From 745224e04a03e4544c58d5d38d3c54f67100f8eb Mon Sep 17 00:00:00 2001 From: David Turner Date: Wed, 18 Jun 2014 01:54:42 -0400 Subject: [PATCH] refs.c: SSE2 optimizations for check_refname_component MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit Optimize check_refname_component using SSE2 on x86_64. git rev-parse HEAD is a good test-case for this, since it does almost nothing except parse refs. For one particular repo with about 60k refs, almost all packed, the timings are: Look up table: 29 ms SSE2: 23 ms This cuts about 20% off of the runtime. Ondřej Bílka suggested an SSE2 approach to the substring searches, which netted a speed boost over the SSE4.2 code I had initially written. Signed-off-by: David Turner Signed-off-by: Junio C Hamano --- git-compat-util.h | 11 +++ refs.c | 234 ++++++++++++++++++++++++++++++++++++++++---- t/t1402-check-ref-format.sh | 15 +++ t/valgrind/default.supp | 8 ++ 4 files changed, 250 insertions(+), 18 deletions(-) diff --git a/git-compat-util.h b/git-compat-util.h index 76910e6cd1..086fd374bb 100644 --- a/git-compat-util.h +++ b/git-compat-util.h @@ -677,6 +677,17 @@ void git_qsort(void *base, size_t nmemb, size_t size, #endif #endif +#if defined(__GNUC__) && defined(__x86_64__) +#include +/* + * This is the system memory page size; it's used so that we can read + * outside the bounds of an allocation without segfaulting. + */ +#ifndef PAGE_SIZE +#define PAGE_SIZE 4096 +#endif +#endif + #ifdef UNRELIABLE_FSTAT #define fstat_is_reliable() 0 #else diff --git a/refs.c b/refs.c index 637a93d695..25117fe0cd 100644 --- a/refs.c +++ b/refs.c @@ -7,21 +7,27 @@ /* * How to handle various characters in refnames: + * This table is used by both the SIMD and non-SIMD code. It has + * some cases that are only useful for the SIMD; these are handled + * equivalently to the listed disposition in the non-SIMD code. * 0: An acceptable character for refs - * 1: End-of-component - * 2: ., look for a preceding . to reject .. in refs - * 3: {, look for a preceding @ to reject @{ in refs - * 4: A bad character: ASCII control characters, "~", "^", ":" or SP + * 1: @, look for a following { to reject @{ in refs (SIMD or = 0) + * 2: \0: End-of-component and string + * 3: /: End-of-component (SIMD or = 2) + * 4: ., look for a preceding . to reject .. in refs + * 5: {, look for a preceding @ to reject @{ in refs + * 6: *, usually a bad character except, once as a wildcard (SIMD or = 7) + * 7: A bad character except * (see check_refname_component below) */ static unsigned char refname_disposition[256] = { - 1, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, - 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, - 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0, 0, 2, 1, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0, 0, 0, 4, + 2, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, + 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0, 0, 0, 4, 3, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7, 0, 0, 0, 0, 7, + 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7, 7, 0, 7, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 4, 0, 4, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 4, 4 + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0, 7, 7 }; /* @@ -33,8 +39,9 @@ static unsigned char refname_disposition[256] = { * - any path component of it begins with ".", or * - it has double dots "..", or * - it has ASCII control character, "~", "^", ":" or SP, anywhere, or - * - it ends with a "/". - * - it ends with ".lock" + * - it has pattern-matching notation "*", "?", "[", anywhere, or + * - it ends with a "/", or + * - it ends with ".lock", or * - it contains a "\" (backslash) */ static int check_refname_component(const char *refname, int flags) @@ -46,17 +53,19 @@ static int check_refname_component(const char *refname, int flags) int ch = *cp & 255; unsigned char disp = refname_disposition[ch]; switch (disp) { - case 1: + case 2: /* fall-through */ + case 3: goto out; - case 2: + case 4: if (last == '.') return -1; /* Refname contains "..". */ break; - case 3: + case 5: if (last == '@') return -1; /* Refname contains "@{". */ break; - case 4: + case 6: /* fall-through */ + case 7: return -1; } last = ch; @@ -79,7 +88,7 @@ out: return cp - refname; } -int check_refname_format(const char *refname, int flags) +static int check_refname_format_bytewise(const char *refname, int flags) { int component_len, component_count = 0; @@ -115,6 +124,195 @@ int check_refname_format(const char *refname, int flags) return 0; } +#if defined(__GNUC__) && defined(__x86_64__) +#define SSE_VECTOR_BYTES 16 + +/* Vectorized version of check_refname_format. */ +int check_refname_format(const char *refname, int flags) +{ + const char *cp = refname; + + const __m128i dot = _mm_set1_epi8('.'); + const __m128i at = _mm_set1_epi8('@'); + const __m128i curly = _mm_set1_epi8('{'); + const __m128i slash = _mm_set1_epi8('/'); + const __m128i zero = _mm_set1_epi8('\000'); + const __m128i el = _mm_set1_epi8('l'); + + /* below '*', all characters are forbidden or rare */ + const __m128i star_ub = _mm_set1_epi8('*' + 1); + + const __m128i colon = _mm_set1_epi8(':'); + const __m128i question = _mm_set1_epi8('?'); + + /* '['..'^' contains 4 characters: 3 forbidden and 1 rare */ + const __m128i bracket_lb = _mm_set1_epi8('[' - 1); + const __m128i caret_ub = _mm_set1_epi8('^' + 1); + + /* '~' and above are forbidden */ + const __m128i tilde_lb = _mm_set1_epi8('~' - 1); + + int component_count = 0; + + if (refname[0] == 0 || refname[0] == '/') { + /* entirely empty ref or initial ref component */ + return -1; + } + + /* + * Initial ref component of '.'; below we look for /. so we'll + * miss this. + */ + if (refname[0] == '.') { + if (refname[1] == '/' || refname[1] == '\0') + return -1; + if (!(flags & REFNAME_DOT_COMPONENT)) + return -1; + } + while(1) { + __m128i tmp, tmp1, result; + uint64_t mask; + + if ((uintptr_t) cp % PAGE_SIZE > PAGE_SIZE - SSE_VECTOR_BYTES - 1) + /* + * End-of-page; fall back to slow method for + * this entire ref. + */ + return check_refname_format_bytewise(refname, flags); + + tmp = _mm_loadu_si128((__m128i *)cp); + tmp1 = _mm_loadu_si128((__m128i *)(cp + 1)); + + /* + * This range (note the lt) contains some + * permissible-but-rare characters (including all + * characters >= 128), which we handle later. It also + * includes \000. + */ + result = _mm_cmplt_epi8(tmp, star_ub); + + result = _mm_or_si128(result, _mm_cmpeq_epi8(tmp, question)); + result = _mm_or_si128(result, _mm_cmpeq_epi8(tmp, colon)); + + /* This range contains the permissible ] as bycatch */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpgt_epi8(tmp, bracket_lb), + _mm_cmplt_epi8(tmp, caret_ub))); + + result = _mm_or_si128(result, _mm_cmpgt_epi8(tmp, tilde_lb)); + + /* .. */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, dot), + _mm_cmpeq_epi8(tmp1, dot))); + /* @{ */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, at), + _mm_cmpeq_epi8(tmp1, curly))); + /* // */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, slash), + _mm_cmpeq_epi8(tmp1, slash))); + /* trailing / */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, slash), + _mm_cmpeq_epi8(tmp1, zero))); + /* .l, beginning of .lock */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, dot), + _mm_cmpeq_epi8(tmp1, el))); + /* + * Even though /. is not necessarily an error, we flag + * it anyway. If we find it, we'll check if it's valid + * and if so we'll advance just past it. + */ + result = _mm_or_si128(result, _mm_and_si128( + _mm_cmpeq_epi8(tmp, slash), + _mm_cmpeq_epi8(tmp1, dot))); + + mask = _mm_movemask_epi8(result); + if (mask) { + /* + * We've found either end-of-string, or some + * probably-bad character or substring. + */ + int i = __builtin_ctz(mask); + switch (refname_disposition[cp[i] & 255]) { + case 0: /* fall-through */ + case 5: + /* + * bycatch: a good character that's in + * one of the ranges of mostly-forbidden + * characters + */ + cp += i + 1; + break; + case 1: + if (cp[i + 1] == '{') + return -1; + cp += i + 1; + break; + case 2: + if (!(flags & REFNAME_ALLOW_ONELEVEL) + && !component_count && !strchr(refname, '/')) + /* Refname has only one component. */ + return -1; + return 0; + case 3: + component_count ++; + /* + * Even if leading dots are allowed, don't + * allow "." as a component (".." is + * prevented by case 4 below). + */ + if (cp[i + 1] == '.') { + if (cp[i + 2] == '\0') + return -1; + if (flags & REFNAME_DOT_COMPONENT) { + /* skip to just after the /. */ + cp += i + 2; + break; + } + return -1; + } else if (cp[i + 1] == '/' || cp[i + 1] == '\0') + return -1; + break; + case 4: + if (cp[i + 1] == '.' || cp[i + 1] == '\0') + return -1; + /* .lock as end-of-component or end-of-string */ + if ((!strncmp(cp + i, ".lock", 5)) + && (cp[i + 5] == '/' || cp[i + 5] == 0)) + return -1; + cp += 1; + break; + case 6: + if (((cp == refname + i) || cp[i - 1] == '/') + && (cp[i + 1] == '/' || cp[i + 1] == 0)) + if (flags & REFNAME_REFSPEC_PATTERN) { + flags &= ~REFNAME_REFSPEC_PATTERN; + /* restart after the * */ + cp += i + 1; + continue; + } + /* fall-through */ + case 7: + return -1; + } + } else + cp += SSE_VECTOR_BYTES; + } +} + +#else + +int check_refname_format (const char *refname, int flags) +{ + return check_refname_format_bytewise(refname, flags); +} + +#endif + struct ref_entry; /* diff --git a/t/t1402-check-ref-format.sh b/t/t1402-check-ref-format.sh index 1a5a5f39fd..9aeb352b3d 100755 --- a/t/t1402-check-ref-format.sh +++ b/t/t1402-check-ref-format.sh @@ -64,6 +64,7 @@ valid_ref "$(printf 'heads/fu\303\237')" invalid_ref 'heads/*foo/bar' --refspec-pattern invalid_ref 'heads/foo*/bar' --refspec-pattern invalid_ref 'heads/f*o/bar' --refspec-pattern +invalid_ref 'heads/foo*//bar' --refspec-pattern ref='foo' invalid_ref "$ref" @@ -128,6 +129,20 @@ valid_ref NOT_MINGW "$ref" '--allow-onelevel --normalize' invalid_ref NOT_MINGW "$ref" '--refspec-pattern --normalize' valid_ref NOT_MINGW "$ref" '--refspec-pattern --allow-onelevel --normalize' + +valid_ref 'refs/heads/a-very-long-refname' +invalid_ref 'refs/heads/.a-very-long-refname' +invalid_ref 'refs/heads/abcdefgh0123..' +invalid_ref 'refs/heads/abcdefgh01234..' +invalid_ref 'refs/heads/abcdefgh012345..' +invalid_ref 'refs/heads/abcdefgh0123456..' +invalid_ref 'refs/heads/abcdefgh01234567..' +valid_ref 'refs/heads/abcdefgh0123.a' +valid_ref 'refs/heads/abcdefgh01234.a' +valid_ref 'refs/heads/abcdefgh012345.a' +valid_ref 'refs/heads/abcdefgh0123456.a' +valid_ref 'refs/heads/abcdefgh01234567.a' + test_expect_success "check-ref-format --branch @{-1}" ' T=$(git write-tree) && sha1=$(echo A | git commit-tree $T) && diff --git a/t/valgrind/default.supp b/t/valgrind/default.supp index 0a6724fcc4..332ab1a3b3 100644 --- a/t/valgrind/default.supp +++ b/t/valgrind/default.supp @@ -49,3 +49,11 @@ Memcheck:Addr4 fun:copy_ref } +{ + ignore-sse-check_refname_format + Memcheck:Addr8 + fun:check_refname_format + fun:cmd_check_ref_format + fun:handle_builtin + fun:main +} -- 2.11.4.GIT