Mercurial > hg > nginx-vendor-0-7
view src/os/unix/ngx_gcc_atomic_x86.h @ 506:b9fdcaf2062b NGINX_0_7_65
nginx 0.7.65
*) Security: now nginx/Windows ignores trailing spaces in URI.
Thanks to Dan Crowley, Core Security Technologies.
*) Security: now nginx/Windows ignores short files names.
Thanks to Dan Crowley, Core Security Technologies.
*) Change: now the "009" status code is written to an access log for
proxied HTTP/0.9 responses.
*) Change: now the default buffer size of the
"large_client_header_buffers" directive is 8K.
Thanks to Andrew Cholakian.
*) Change: now default SSL ciphers are "HIGH:!ADH:!MD5".
*) Change: now SSLv2 protocol is disabled by default.
*) Change: now $host variable value is always low case.
*) Feature: the conf/fastcgi.conf for simple FastCGI configurations.
*) Feature: now URI part is not required a "proxy_pass" directive if
variables are used.
*) Feature: the $ssl_session_id variable.
*) Bugfix: if a proxied or FastCGI request was internally redirected to
another proxied or FastCGI location, then $upstream_response_time
variable may have abnormally large value; the bug had appeared in
0.7.63.
*) Bugfix: if the "expires modified" set date in the past, then a
negative number was set in the "Cache-Control" response header
line.
Thanks to Alex Kapranoff.
*) Bugfix: nginx closed a connection if a cached response had an empty
body.
Thanks to Piotr Sikora.
*) Bugfix: nginx cached a 304 response if there was the "If-None-Match"
header line in a proxied request.
Thanks to Tim Dettrick and David Kostal.
*) Bugfix: nginx did not treat a comma as separator in the
"Cache-Control" backend response header line.
*) Bugfix: cached HTTP/0.9 responses were handled incorrectly.
*) Bugfix: nginx sent gzipped responses to clients those do not support
gzip, if "gzip_static on" and "gzip_vary off"; the bug had appeared
in 0.8.16.
*) Bugfix: nginx always added "Content-Encoding: gzip" response header
line in 304 responses sent by ngx_http_gzip_static_module.
*) Bugfix: the "!-x" operator did not work.
Thanks to Maxim Dounin.
*) Bugfix: a segmentation fault might occur in a worker process, if
limit_rate was used in HTTPS server.
Thanks to Maxim Dounin.
*) Bugfix: a segmentation fault might occur in a worker process while
$limit_rate logging.
Thanks to Maxim Dounin.
*) Bugfix: nginx did not support dates in 2038 year on 32-bit platforms;
*) Bugfix: nginx/Windows tried to delete a temporary file twice if the
file should replace an already existent file.
*) Bugfix: nginx/Windows tried to rename a temporary file twice if the
file should replace an already existent file.
*) Bugfix: nginx/Windows might not create temporary file, a cache file,
or "proxy/fastcgi_store"d file if a worker had no enough access
rights for top level directories.
*) Bugfix: in UTF-8 encoding support by "try_files" directive in
nginx/Windows.
*) Bugfix: UTF-8 encoding usage in the ngx_http_autoindex_module.
Thanks to Maxim Dounin.
*) Bugfix: the ngx_http_autoindex_module did not show the trailing
slash in links to a directory; the bug had appeared in 0.7.15.
*) Bugfix: nginx did not close a log file set by the --error-log-path
configuration option; the bug had appeared in 0.7.53.
*) Bugfix: "addition_types" directive was incorrectly named
"addtion_types".
*) Bugfix: invalid request line in $request variable was written in
access_log only if error_log was set to "info" or "debug" level.
author | Igor Sysoev <http://sysoev.ru> |
---|---|
date | Mon, 01 Feb 2010 00:00:00 +0000 |
parents | 052a7b1d40e5 |
children |
line wrap: on
line source
/* * Copyright (C) Igor Sysoev */ #if (NGX_SMP) #define NGX_SMP_LOCK "lock;" #else #define NGX_SMP_LOCK #endif /* * "cmpxchgl r, [m]": * * if (eax == [m]) { * zf = 1; * [m] = r; * } else { * zf = 0; * eax = [m]; * } * * * The "r" means the general register. * The "=a" and "a" are the %eax register. * Although we can return result in any register, we use "a" because it is * used in cmpxchgl anyway. The result is actually in %al but not in %eax, * however, as the code is inlined gcc can test %al as well as %eax, * and icc adds "movzbl %al, %eax" by itself. * * The "cc" means that flags were changed. */ static ngx_inline ngx_atomic_uint_t ngx_atomic_cmp_set(ngx_atomic_t *lock, ngx_atomic_uint_t old, ngx_atomic_uint_t set) { u_char res; __asm__ volatile ( NGX_SMP_LOCK " cmpxchgl %3, %1; " " sete %0; " : "=a" (res) : "m" (*lock), "a" (old), "r" (set) : "cc", "memory"); return res; } /* * "xaddl r, [m]": * * temp = [m]; * [m] += r; * r = temp; * * * The "+r" means the general register. * The "cc" means that flags were changed. */ #if !(( __GNUC__ == 2 && __GNUC_MINOR__ <= 7 ) || ( __INTEL_COMPILER >= 800 )) /* * icc 8.1 and 9.0 compile broken code with -march=pentium4 option: * ngx_atomic_fetch_add() always return the input "add" value, * so we use the gcc 2.7 version. * * icc 8.1 and 9.0 with -march=pentiumpro option or icc 7.1 compile * correct code. */ static ngx_inline ngx_atomic_int_t ngx_atomic_fetch_add(ngx_atomic_t *value, ngx_atomic_int_t add) { __asm__ volatile ( NGX_SMP_LOCK " xaddl %0, %1; " : "+r" (add) : "m" (*value) : "cc", "memory"); return add; } #else /* * gcc 2.7 does not support "+r", so we have to use the fixed * %eax ("=a" and "a") and this adds two superfluous instructions in the end * of code, something like this: "mov %eax, %edx / mov %edx, %eax". */ static ngx_inline ngx_atomic_int_t ngx_atomic_fetch_add(ngx_atomic_t *value, ngx_atomic_int_t add) { ngx_atomic_uint_t old; __asm__ volatile ( NGX_SMP_LOCK " xaddl %2, %1; " : "=a" (old) : "m" (*value), "a" (add) : "cc", "memory"); return old; } #endif /* * on x86 the write operations go in a program order, so we need only * to disable the gcc reorder optimizations */ #define ngx_memory_barrier() __asm__ volatile ("" ::: "memory") /* old as does not support "pause" opcode */ #define ngx_cpu_pause() __asm__ (".byte 0xf3, 0x90")