| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
An idle connection was not removed from idle connection list
if the connections detected that listening socket had been closed.
|
|
Messed up return values in nxt_upstream_find() caused error in applying any
configuration with a valid "pass" value in router configuration pointing to
upstream. That wasn't the case in "listeners" objects, where the return value
wasn't checked.
Also, it caused segfault in cases where the "pass" option was configured with
variables and resulting value was pointing to a non-existent upstream.
Added missing return checks as well to catch possible memory allocation errors.
The bug was introduced in d32bc428f46b.
This closes #472 issue on GitHub.
|
|
|
|
|
|
The coming ASGI support requires raw HTTP headers format. Headers grouping
and upcase code were moved to WSGI module.
|
|
|
|
This is required for futher ASGI implementation.
|
|
No functional changes. Get ready for an increase in file number.
|
|
Some of the pointers were not adjusted after frame's memory re-allocation.
Fortunately, this function was not used and the bug has no effect.
|
|
|
|
Also added few tests for $uri and minor style fixes.
|
|
This closes #407 issue on GitHub.
|
|
|
|
Now it's possible to disable default bind mounts of
languages by setting:
{
"isolation": {
"automount": {
"language_deps": false
}
}
}
In this case, the user is responsible to provide a "rootfs"
containing the language libraries and required files for
the application.
|
|
|
|
|
|
It's not used since cbcd76704c90.
This option is a leftover from previous IPC between router and applications
processes. It was never documented, though.
Thanks to 洪志道 (Hong Zhi Dao).
|
|
|
|
|
|
|
|
Thanks to 洪志道 (Hong Zhi Dao).
This closes #466 issue on GitHub.
|
|
This is a quick and dirty sendfile() replacement.
This closes #452 PR on GitHub.
|
|
Using #if directives instead of #ifdef the same way as in other places.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Buffer for application prefork request allocated from temp conf mem_pool.
If error response from main process received before buffer completion handler,
temp conf mem_pool destroyed and router may crash in completion handler.
Assertion "src/nxt_buf.c:208 assertion failed: data == b->parent" triggered
when NXT_DEBUG_ALLOC enabled in configure.
This patch disables completion handler and memory allocated for buffer
released with memory pool.
|
|
Previously, an error during the prefork phase triggered assert:
src/nxt_port.c:27 assertion failed: port->pair[0] == -1
and resulted in exiting of the main process.
This could be easily reproduced by pushing a configuration with "rootfs",
when daemon is running without required permissions.
|
|
Found by Coverity (CID 361277).
|
|
|
|
This closes #441 PR on GitHub.
|
|
After shared application port introducing, request queue in router was
removed and requests may stuck forever waiting for another process start.
|
|
|
|
|
|
Two consecutive fd and fd2 fields replaced with array.
|
|
The default libunit behavior relies on blocking the recv() call for port file
descriptors, which an application may override if needed. For external
applications, port file descriptors were toggled to blocking mode before the
exec() call. If the exec() call failed, descriptor remained blocked, so the
process hanged while trying to read from it.
This patch moves file descriptor mode switch inside libunit.
|
|
|
|
The goal is to minimize the number of syscalls needed to deliver a message.
|
|
- naive circular queue, described in the article "A Scalable, Portable, and
Memory-Efficient Lock-Free FIFO Queue" by Ruslan Nikolaev:
https://drops.dagstuhl.de/opus/volltexte/2019/11335/pdf/LIPIcs-DISC-2019-28.pdf
- circular queue, proposed by Valentin Bartenev in the "Unit router application
IPC" design draft
|
|
|
|
Mostly harmless.
|
|
Each application initializes a shared port with 2 file descriptors, so the test
fails because the router reaches the open files limit.
|
|
This is a temporary solution after the 'shared port' patch. The application
process becomes idle immediately after creation. Even if it starts processing
a request (without acknowledging it yet), it is stopped by the router because
an 'out-of-idle-time' event occurs.
|
|
|
|
Generic process-to-process shared memory exchange is no more required. Here,
it is transformed into a router-to-application pattern. The outgoing shared
memory segments collection is now the property of the application structure.
The applications connect to the router only, and the process only needs to group
the ports.
|
|
This is the port shared between all application processes which use it to pass
requests for processing. Using it significantly simplifies the request
processing code in the router. The drawback is 2 more file descriptors per each
configured application and more complex libunit message wait/read code.
|
