Gregory Nutt
|
2caeead612
|
Fix an important bug in the watchdog creation logic
|
2014-11-10 18:15:26 -06:00 |
|
Gregory Nutt
|
fcc2a1bb4b
|
Fix incorrect comment
|
2014-11-07 15:54:24 -06:00 |
|
Gregory Nutt
|
2da26f7884
|
More fixes to tickless operation, alarm/mode code. From Brandon Warhurst
|
2014-10-23 07:13:57 -06:00 |
|
Gregory Nutt
|
a7a5c8cff3
|
Fixes to tickless operation code. From Brandon Warhurst
|
2014-10-23 06:41:31 -06:00 |
|
Gregory Nutt
|
928b0bda3d
|
Default for CONFIG_SCHED_HPWORK should be no
|
2014-10-14 16:56:37 -06:00 |
|
Gregory Nutt
|
179fabb019
|
Add description of work queues to the porting guide. Update comments
|
2014-10-14 10:21:18 -06:00 |
|
Gregory Nutt
|
c9563da832
|
Missing workqueue logic
|
2014-10-12 17:19:51 -06:00 |
|
Gregory Nutt
|
e702d9396b
|
Fix one missed name change from last big commit
|
2014-10-12 07:09:57 -06:00 |
|
Gregory Nutt
|
c9d7472d68
|
Change naming of HP work queue configuration varaibles to be symmetric with LP work queue naming
|
2014-10-11 17:03:44 -06:00 |
|
Gregory Nutt
|
1a27a4d4c0
|
Eliminate warnings
|
2014-10-11 17:02:35 -06:00 |
|
Gregory Nutt
|
b78fe7c270
|
Rename CONFIG_SCHED_USRWORK to CONFIG_LIB_USRWORK
|
2014-10-11 15:59:40 -06:00 |
|
Gregory Nutt
|
9292e3d9de
|
Decouple the user-space work queue from the kernel space work queues
|
2014-10-11 15:50:22 -06:00 |
|
Gregory Nutt
|
bb1d91bd83
|
Make building of low-priority work queue independent of also building the high-priority work queue
|
2014-10-11 14:43:24 -06:00 |
|
Gregory Nutt
|
c0b703ac73
|
Fix some missing conditional compilation that caused build failures in some configurations
|
2014-10-11 08:50:00 -06:00 |
|
Gregory Nutt
|
d2424ac65c
|
Fix a couple of problems introduced in the last commmits
|
2014-10-11 08:29:26 -06:00 |
|
Gregory Nutt
|
0218f01f12
|
If there mutliple low-priority worker threads, only one needs to perform garbage collection
|
2014-10-11 07:09:21 -06:00 |
|
Gregory Nutt
|
234f73cb09
|
Add logic for priority inheritance with multiple worker threads. How should this work? Tentatively, the all get reprioritized together. It would be best to just reprioritize the single thread doing the high priority work, but we don't know which that will be in advance
|
2014-10-10 17:47:39 -06:00 |
|
Gregory Nutt
|
4a4b3ac537
|
Add support for multiple low-priority worker threads
|
2014-10-10 16:24:50 -06:00 |
|
Gregory Nutt
|
438e3e1a90
|
Fix a few bugs introduced in the last checkin
|
2014-10-10 15:06:13 -06:00 |
|
Gregory Nutt
|
cf59a195ba
|
User-mode work queue logic should not disable interrupts
|
2014-10-10 14:52:04 -06:00 |
|
Gregory Nutt
|
16a3e83258
|
Add support for delays of different durations in work queue processing
|
2014-10-10 13:21:37 -06:00 |
|
Gregory Nutt
|
75e7a4dbfc
|
Minor changes to work queue timing logic
|
2014-10-10 12:27:11 -06:00 |
|
Gregory Nutt
|
b2cebaa9d4
|
Modularize starting of worker threads to better isolate individual initialization characteristics
|
2014-10-10 09:38:28 -06:00 |
|
Gregory Nutt
|
2015fd76e2
|
Fix some conditional logic in last work queue repartitioning change
|
2014-10-10 08:47:41 -06:00 |
|
Gregory Nutt
|
1afc9773ac
|
Decoupling work queue data structures. This is part of the preparation to support multiple low-priority worker threads
|
2014-10-10 08:35:58 -06:00 |
|
Gregory Nutt
|
6220256a09
|
Repartition work queue code. Move kernel-specific parts from libc/wqueue to sched/wqueue
|
2014-10-10 06:22:51 -06:00 |
|
Gregory Nutt
|
7f36c4e97b
|
Update some Kconfig comments; Add a upper limit on the lower priority worker thread for priority inheritance
|
2014-10-07 17:11:26 -06:00 |
|
Gregory Nutt
|
322f9f401c
|
Simplify how C source files are selected in the build
|
2014-10-07 07:42:36 -06:00 |
|
Gregory Nutt
|
d7843c8e24
|
Add support for priority inheritance on the low priority worker queue
|
2014-10-07 07:41:52 -06:00 |
|
Gregory Nutt
|
38ddbc774f
|
Cosmetic change to comments
|
2014-10-06 18:03:01 -06:00 |
|
Gregory Nutt
|
37cf4135c4
|
fs_initialize() is no longer weak
|
2014-10-06 11:45:35 -06:00 |
|
Gregory Nutt
|
d914f3ceec
|
Major structure of file system functions to better support asynchronous I/O. Respository should not be trusted until I have a chance to verify everything
|
2014-10-06 10:53:25 -06:00 |
|
Gregory Nutt
|
87eb1d47ed
|
Fix some recently introduced typos, build problems, and warnings
|
2014-10-05 16:58:52 -06:00 |
|
Gregory Nutt
|
652d3ed29d
|
Rename CONFIG_LIBC_AIO to CONFIG_FS_AIO since it is now an OS property
|
2014-10-05 15:44:43 -06:00 |
|
Gregory Nutt
|
31cdcde73a
|
Add definitions for SIGPOLL
|
2014-10-05 06:02:37 -06:00 |
|
Gregory Nutt
|
da29907ac9
|
Update TODO list + cosmetic changes to clock_systimer.c
|
2014-10-03 11:21:35 -06:00 |
|
Gregory Nutt
|
8d43a6bdcd
|
Fix build error with CONFIG_SYSTEM_TIME64
|
2014-10-03 10:20:46 -06:00 |
|
Gregory Nutt
|
71b574f26c
|
Repartition some message queue logic: sched/mqueue should have all mqueue knowledge; fs/mqueue should deal only with inodes
|
2014-09-30 08:03:39 -06:00 |
|
Gregory Nutt
|
fcfe877e96
|
Cosmetic update to comments
|
2014-09-29 16:22:21 -06:00 |
|
Gregory Nutt
|
1f2cc9f4fe
|
Complete re-implementation of mq_close
|
2014-09-29 15:33:34 -06:00 |
|
Gregory Nutt
|
584d0fe4ad
|
Complete re-implementation of mq_open()
|
2014-09-29 14:59:31 -06:00 |
|
Gregory Nutt
|
9e975a217d
|
Separate mqueue allocation logic from mq_open() and put it in sched/mqueue/mq_msgqalloc.c
|
2014-09-29 14:09:31 -06:00 |
|
Gregory Nutt
|
b0f80cc8db
|
Move mq_open.c, mq_close.c, and mq_unlink.c from sched/mqueue to fs/mqueue
|
2014-09-29 13:35:32 -06:00 |
|
Gregory Nutt
|
e3fa34681b
|
Convert mqueue structure for use in VFS as inode data; rename mqueue_inode_s; remove links, reference counts and name from mqueue structure. These will be replaced by VFS data. Remove g_msgqueues and mq_findnamed.c; these will be replace with VFS logic
|
2014-09-29 13:19:11 -06:00 |
|
Gregory Nutt
|
542e706d7d
|
Fix a perverse case where vfork() is called from a pthread. Still not recommended
|
2014-09-29 12:21:34 -06:00 |
|
Gregory Nutt
|
1dc9768c1a
|
Fix vfork(). Now that arguments are kept on the stack, the way that arguments are passed from parent to child in vfork() must change. This bug has always been present, but was not visible with the old strdup() way of passing arguments
|
2014-09-29 10:45:44 -06:00 |
|
Gregory Nutt
|
9a3f18db37
|
Fix missing header file for bug introduced yesterday
|
2014-09-29 07:48:45 -06:00 |
|
Gregory Nutt
|
8c958be5ab
|
Semaphore initialization is now only required if priority inheritance is initialized
|
2014-09-29 06:50:48 -06:00 |
|
Gregory Nutt
|
650a0d0615
|
Completes VFS-based named semaphore implemetation. Still a little buggy
|
2014-09-28 15:58:56 -06:00 |
|
Gregory Nutt
|
30694f064c
|
Add VFS-based sem_open() implementation
|
2014-09-28 14:35:17 -06:00 |
|