sambuc/codezero
1
0
Fork 0
mirror of https://github.com/drasko/codezero.git synced 2026-01-12 10:53:16 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity
506 Commits 1 Branch 0 Tags
97b15e6e3715ebf4a351022d5cc799a400190c86
Commit Graph

157 Commits

Author SHA1 Message Date
Bahadir Balban
8ffd4537ea Added newer platform files. 
The build can now be done with fine-grain conditions on how to determine
drivers, platforms, arch, subarch etc.
 2009-09-11 14:29:24 +03:00
Bahadir Balban
942865e6fc Working compilation, but adding variant_dir = BUILD it doesn't work 2009-09-11 13:55:46 +03:00
Bahadir Balban
ccd8a61a89 Added container physical regions to cml configuration 2009-09-04 11:41:44 +03:00
Russel Winder
399f1d530d Pull out the toolchains specification as a separate file and extend to cover the new options in the configuration file. 2009-08-30 13:47:16 +01:00
Bahadir Balban
bda0e064af Added new subarch, platform, cpu, driver variants for configuration testing. 
New files are an exact copy of pb926 files, but will help in testing different
configurations and compilations.
 2009-08-30 13:21:25 +03:00
Russel Winder
9e894274a3 Refactor the build to try and get as much material as possible out of SConstruct. 2009-08-28 14:10:24 +01:00
Russel Winder
03205855cc Generalize the configuration processing so that the top-level SConstruct only processes the ARCH symbol and prepares the PLATFORM and SUBARCH entries so as to write the config.h. Everything is now entered in the environment so the SConscript files can be responsible fo rthe processing. 2009-08-28 10:39:19 +01:00
Russel Winder
079b031616 Add check to ensure the source for the required device driver is actually present. 2009-08-28 09:55:55 +01:00
Russel Winder
187b70a829 Remove redundant comments. 2009-08-28 09:23:56 +01:00
Russel Winder
6ca0f2607a Switch to using a different strategy for building the kernel. 2009-08-27 16:43:47 +01:00
Bahadir Balban
117e3482b6 Capabilities for quantitative resources working. 
Status:
- Capability initialization is a bit hacky with dummy current etc.
- All container caps belong to the pager
- Tasks refer to their pager's capabilities for mutex allocation - Hacky.
- Kernel container keeps quantitative caps and memory caps in separate lists - Hacky.

These will all evolve and get fixed.
 2009-08-22 23:41:53 +03:00
Bahadir Balban
d3d072e47d Capability accounting for quantitative capabilities - First part done. 
Need to make sure accounting charges correct containers during init.
Therefore kernel resource spending must also be accounted for.
 2009-08-22 12:09:51 +03:00
Bahadir Balban
de087eb1f8 Clean up to kernel code, Added `used' field to capabilities 
With the addition of used field, capability structure can be
used for resource accounting over the capability.
 2009-08-21 12:10:34 +03:00
Bahadir Balban
dabc448c4e Added freeing of pager init memory, but this memory is currently not utilized. 2009-08-19 16:26:44 +03:00
Bahadir Balban
95b7b65c89 Pager virtual address pool bookkeeping added for all pager virtual addresses 
Previously virt_to_phys/phys_to_virt macros were used such that they did a
blind offset translation for creating a pager internal virtual address for
user mappings.

This is now changed such that a properly bookkeeped virtual address pool is
used which will avoid any clashes on the virtual space.
 2009-08-19 14:44:36 +03:00
Bahadir Balban
5bee43ae99 Merge branch 'newBuild' of /opt/c0-trees/codezero-russel into devel 
Conflicts:

	SConstruct
	src/api/SConscript
	src/arch/arm/SConscript
	src/generic/SConscript
	tasks/bootdesc/SConstruct
 2009-08-14 17:09:58 +03:00
Bahadir Balban
6a42491751 Fixed uninitialized local vars. 2009-08-14 12:54:54 +03:00
Bahadir Balban
c7404c7322 Fixed boot messages 2009-08-11 15:28:39 +03:00
Bahadir Balban
f82db61519 Added freeing/unmapping of boot memory 2009-08-11 15:01:27 +03:00
Bahadir Balban
f4d9520fef First working Codezero & POSIX Services with container/capability changes 2009-08-11 11:23:26 +03:00
Bahadir Balban
4f4532210a Fixed a bug with page_map size by dynamically allocating it. 2009-08-10 23:47:58 +03:00
Bahadir Balban
15b659eaf0 Removed any refs to global id pools. 
Also added container reference setup to thread creation
 2009-08-10 15:16:57 +03:00
Bahadir Balban
aa71161a04 Seems phys_to_page/page_to_phys macros are fixed in mm0 2009-08-09 17:59:32 +03:00
Bahadir Balban
02a3f1ac91 Pager works until end of init_physmem_secondary 2009-08-09 17:22:13 +03:00
Bahadir Balban
961b255257 Replaced kread and kmem_control syscalls with capability_control and container_control respectively. 2009-08-06 16:22:55 +03:00
Bahadir Balban
dccfff80a3 Removed commented out scheduler globals. 2009-08-06 15:02:42 +03:00
Bahadir Balban
e329a7c91a Removed kmalloc. Initialization path resolved, almost done. 2009-08-06 14:15:39 +03:00
Russel Winder
8b9b254f6c Make the coding style more consistent with PEP-8. 2009-08-05 11:15:26 +01:00
Bahadir Balban
519bfba8c7 Container/Pager/Capability initialization that works. 
Need to safely free boot memory and jump to first task's stack.
Need to test scheduler and all syscall entries.
 2009-08-04 13:56:11 +03:00
Bahadir Balban
7e8845abf8 Code that compiles until initialization of containers and pagers. 2009-08-02 23:43:14 +03:00
Russel Winder
9b2ff087d0 Add in all the amendments and new files needed to get the kernel building. 2009-08-01 17:50:59 +01:00
Bahadir Balban
82807c2f0a Code that works until after allocating pgd and pmd caches. 2009-07-29 20:01:40 +03:00
Bahadir Balban
723cf7bde9 Code that compiles and works up to initializing the first memcache. 2009-07-29 19:08:29 +03:00
Bahadir Balban
dd8f773f10 Some more progress on resource management and boot up. 2009-07-29 13:32:38 +03:00
Russel Winder
d40c76de84 Remove all the SCons files pending starting a new build system. 2009-07-28 14:47:00 +01:00
Bahadir Balban
2b0ea24e94 Added 2 more files. 2009-07-25 17:52:06 +03:00
Bahadir Balban
f7b768ee16 More changes for cleaner initialization and support for containers. 2009-07-25 17:44:29 +03:00
Bahadir Balban
0dd8918ae5 Remove all references to ARM registers in system call arguments. 2009-06-10 17:30:02 +03:00
Bahadir Balban
4e43d09325 wait_on_prepare and wait_on_prepared_wait with preemption considered 2009-06-10 15:48:34 +03:00
Bahadir Balban
7ef479733f Fix to l4_mutex_control for the unexpected sleeping with mutex held 
when preemption occurs after call to wait_on_prepare
 2009-06-10 15:41:30 +03:00
Bahadir Balban
6b3ddadcf5 Cleaned up code for userspace mutexes. 
- Issues: wait_on_prepare and wait_on_prepared_wait use cases
  are erroneous because they may sleep with mutex_control_lock
  taken if preemption occurs.
 2009-06-10 14:37:23 +03:00
Bahadir Balban
49d47abe97 Evaluated license issues, made some other enhancements. 2009-06-08 17:00:21 +03:00
Bahadir Balban
276b4643c6 Removed linux linked list dependency. 2009-06-02 13:19:17 +03:00
Bahadir Balban
4757f46f71 Minor comment fixes. 2009-06-02 10:26:21 +03:00
Bahadir Balban
13d469f659 Removed all unecessary prints from the mutex test and the kernel 
- removed (%d) Sleeping print from contended kernel mutexes.
- removed (%d) Waiting print from WAIT_EVENT used by the pager for suspending tasks.
- removed prints from the mutex_control syscall and user mutex test.
 2009-06-01 16:02:10 +03:00
Bahadir Balban
3b9cda968d Added mutex locking to kmalloc/kfree. 2009-06-01 15:18:44 +03:00
Bahadir Balban
1de77f1b93 Merge branch 'devel' into fix-kmalloc 2009-06-01 15:08:28 +03:00
Bahadir Balban
4206c6e3fc An extra mutex_unlock has been removed 2009-06-01 15:06:07 +03:00
Bahadir Balban
53f54f2328 Modified kmalloc to cater for dynamic memory allocation on a previously allocated size 2009-06-01 14:48:26 +03:00
Bahadir Balban
33200c92df Mutex implementation almost working. 
- Fixed a wrong instruction in mutex.S user library
- Added support for blocking lock/unlock
- Divided waiting into wait_on_prepare and wait_on_prepared_wait
  so that mutex_control lock is released after getting in the waitqueue.
- Declaring waitqueue on the stack should be done outside wait_on_prepare

Issues:
- Tests can be simplified for atomic data access instead of producer/consumer.
- kmalloc variable sized memory caches are not freed properly. Currently only the
  last slot can be freed, occupied correctly. it should be done in any slot, i.e.
  1, 2, 3, 4 instead of just 5.
- Need to add a mutex to kmalloc.
 2009-06-01 14:11:40 +03:00