 51: ... >> >> with this patch: >> a: 8b 45 fc mov 0xfffffffc(%ebp),%eax >> d: 83 f8 00 cmp $0x0,%eax >> 10: 0f 84 0c 00 00 00 je 22 <main+0x22> >> 16: 8b 45 f8 mov 0xfffffff8(%ebp),%eax >> 19: 83 f8 00 cmp $0x0,%eax >> 1c: 0f 85 18 00 00 00 jne 3a <main+0x3a> >> 22: 8b 45 f4 mov 0xfffffff4(%ebp),%eax >> 25: 83 f8 00 cmp $0x0,%eax >> 28: 0f 84 14 00 00 00 je 42 <main+0x42> >> 2e: 8b 45 f0 mov 0xfffffff0(%ebp),%eax >> 31: 83 f8 00 cmp $0x0,%eax >> 34: 0f 84 08 00 00 00 je 42 <main+0x42> >> 3a: b8 00 00 00 00 mov $0x0,%eax >> 3f: 89 45 ec mov %eax,0xffffffec(%ebp) >> 42: ... >> >> >> >> >> ________________________________ >> >> >> kbkp...@sina.com >> >>> >>> From: kbkp...@sina.com >>> Date: 2024-12-20 10:15 >>> To: tinycc-devel >>> Subject: [Tinycc-devel] [tcc-win32] compiler generate wrong instruction >>> sequence >>> >>> It was only detected under Windows. >>> >>> bug_tcc.c: >>> >>> typedef unsigned char* byteptr; >>> typedef unsigned char u8; >>> >>> struct string { >>> u8* str; >>> int len; >>> }; >>> typedef struct string string; >>> >>> #define _SLIT(s) ((string){ \ >>> .str=(byteptr)("" s), \ >>> .len=(sizeof(s)-1)}) >>> >>> int _v_dump_expr_int(string fpath, int dump_arg) { >>> return 0; >>> } >>> void main() { >>> int a = 1; >>> _v_dump_expr_int(_SLIT("gg.v"), (a==0 ? (0) : (1))); >>> } >>> >>> compile it with following command: >>> >>> tcc.exe -bt bug_tcc.c >>> >>> run it, will cause an Error: >>> >>> bug_tcc.c:19: at main: RUNTIME ERROR: invalid memory access >>> >>> By disassemble the bug_tcc.exe file, it is shown that the compiler generate >>> a wrong instruction sequence: >>> >>> .text:0000000000401024 var_30 = byte ptr -30h >>> .text:0000000000401024 var_18 = qword ptr -18h >>> .text:0000000000401024 var_10 = dword ptr -10h >>> .text:0000000000401024 var_4 = dword ptr -4 >>> .text:0000000000401024 >>> .text:0000000000401024 push rbp >>> .text:0000000000401025 mov rbp, rsp >>> .text:0000000000401028 sub rsp, 50h >>> .text:000000000040102F mov eax, 1 >>> .text:0000000000401034 mov [rbp+var_4], eax >>> .text:0000000000401037 mov eax, 10h >>> .text:000000000040103C mov r8, rax ; Size >>> .text:000000000040103F mov eax, 0 >>> .text:0000000000401044 mov r11, rax >>> .text:0000000000401047 lea rax, [rbp+var_18] >>> .text:000000000040104B mov r10, rax >>> .text:000000000040104E mov rcx, r10 ; void * >>> .text:0000000000401051 mov rdx, r11 ; Val >>> .text:0000000000401054 call memset >>> .text:0000000000401059 lea rax, aGgV ; "gg.v" >>> .text:0000000000401060 mov [rbp+var_18], rax >>> .text:0000000000401064 mov eax, 4 >>> .text:0000000000401069 mov [rbp+var_10], eax >>> .text:000000000040106C mov eax, [rbp+var_4] >>> .text:000000000040106F cmp eax, 0 >>> .text:0000000000401072 jnz loc_40107D >>> .text:0000000000401078 jmp loc_401082 >>> .text:000000000040107D ; >>> --------------------------------------------------------------------------- >>> .text:000000000040107D >>> .text:000000000040107D loc_40107D: ; CODE XREF: >>> sub_401024+4E↑j >>> .text:000000000040107D jmp loc_40108E >>> .text:0000000000401082 ; >>> --------------------------------------------------------------------------- >>> .text:0000000000401082 >>> .text:0000000000401082 loc_401082: ; CODE XREF: >>> sub_401024+54↑j >>> .text:0000000000401082 lea rax, [rsp+50h+var_30] >>> // this instruction should move >>> outside... >>> .text:0000000000401087 mov eax, 0 >>> // BUG!!! overwrite the >>> rax/eax, this will cause invalid memory access >>> .text:000000000040108C jmp short loc_401093 >>> .text:000000000040108E ; >>> --------------------------------------------------------------------------- >>> .text:000000000040108E >>> .text:000000000040108E loc_40108E: ; CODE XREF: >>> sub_401024:loc_40107D↑j >>> .text:000000000040108E mov eax, 1 >>> .text:0000000000401093 >>> .text:0000000000401093 loc_401093: ; CODE XREF: >>> sub_401024+68↑j >>> .text:0000000000401093 push rsi >>> .text:0000000000401094 push rdi >>> .text:0000000000401095 lea rsi, [rbp+var_18] >>> .text:0000000000401099 mov rdi, rax >>> // Because rax was overwrite >>> to zero, it will fail at movsq >>> .text:000000000040109C movsq >>> .text:000000000040109E movsq >>> .text:00000000004010A0 pop rdi >>> .text:00000000004010A1 pop rsi >>> .text:00000000004010A2 mov r11, rax >>> .text:00000000004010A5 lea r10, [rsp+50h+var_30] >>> .text:00000000004010AA mov rcx, r10 >>> .text:00000000004010AD mov rdx, r11 >>> .text:00000000004010B0 call sub_401000 >>> .text:00000000004010B5 leave >>> .text:00000000004010B6 retn >>> .text:00000000004010B6 sub_401024 endp >>> >>> >>> Hope someone can fix it. >>> >> > _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
