| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
| |
Generated with
sed -i 's|Copyright.*Andre Noll.*|SPDX-License-Identifier: GPL-2.0 */|g' *.c *.h
followed by manually tweaking the result a bit. No license change intended.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The moderately recent commit which revamped the stdin/stdout API missed to
patch the client code to call the cleanup functions {stdin,stdout}_free()
introduced by that commit.
Also, both {stdin,stdout}.c missed to remove the buffer tree node in certain
cases. The additional btr_remove_node() calls fix this, and they don't hurt
since this function is idempotent.
Fixes: c30b3cf7b9eb76f6b0a6259a4f630017de804608
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With the current API it is impossible to free the memory that was allocated
by stdin_task_register() after schedule() returns. This commit replaces the
APIs defined in stdin.h and stdout.h to remove the memory leaks which are
inevitable with the old API. We also make struct stdin_task and stdout_task
opaque because the API users (client.c, filter.c, recv.c and write.c) should
not be able to manipulate the fields of these structures directly.
The new stdout API differs from the new stdin API only in that stdout_new()
needs an additional argument to let the caller pass in a pointer to the
parent buffer tree node. Other than that, the APIs are symmetric.
Use the opportunity to do the non-blocking dance in std{in,out}_new()
and std{in,out}_free() rather than in the post-monitor functions. This
way we may get rid of the silly ->must_set_nonblock_flag boolean of struct
std{in,out}_task.
|
| |
|
|
|
|
|
| |
The word "monitor" is neutral and continues to be correct after the
switch from select(2) to poll(2).
Pure rename, nothing to see here.
|
| |
|
|
| |
Don't state the obvious and avoid talking about fd sets.
|
| |
|
|
|
|
|
|
| |
This preparatory patch for replacing select() renames para_fd_set()
to sched_fd_set(), moves it to sched.c and makes it static. All
users are modified to call either of the two new public functions
sched_monitor_{read,write}fd() which take a pointer to struct sched
rather than an fd set pointer.
|
| |
|
|
|
| |
Two trivial wrappers for FD_ISSET() which hide the fact that we're
still using the select(2) API.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The GPLv2 line does not add any additional information, so drop
it. This leaves a single line of legalese text for most files, which
is about the amount of screen real estate it deserves.
This patch was created with the following script (plus some manual
fixups):
awk '{
if (NR <= 5) {
gs = gensub(/.*Copyright.* ([0-9]+).*Andre Noll.*/, "\\1", "g")
if (gs != $0)
year = gs
next
}
if (NR == 6 && year != "")
printf("/* Copyright (C) %s Andre Noll <maan@tuebingen.mpg.de>, see file COPYING. */\n", year)
print
}'
|
| |\
| |
| |
| |
| | |
* maint:
stdin/stdout: Only set nonblock flags for non-tty fds.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Although the ->post_select methods of all paraslash executables perform
I/O only when select(2) reports that the file descriptor is ready,
we set the O_NONBLOCK flag for all monitored fds. This is considered
good practice because, in general, it might happen that a subsequent
read(2) call blocks even if select(2) indicates that the fd is ready
for reading. For example, an fd corresponding to a TCP socket might
be flagged as ready for reading if a network packet with incorrect
checksum has arrived, but a subsequent read(2) blocks until the packet
has been retransmitted.
However, stdin and stdout often correspond to a terminal device where
the above scenario won't happen. Moreover, for terminals it's essential
to reset the O_NONBLOCK flag to the old value on exit because the
shell refers to the same file description and thus shares the file
status flags, including O_NONBLOCK. Many terminal applications, for
example dialog(1), expect stdout to be set to blocking mode and fail
in arcane ways if O_NONBLOCK is set.
When the stdin and stdout tasks are about to exit, they reset the
file status flags back to the original values. However if "para_client
stat" is killed with SIGINT, SIGTERM or SIGKILL, or put to sleep with
SIGSTOP, the O_NONBLOCK flag remains set because para_client does not
handle signals at all. para_recv, para_filter and para_write suffer
from the same issue. Adding signal handling to these programs would
not help in the SIGSTOP case because this signal can not be caught.
This patch modifies stdin.c and stdout.c to no longer set O_NONBLOCK
for fd 0 and fd 1 if these fds are associated with a terminal
device. This is much easier and should do the job as well.
|
| |/
|
|
|
| |
This way doxygen issues a warning if the file/function/structure no
longer exists and a stale reference remains.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Done with
files=$(git grep -l 'Copyright (C) [0-9]\{4\}\(-2014\)* Andre Noll')
sed --in-place= -e 's/Copyright (C) \([0-9]\{4\}\)-2014 Andre Noll/Copyright (C) \1 Andre Noll/1' $files
In previous years we ran a similar script to set the second year in
the range to the current year. This is kind of silly, so let's get
rid of this useless information.
This commit replaces "Copyright (C) A-B" by "Copyright (C) A" in
all file headers, i.e. only the first year (A) is left in. Accurate
information including time stamps for each change can be obtained
from the git history.
|
| |
|
|
| |
All these headers get included from para.h.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The mail server on systemlinux.org was down for more than a week
lately, so let's use an alternative official address. This commit
changes all maan@systemlinux.org addresses to maan@tuebingen.mpg.de.
Most .c and .h files contain the email address in the copyright header,
so they must all be patched. Three other files contain the address
for a different reason:
* README lists email and git, gitweb and home page URLs
* configure.ac needs it for configure -h
* version.c contains it for the -V option of all commands
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The patch is large, but it's fairly straight forward: Instead of
a task pointer all ->pre_select() and ->post_select() methods now
receive the context pointer that was passed to the scheduler when
the task was registered. This allows to kill the public task_context().
Two pre_select/post_select functions are not directly called by the
scheduler: session_post_select(), generic_recv_pre_select(). These
are changed to receive a proper struct rather than a void pointer.
Note that generic_filter_pre_select() is not changed in this
manner because some filters do not provide a pre_select wrapper but set
task->pre_select to generic_filter_pre_select().
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the current implementation struct task is public so users of this
structure can mess with internal scheduler details as they please. This
has led to many bugs and questionable code in the past. This commit
is the first step to overcome this design mistake. At the end of this
patch series struct task can be made private to sched.c.
This commit introduces the following new public functions:
* task_register,
* task_context,
* sched_shutdown.
It also adds the new public task_info structure which carries the
information passed to the scheduler when a new task is registered. This
structure will stay public while struct task will become private once
all users have been converted.
task_register() is supposed to eventually replace register_task(). The
main difference of the two is that the new function returns a _pointer_
to a task structure which is allocated dynamically. Users are not
supposed to look at the fields of this pointer directly.
task_context() is a temporary helper which can be removed again at the
end of the series. Its sole purpose is to return the context pointer
which was passed at task register time as part of struct task_info.
The final new function, sched_shutdown(), deallocates the task
structures allocated during task_register() to cleanly shut down the
scheduler after all tasks have terminated.
All users need to be converted to the new API. This patch only converts
the stdout task though. The other tasks will be converted in subsequent
patches. The scheduler can tell if a task was registered using the
new API by means of the new ->owned_by_sched bit of struct task.
This boolean variable can also be removed after all tasks have been
converted.
Some users will need to query the exit status of a terminated
task. Hence we keep all tasks on the task list after ->post_select()
returned negative but call neither ->pre_select() nor ->post_select()
any more for such tasks. This leads to the concept of zombie tasks.
|
| |
|
|
|
|
|
|
|
| |
This year, we're really on time. The changes in this patch were
created by the following silly script:
files=$(git grep -l 'Copyright (C) [0-9]\{4\}\(-2013\)* Andre Noll')
sed --in-place= -e 's/Copyright (C) \([0-9]\{4\}\)-2013 Andre Noll/Copyright (C) \1-2014 Andre Noll/1' $files
sed --in-place= -e 's/Copyright (C) 2013 Andre Noll/Copyright (C) 2013-2014 Andre Noll/1' $files
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Was cooking for almost a month.
b99b4a stdin/stdout: Restore fd flags on shutdown.
Conflicts:
stdin.c
stdout.c
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
The stdin/stdout code should restore the flags to the old value to
avoid surprises, for example in shell scripts.
This changes stdin.c and stdout.c to save the old value. It is
restored in ->post_select when the buffer tree node is removed and
no more I/O takes place.
|
| | | |
|
| | |
| |
| |
| | |
It has no more users.
|
| |/ |
|
| |
|
|
| |
Better late than never.
|
| |
|
|
|
|
|
|
|
| |
This has turned out to be source for bugs. Deallocate everything
in btr_remove_node() hence making removing the node and freeing
its resources an atomic operation.
To avoid dangling pointers to freed btrn nodes, the argument of
btr_remove_node() is changed to to struct btr_node **btrnp.
|
| |
|
|
|
|
| |
This function is not only useful for non-blocking file descriptors,
so the name was misleading. Rename it to xwrite() for the lack of
a better name.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
This adds the #include statement for stdbool.h to para.h. This allows
to get rid of the individual includes in *.c.
More importantly, since all *.c files include para.h, booleans will
now be available everywhere so that we won't need to touch dozens of
files anymore whenever a boolean is added to a public structure.
|
| |
|
|
|
|
|
|
| |
This gets rid of quite some inclusions of <dirent.h> which are not needed
as most .c files do not deal with directories at all.
afs.c doe not mmap anything and needs nothing from sys/time.h so remove
these includes as well.
|
| | |
|
| |
|
|
|
|
|
|
| |
This removes the redundant 'max_size_bytes' argument of
* write_nonblocking(),
* send_queued_chunks(), and
* send_chunk(),
since it was set to 0 in all cases.
|
| |
|
|
| |
Hey, this is earlier than last year :)
|
| |
|
|
|
| |
para_client was the last user of the non-btr code in stdin.[ch]
and stdout.[ch]. Remove this unused code and update the documentation.
|
| | |
|
| |
|
|
|
|
| |
This allows to have a single function which can be called from
both the pre_select and the post_select methods of all receivers/
filters/writers and from the stdin and stdout tasks.
|
| |
|
|
|
| |
Also, close filter/receiver nodes in filter.c/recv.c instead of in
each post_select().
|
| | |
|
| | |
|
| |
|
|
| |
Not used yet.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
A wrapper for regcomp() that logs an error message if the regcomp() failed
and uses a return value according to the paraslash rules. Currently there is
only one user of regular expressions, is_v4_dot_quad(), which is converted
to para_regex(), but new callers will be added soon.
Unfortunately, this change made it necessary to include regex.h in all .c files
that use string.h. Clean up the order in which headers are included a bit while
we're at it.
Maybe I should rethink the rule "Only .c files shall include header files"...
|
| |
|
|
| |
This allows filters to change the output buffer size on the fly.
|
| | |
|
| |
|
|
| |
This one caused the first chunk of data to be swallowed.
|
| | |
|
| |
|
|
| |
It contained some outdated comments.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Get rid of sched->select_ret. Tasks shouldn't care about the
return value of the select call.
- Kill task->private_data. Use container_of() instead.
- Remove task->event_handler. It is never necessary and only
makes the code more convoluted. The scheduler unregisters
tasks as soon as either the pre_select or the post_select
functions return an error.
- Rename task->ret to task->error and get rid of a couple
of error fields in other structs that usually only contained
a copy of the task's error value.
This conversion likely introduces many bugs that have to be
shaken out in subsequent patches. Hopefully it will result in
less error-prone code in the long run.
|
| | |
|
| | |
|
| |
|
|
| |
Add documentation and an assertion for (num > 0).
|
| |
|
|
|
|
|
|
|
|
| |
This way it's possible to tell at a later time why
the receiver/filter/writer terminated.
This allows to increase the delay for reconnecting in case
the receiver failed to connect to para_server: Let the receivers
set the error value to -E_RECV_EOF in case a normal end of file
event occurred and check this value when calculating the restart
barrier.
|
