summaryrefslogtreecommitdiffhomepage
path: root/src/os/unix (follow)
AgeCommit message (Collapse)AuthorFilesLines
2014-11-19Style.Valentin Bartenev1-2/+2
2014-11-19Fixed type of sendfile() return value on Linux.Valentin Bartenev1-6/+6
There was no real problem since the amount of bytes can be sent is limited by NGX_SENDFILE_MAXSIZE to less than 2G. But that can be changed in the future
2014-08-13Moved writev() handling code to a separate function.Valentin Bartenev5-102/+60
This reduces code duplication and unifies debug logging of the writev() syscall among various send chain functions.
2014-08-13Merged conditions in the ngx_*_sendfile_chain() functions.Valentin Bartenev3-27/+10
No functional changes.
2014-08-13Moved the code for coalescing file buffers to a separate function.Valentin Bartenev3-72/+9
2014-11-19Refactored ngx_solaris_sendfilev_chain().Valentin Bartenev1-18/+7
Though ngx_solaris_sendfilev_chain() shouldn't suffer from the problem mentioned in d1bde5c3c5d2 since currently IOV_MAX on Solaris is 16, but this follows the change from 3d5717550371 in order to make the code look similar to other systems and potentially eliminates the problem in the future.
2014-10-07Introduced the ngx_output_chain_to_iovec() function.Valentin Bartenev5-318/+170
It deduplicates code of the send chain functions and uses only preallocated memory, which completely solves the problem mentioned in d1bde5c3c5d2.
2014-11-19Renamed ngx_handle_sent_chain() to ngx_chain_update_sent().Valentin Bartenev5-5/+5
No functional changes.
2014-08-13Events: introduced cancelable timers.Valentin Bartenev1-0/+2
2014-08-25Events: removed broken thread support from event timers.Valentin Bartenev1-1/+0
It's mostly dead code. And the idea of thread support for this task has been deprecated.
2014-10-28Core: added limit to recv_chain().Roman Arutyunyan3-8/+21
2014-08-27Fixed counting of sent bytes in the send chain functions on EINTR.Valentin Bartenev5-0/+5
Previously, a value of the "send" variable wasn't properly adjusted in a rare case when syscall was interrupted by a signal. As a result, these functions could send less data than the limit allows.
2014-09-22Removed duplicate initialization of the "rev" variable.Valentin Bartenev1-2/+0
2014-08-13Generalized definitions of the number of preallocated iovec's.Valentin Bartenev7-57/+24
No functional changes.
2014-08-13Reduced difference between the send chain functions.Valentin Bartenev1-6/+4
No functional changes. This follows the change from ad137a80919f.
2014-08-13Merged implementations of ngx_readv_chain().Valentin Bartenev1-97/+9
There's no real need in two separate implementations, with and without kqueue support.
2014-08-13Removed the "complete" variable from various send chain functions.Valentin Bartenev5-35/+10
It was made redundant by the previous change, since the "sent" variable is no longer modified.
2014-08-13Moved the code for adjusting sent buffers in a separate function.Valentin Bartenev5-167/+5
2014-08-13Fixed writev() debug log message in ngx_darwin_sendfile_chain().Valentin Bartenev1-1/+1
2014-09-08Added ngx_init_setproctitle() return code check.Maxim Dounin2-3/+5
The ngx_init_setproctitle() function, as used on systems without setproctitle(3), may fail due to memory allocation errors, and therefore its return code needs to be checked. Reported by Markus Linnala.
2014-09-01Events: removed broken thread support from posted events.Valentin Bartenev2-12/+4
It's mostly dead code. And the idea of thread support for this task has been deprecated.
2014-08-01Core: exit on ngx_pnalloc() failure.Piotr Sikora1-0/+4
Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
2014-07-28Bigger iovec buffer in ngx_readv_chain().Maxim Dounin1-1/+5
This helps to reduce likelyhood of memory allocations in ngx_readv_chain(), which are known to lead to noticeable effects in some cases, see http://mailman.nginx.org/pipermail/nginx/2014-July/044512.html.
2014-06-19FreeBSD has migrated to Bugzilla.Sergey Kandaurov2-2/+2
2014-05-28Configure: added -Wno-deprecated-declarations on OS X.Maxim Dounin1-3/+0
Previous workaround to avoid warnings on OS X due to deprecated system OpenSSL library (introduced in a3870ea96ccd) no longer works, as the MAC_OS_X_VERSION_MIN_REQUIRED macro is ignored on OS X 10.9 if a compiler used supports __attribute__(availability).
2014-05-23Events: use eventfd() instead of syscall(SYS_eventfd) if possible.Ruslan Ermilov1-0/+3
This fixes --with-file-aio support on systems that lack eventfd() syscall, notably aarch64 Linux. The syscall(SYS_eventfd) may still be necessary on systems that have eventfd() syscall in the kernel but lack it in glibc, e.g. as seen in the current CentOS 5 release.
2014-05-12Added syslog support for error_log and access_log directives.Vladimir Homutov1-4/+7
2014-05-20Configure: the --build= option.Ruslan Ermilov1-1/+1
If set, its value is output in "nginx -v" and in the error log.
2014-03-31Added #ifndef around NGX_HAVE_CASELESS_FILESYSTEM define.Maxim Dounin1-0/+2
This brings Cygwin compilation in line with other case-insensitive systems (notably win32 and OS X) where one can force case sensitivity using -DNGX_HAVE_CASELESS_FILESYSTEM=0.
2014-03-06Added connection serial number in logging of left open sockets.Sergey Kandaurov1-2/+2
2014-02-03Use ngx_socket_errno where appropriate.Piotr Sikora2-3/+3
Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
2014-01-30Core: handle getsockopt(TCP_FASTOPEN) failures.Piotr Sikora1-0/+1
Linux returns EOPNOTSUPP for non-TCP sockets and ENOPROTOOPT for TCP sockets, because getsockopt(TCP_FASTOPEN) is not implemented so far. While there, lower the log level from ALERT to NOTICE to match other getsockopt() failures. Signed-off-by: Piotr Sikora <piotr@cloudflare.com>
2014-01-04Added per-process random seeding (ticket #456).Maxim Dounin1-0/+2
2013-10-31Core: handling of getsockopt(TCP_DEFER_ACCEPT) failures.Maxim Dounin1-0/+1
Recent Linux versions started to return EOPNOTSUPP to getsockopt() calls on unix sockets, resulting in log pollution on binary upgrade. Such errors are silently ignored now.
2013-09-23Added ngx_filename_cmp() with "/" sorted to the left.Maxim Dounin6-11/+5
This patch fixes incorrect handling of auto redirect in configurations like: location /0 { } location /a- { } location /a/ { proxy_pass ... } With previously used sorting, this resulted in the following locations tree (as "-" is less than "/"): "/a-" "/0" "/a/" and a request to "/a" didn't match "/a/" with auto_redirect, as it didn't traverse relevant tree node during lookup (it tested "/a-", then "/0", and then falled back to null location). To preserve locale use for non-ASCII characters on case-insensetive systems, libc's tolower() used.
2013-09-05Fixed handling of the ready flag with kqueue.Valentin Bartenev2-0/+2
There is nothing to do more when recv() has returned 0, so we should drop the flag.
2013-09-02Disable symlinks: use O_PATH to open path components.Valentin Bartenev1-0/+3
It was introduced in Linux 2.6.39, glibc 2.14 and allows to obtain file descriptors without actually opening files. Thus made it possible to traverse path with openat() syscalls without the need to have read permissions for path components. It is effectively emulates O_SEARCH which is missing on Linux. O_PATH is used in combination with O_RDONLY. The last one is ignored if O_PATH is used, but it allows nginx to not fail when it was built on modern system (i.e. glibc 2.14+) and run with a kernel older than 2.6.39. Then O_PATH is unknown to the kernel and ignored, while O_RDONLY is used. Sadly, fstat() is not working with O_PATH descriptors till Linux 3.6. As a workaround we fallback to fstatat() with the AT_EMPTY_PATH flag that was introduced at the same time as O_PATH.
2013-09-02Added the NGX_EBADF define.Valentin Bartenev1-0/+1
2013-07-30Autoindex: improved ngx_de_info() error handling.Sergey Kandaurov1-1/+1
This allows to build a directory listing whenever a loop exists in symbolic link resolution of the path argument.
2013-08-20Format specifier fixes in error logging.Sergey Kandaurov1-1/+1
2013-08-15Unbreak building on Linux without sendfile64() support.Valentin Bartenev1-3/+3
It was broken in 8e446a2daf48 when the NGX_SENDFILE_LIMIT constant was added to ngx_linux_sendfile_chain.c having the same name as already defined one in ngx_linux_config.h. The newer is needed to overcome a bug in old Linux kernels by limiting the number of bytes to send per sendfile() syscall. The older is used with sendfile() on ancient kernels that works with 32-bit offsets only. One of these renamed to NGX_SENDFILE_MAXSIZE.
2013-08-08Don't lose pointer to first nonempty buf in ngx_*_sendfile_chain().Gleb Smirnoff4-66/+58
In ngx_*_sendfile_chain() when calculating pointer to a first non-zero sized buf, use "in" as iterator. This fixes processing of zero sized buf(s) after EINTR. Otherwise function can return zero sized buf to caller, and later ngx_http_write_filter() logs warning.
2013-07-25On DragonFlyBSD, TCP_KEEPIDLE and TCP_KEEPINTVL are in msecs.Ruslan Ermilov1-0/+5
Based on a patch by Sepherosa Ziehau.
2013-07-13Events: honor NGX_USE_GREEDY_EVENT when kqueue support is enabled.Valentin Bartenev2-2/+4
Currently this flag is needed for epoll and rtsig, and though these methods usually present on different platforms than kqueue, nginx can be compiled to support all of them.
2013-07-11Style.Maxim Dounin1-1/+1
2013-07-05Make macros safe.Gleb Smirnoff1-5/+5
2013-06-26Fixed build on Linux with x32 ABI.Maxim Dounin1-2/+7
On Linux x32 inclusion of sys/sysctl.h produces an error. As sysctl() is only used by rtsig event method code, which is legacy and not compiled in by default on modern linuxes, the sys/sysctl.h file now only included if rtsig support is enabled. Based on patch by Serguei I. Ivantsov.
2013-06-20Core: support several "error_log" directives.Vladimir Homutov1-0/+2
When several "error_log" directives are specified in the same configuration block, logs are written to all files with a matching log level. All logs are stored in the singly-linked list that is sorted by log level in the descending order. Specific debug levels (NGX_LOG_DEBUG_HTTP,EVENT, etc.) are not supported if several "error_log" directives are specified. In this case all logs will use debug level that has largest absolute value.
2013-06-12Valgrind: another complaint about uninitialized bytes.Tatsuhiko Kubo1-0/+2
2013-06-05Valgrind: supressed complaints about uninitialized bytes.Maxim Dounin1-0/+8
Valgrind complains if we pass uninitialized memory to a syscall: ==36492== Syscall param sendmsg(msg.msg_iov[0]) points to uninitialised byte(s) ==36492== at 0x6B5E6A: sendmsg (in /usr/lib/system/libsystem_kernel.dylib) ==36492== by 0x10004288E: ngx_signal_worker_processes (ngx_process_cycle.c:527) ==36492== by 0x1000417A7: ngx_master_process_cycle (ngx_process_cycle.c:203) ==36492== by 0x100001F10: main (nginx.c:410) ==36492== Address 0x7fff5fbff71c is on thread 1's stack Even initialization of all members of the structure passed isn't enough, as there is padding which still remains uninitialized and results in Valgrind complaint. Note there is no real problem here as data from uninitialized memory isn't used.
generated by cgit (git 2.34.1) at 2025-12-07 03:47:49 +0000