Rev |
Age |
Author |
Path |
Log message |
Diff |
2636 |
6102 d 10 h |
jermar |
/trunk/ |
Update comments wrt the previous commit.
Minor formatting fixes. |
|
2635 |
6102 d 12 h |
cejka |
/trunk/ |
Function ipc_connect_me_to sends 3 user defined arguments now.
One argument added also to ipc_forward_fast.
Fixed devmap and improved its test. |
|
2622 |
6110 d 17 h |
jermar |
/trunk/ |
Add mode argument to IPC forward.
This argument can be used to modify the way forward behaves. |
|
2621 |
6112 d 16 h |
jermar |
/trunk/uspace/ |
Modify the async framework to make use of all six syscall arguments.
Supply user-friendly macros as in previous cases. |
|
2620 |
6112 d 22 h |
jermar |
/trunk/ |
Be more deterministic when a user accidently uses fast version of IPC
call/answer instead of the full one and passes fewer arguments than required by
the recipient of the call/response.
and the recipient interprets arguments that
were actually not passed by the sender. |
|
2619 |
6114 d 10 h |
jermar |
/trunk/ |
Modify ipc_answer_*() to make use of all six syscall arguments. The recommended
means of answering calls is via the ipc_answer_m() macros (where m denotes the
number of return arguments) that automatically decide between the fast register
version or the slow universal version of ipc_answer(). |
|
2618 |
6114 d 22 h |
jermar |
/trunk/ |
Modify asynchronous IPC to make use of all six syscall arguments. The preferred
means of asynchronous communication is now via the set of ipc_call_async_m()
macros, where m is the number of payload arguments passed to the kernel. These
macros will automatically decide between the fast and the universal slow version
of ipc_call_async. |
|
2616 |
6115 d 15 h |
jermar |
/trunk/uspace/lib/libc/ |
usleep() should return int. Non-void functions really should return a value, so
fix sleep() to always return 0. |
|
2615 |
6115 d 19 h |
jermar |
/trunk/ |
Modify synchronous IPC to make use of all six syscall arguments. The preferred
means of synchronous communication is now via the set of ipc_call_sync_m_n()
macros, where m is the number of payload arguments passed to the kernel and n is
the number of return values. These macros will automatically decide between the
fast and the universal slow version of ipc_call_sync. |
|
2586 |
6135 d 9 h |
jermar |
/trunk/uspace/lib/libc/ |
Unify implementations of TLS variant I and variant II __alloc_tls() and
__free_tls_arch(). |
|
2570 |
6161 d 17 h |
jermar |
/trunk/uspace/lib/libc/generic/io/ |
Fix a bug in interpretation of precision when printing strings. Fix cstyle. |
|
2569 |
6161 d 19 h |
jermar |
/trunk/uspace/ |
Tiny refinement of if-else_if-else statement in vfs_register.c. |
|
2568 |
6163 d 1 h |
jermar |
/trunk/uspace/ |
Rename fibril_schedule_next_adv() to fibril_switch(). Rename
fibril_schedule_next() to fibril_yield(). Some fibril structures could be
uninitialized, set them to zero in fibril_setup(). For some fibrils, the stack
member can be NULL (e.g. every thread's first/main fibril); don't do free on
these stacks when cleaning up after a dead fibril. |
|
2564 |
6163 d 1 h |
jermar |
/trunk/uspace/lib/libc/ |
Support for time() function. |
|
2541 |
6168 d 19 h |
jermar |
/trunk/uspace/ |
Header cleanup.
Move off_t, size_t and ssize_t into sys/types.h.
Get rid off the non-standard and dummy types.h. |
|
2534 |
6175 d 15 h |
jermar |
/trunk/uspace/ |
In async_new_connection(), be more robust in the case that the function
is called by the server directly after its client accepts the
IPC_M_CONNECT_TO_ME call.
In fat.c main(), kill the "main" fibril and switch to the connection
fibril. |
|
2531 |
6177 d 12 h |
jermar |
/trunk/uspace/ |
VFS work.
Implement VFS_REGISTER part of the protocol in the FAT file system.
IPC work.
Rename ipc_data_send_accept() to ipc_data_receive() and ipc_data_send_answer()
to ipc_data_deliver(). Introduce ipc_data_send(). |
|
2530 |
6178 d 3 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
Pseudo thread -> fibril change in several comments. |
|
2522 |
6181 d 1 h |
jermar |
/trunk/uspace/lib/libc/ |
Add wrappers facilitating easy use of IPC_M_DATA_SEND calls and answers. |
|
2495 |
6245 d 16 h |
jermar |
/trunk/uspace/lib/libc/generic/io/ |
Formatting changes. |
|
2492 |
6251 d 14 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
Add comment clarifying how a fibril waiting for a message gets rescheduled. |
|
2490 |
6253 d 15 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
Improve comments in async.c |
|
2488 |
6255 d 9 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
Improve comments in async.c |
|
2486 |
6255 d 13 h |
jermar |
/trunk/uspace/lib/libc/ |
Remove duplicit and empty time.h from libc.
Move timeval functions from async.c to time.c. |
|
2485 |
6259 d 11 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
More fibril/psthread cleanup. |
|
2484 |
6259 d 12 h |
jermar |
/trunk/uspace/lib/libc/generic/ |
More psthread/fibril cleanup. |
|
2483 |
6259 d 16 h |
jermar |
/trunk/uspace/lib/libc/ |
Remove fibril_join().
We cannot guarantee our assumptions that easily.
This broken feature is removed instead of fixing because there are now users of it. |
|
2482 |
6260 d 7 h |
jermar |
/trunk/ |
Goodbye pseudo threads, welcome fibrils.
The renaming might still be incomplete. |
|
2481 |
6260 d 8 h |
jermar |
/trunk/uspace/lib/libc/ |
1. Implement simple joining of pseudo threads. Only one pseudo thread is allowed to join another pseudo
thread. The restriction is that both pseudo threads must be from the same thread (this is to simplify
synchronization). The joiner doesn't free anything from the joinee. It only gets its return value. Not
tested thoroughly yet.
2. Cleanup of a dead pseudo thread is done by the next scheduled pseudo thread. Not tested thoroughly yet. |
|
2479 |
6264 d 18 h |
jermar |
/trunk/ |
New, better-structured, directory layout for uspace. |
|
2471 |
6270 d 12 h |
jermar |
/trunk/ |
Greatly improve comments in the IPC layer.
Now I think I finally start to understand our IPC internals :-) |
|
2470 |
6274 d 10 h |
jermar |
/trunk/ |
Fix a double-free recently introduced in revision r2436.
Add a comment clarifying the life-cycle of the kernel_uarg structure.
Comment out the freeing of userspace thread stack while still running on it.
This should be fixed by implementing userspace version of thread_join(). |
|
2469 |
6274 d 11 h |
jermar |
/trunk/uspace/libc/generic/ |
Fix a bug in userspace thread_create() that confused success of the
syscall with failure. |
|
2445 |
6286 d 17 h |
decky |
/trunk/ |
initial merge of branches/fs
(not finished, huge cleanup is needed) |
|
2436 |
6288 d 12 h |
jermar |
/trunk/ |
Fix two memory leaks.
In kernel, kernel_uarg structure needs to be deallocated when a thread
with userspace context is destroyed.
In userspace, the return value of the SYS_THREAD_CREATE must be checked
for error conditions and in case of error, uarg and stack must be freed
up. |
|
2359 |
6300 d 21 h |
jermar |
/trunk/ |
Improve comments for the IPC subsystem.
Fix formatting and indentation. |
|
2230 |
6339 d 15 h |
jermar |
/trunk/uspace/ |
Remove printf() serialization from thread1 test in tester.
The agreement among developers seems to be that:
- there _must_ be a futex to serialize access to printf()
- there _must_ be pseudo thread serialization in printf()
- the best place for this is vprintf(), the one that goes to console |
|
2229 |
6339 d 18 h |
jermar |
/trunk/ |
Revert some of the changes introduced in revision 2209.
I think it is not correct to remove serialization of pseudo threads in printf_core.c.
With thread-level futex serialization, several pseudo threads running in one thread could easily deadlock
the task.
Add a dedicated futex serialization to thread1.c test only. |
|
2216 |
6341 d 12 h |
decky |
/trunk/ |
make thread ID 64 bit (task ID is 64 bit already)
cleanup thread syscalls |
|
2209 |
6341 d 14 h |
decky |
/trunk/uspace/ |
use futex instead of pthread serialization
synchronize only output to stdout
cleanup |
|