HelenOS-historic
HelenOS
HelenOS-doc
Català-Valencià – Catalan
中文 – Chinese (Simplified)
中文 – Chinese (Traditional)
Česky – Czech
Dansk – Danish
Nederlands – Dutch
English – English
Suomi – Finnish
Français – French
Deutsch – German
עברית – Hebrew
हिंदी – Hindi
Magyar – Hungarian
Bahasa Indonesia – Indonesian
Italiano – Italian
日本語 – Japanese
한국어 – Korean
Македонски – Macedonian
मराठी – Marathi
Norsk – Norwegian
Polski – Polish
Português – Portuguese
Português – Portuguese (Brazil)
Русский – Russian
Slovenčina – Slovak
Slovenščina – Slovenian
Español – Spanish
Svenska – Swedish
Türkçe – Turkish
Українська – Ukrainian
Oëzbekcha – Uzbek
Subversion Repositories
HelenOS-historic
(root)
/
kernel/
– Rev 1163
Rev
Go to most recent revision
|
Show changed files
|
Directory listing
|
RSS feed
Filtering Options
From rev
To rev
Max revs
Search history for
Show All
Clear current filter
Rev
Age
Author
Path
Log message
Diff
1163
6625 d 18 h
decky
/kernel/trunk/arch/
fix types
1162
6625 d 19 h
decky
/kernel/trunk/arch/
make the ppc ports generate proper 32b or 64b code
1161
6625 d 19 h
decky
/kernel/trunk/
remove AFLAGS from C argument list
1160
6625 d 19 h
decky
/kernel/trunk/
start ppc64 port
the code is currently a wild mixture of 32/64b code, will be fixed
1159
6625 d 19 h
jermar
/kernel/trunk/generic/
Replace list of all tasks with B+tree of all tasks.
1158
6625 d 20 h
jermar
/kernel/trunk/generic/
Replace list of all threads with B+tree of all threads.
Add function thread_exists() for querying existence of thread.
1156
6625 d 20 h
jermar
/kernel/trunk/generic/
Sleeping in a wait queue can be now interrupted with
waitq_interrupt_sleep().
1154
6629 d 17 h
jermar
/kernel/trunk/generic/src/adt/
Fix critical bug in hash table initialization.
Only half of the table was actually allocated and initialized.
1153
6629 d 22 h
jermar
/kernel/trunk/generic/src/synch/
Because of another race condition, futex_wakeup() needs to be able to allocate and initialize the kernel futex structure too.
1151
6629 d 23 h
decky
/kernel/trunk/arch/ppc32/src/mm/
ppc32: mark four first frames as unavailable
1150
6631 d 2 h
jermar
/kernel/trunk/generic/
Rename btree_node_left_sibling() to btree_leaf_node_left_neighbour()
and btree_node_right_sibling() to btree_leaf_node_right_neighbour()
as the left and right neighbours of a node needn't be its siblings.
1149
6631 d 12 h
jermar
/kernel/trunk/generic/include/mm/
s/B+-tree/B+tree/
1148
6631 d 12 h
jermar
/kernel/trunk/generic/src/
s/B+-tree/B+tree/
1147
6631 d 12 h
jermar
/kernel/trunk/generic/
Deploy B+tree in address space area management.
Change as_remap() to check for conflicts with other address space areas only when the area in question grows.
1145
6632 d 20 h
decky
/kernel/trunk/arch/ppc32/
align ppc32 framebuffer on 128K boundary
1144
6633 d 16 h
jermar
/kernel/trunk/
Completed B+-tree support.
Enable btree_remove().
Reorder some static functions and group them together.
Fix order of nodes in the leaf_head list.
1142
6634 d 0 h
jermar
/kernel/trunk/
Update B+-tree code.
The code is there, btree_remove() has not been tested yet.
(Fixes, if any, are to come later today.)
1141
6634 d 21 h
palkovsky
/kernel/trunk/generic/src/ipc/
Completed ipc_cleanup, it should be somehow integrated into
cleanup of task. The function can sleep.
1140
6635 d 16 h
jermar
/kernel/trunk/generic/
Change B+-tree from 2-3-4 tree to 2-3-4-5 tree by adding space for the fourth key.
This should make key removal easier.
1139
6635 d 16 h
jermar
/kernel/trunk/generic/include/
Fix comments.
1138
6638 d 18 h
jermar
/kernel/trunk/
CPU stacks must have two frames on ia64.
Make sure both thread stack frames are mapped in before_thread_runs_arch().
Take STACK_FRAMES into account during kernel memory layout initialization in main_bsp().
1137
6638 d 21 h
vana
/kernel/trunk/arch/ia64/include/
context_set patch in kernel for changed allocation of ia64 RSE and stack
1136
6639 d 15 h
jermar
/kernel/trunk/generic/src/adt/
Try to avoid splitting full B+-tree nodes by trying left or right rotation first.
(This improved memory consumption of this algorithm by some 40% - meassured on 101-item set).
1135
6640 d 12 h
decky
/kernel/trunk/
framebuffer code cleanup, support for non-standard scanline sizes
ppc32: get framebuffer parameters from boot loader
1134
6640 d 19 h
jermar
/kernel/trunk/generic/src/adt/
Change B+-tree to:
- store lesser keys in a key's left subtree
- propagate the bigger of medians (if there are two medians) when splitting a node
1132
6641 d 15 h
decky
/kernel/trunk/arch/ppc32/
kernel boot API change
1126
6641 d 22 h
palkovsky
/kernel/trunk/arch/mips32/src/
Fix bad order of EPC & STATUS change in syscall_shortcut.
1124
6641 d 23 h
jermar
/kernel/trunk/arch/
Small changes.
Make sparc64 and ppc32 compile again.
1122
6642 d 10 h
palkovsky
/kernel/trunk/arch/mips32/
Set up T9 to entry point on entry to userspace to facilitate better
PIC running.
1121
6642 d 13 h
jermar
/kernel/trunk/
Small textual changes.
1119
6642 d 13 h
palkovsky
/kernel/trunk/arch/
Fixed bad rflags masking in amd64.
Added TLS get address instruction emulation for mips.
1118
6642 d 13 h
jermar
/kernel/trunk/generic/src/syscall/
Change syscall table.
1117
6642 d 13 h
jermar
/kernel/trunk/generic/
Improved futexes. (kernel part)
1115
6642 d 23 h
jermar
/kernel/trunk/generic/src/proc/
Fix kcpulb() to change state of the migrated thread to Entering so that thread_ready() won't panic() on violated assertion.
Make print out of threads and tasks more dense again.
1112
6643 d 0 h
palkovsky
/kernel/trunk/
Added basic kernel infrastructure for ThreadLocalStorage(TLS) for
ia32(complete),amd64(complete),mips32(missing emulation of rdhwr instruction).
1109
6643 d 12 h
jermar
/kernel/trunk/
Basic futex. Prototype implementation.
1108
6643 d 12 h
jermar
/kernel/trunk/
Small PTE_* macros and SET_PTL0_ADDRESS macro changes.
1107
6643 d 12 h
jermar
/kernel/trunk/
Small atomic #include and typedef changes.
1106
6643 d 12 h
jermar
/kernel/trunk/arch/
Pass either -DCONFIG_PAGE_PT or -DCONFIG_PAGE_HT to the compiler.
1104
6643 d 17 h
jermar
/kernel/trunk/
Define atomic_t only once in atomic.h
Change the encapsulated counter type to long so that it supports negative values as well.