Jan Stoess
|
619d1425d8
|
- Added SMP synchronization of KDB via NMIs and an atomic variable
|
2007-12-07 19:00:30 +01:00 |
Jan Stoess
|
197b9459bb
|
- Don't zero terminate input buffer
|
2007-12-07 18:59:59 +01:00 |
Jan Stoess
|
87497f0396
|
- Make kdb object cpulocal, with selected static members being global
|
2007-12-07 18:59:43 +01:00 |
Jan Stoess
|
0e69cf3e1c
|
- Remove global KDB lock, x86 uses atomic variables an NMIs
|
2007-12-07 18:59:25 +01:00 |
Jan Stoess
|
d7f19ceb41
|
- Move reset function to init section
- Add a command to switch CPUs
|
2007-12-07 18:59:07 +01:00 |
Jan Stoess
|
66b39218cc
|
- Add a arch-specific spinlock, since the global kdb spinlock has been removed
|
2007-12-07 18:57:04 +01:00 |
Jan Stoess
|
e58dccbbda
|
- Add KDB's TCB to the list of known TCBs
|
2007-12-07 18:56:40 +01:00 |
Jan Stoess
|
004fb6daf7
|
- Make kdb_check_breakin more generic, since the hlt/input semantics need
it, too
|
2007-12-07 18:56:21 +01:00 |
Jan Stoess
|
cfdb7a6c5e
|
- Add an option to halt the processor while waiting for input
|
2007-12-07 18:55:56 +01:00 |
Jan Stoess
|
11ec7dcb9a
|
- Remove WORDSIZE_FORMAT, use %wx instead
|
2007-12-05 10:46:22 +01:00 |
Jan Stoess
|
c9713033cf
|
- Use %wx instead of WORDSIZE_FORMAT
|
2007-12-05 10:46:06 +01:00 |
Jan Stoess
|
3fdc0193bc
|
- Use %wx instead of WORDSIZE_FORMAT
|
2007-12-05 10:45:54 +01:00 |
Jan Stoess
|
58a0953b07
|
- Use %wx instead of WORDSIZE_FORMAT
|
2007-12-05 10:45:41 +01:00 |
Jan Stoess
|
0cba4397d5
|
- Remove smp_reference_bits on UP configurations
|
2007-12-05 10:10:26 +01:00 |
Jan Stoess
|
5e9a842028
|
- debug output fix
- remove double comment
|
2007-12-05 10:09:03 +01:00 |
Jan Stoess
|
f438849848
|
- Update config: allow exregs from pager
|
2007-12-05 10:08:19 +01:00 |
Jan Stoess
|
395e4267a5
|
- Update config: allow exregs from pager
|
2007-12-05 10:08:15 +01:00 |
Jan Stoess
|
73334e10e3
|
- Don't assert non-NULL top pdir, it could be that we want to find out that
there's no top pdir on that CPU
|
2007-12-04 12:20:42 +01:00 |
Jan Stoess
|
ea82e9ef2c
|
- Limit cpu no input to realistic values
|
2007-12-04 12:20:10 +01:00 |
Jan Stoess
|
d8ad550fe3
|
- Only lookup global id of a potential tcb if it points into the TCB area
|
2007-12-04 12:17:52 +01:00 |
Jan Stoess
|
4a5d77cfd9
|
- Remove debug output
|
2007-12-04 11:55:21 +01:00 |
Jan Stoess
|
4df2cfb207
|
- Added a catcher for smp_reference_bits on X64 -- at the moment, L4 doesn't
have any non-shared, global pages or superpages.
|
2007-12-04 11:51:55 +01:00 |
Jan Stoess
|
e926c2a924
|
- Implementation of smp_reference_bits for X32
|
2007-12-04 11:51:07 +01:00 |
Jan Stoess
|
54a03b8201
|
- Use an available ptab bit to signal that a pt entry is cpulocal (needed
for smp_reference_bits)
|
2007-12-04 11:50:55 +01:00 |
Jan Stoess
|
0acfcf461a
|
- Introduce smp_reference_bits with a real implementation
|
2007-12-04 11:50:30 +01:00 |
Jan Stoess
|
53d0f05474
|
- Make grubdisk application configurable
|
2007-12-04 10:37:48 +01:00 |
Jan Stoess
|
632928044b
|
- Make grubdisk application configurable
|
2007-12-04 10:37:37 +01:00 |
Jan Stoess
|
2dfea0dd5f
|
- Introduce local ID handling by default
|
2007-12-04 10:37:20 +01:00 |
Jan Stoess
|
a20be2e1f5
|
- Use l4test helpers to create and kill threads
|
2007-12-04 10:18:54 +01:00 |
Jan Stoess
|
39216f058c
|
- Added a variable determining if failed tests should bailout
- Zero mem before handing it out
|
2007-12-04 10:01:32 +01:00 |
Jan Stoess
|
babdca5eb9
|
- Enable -fno-stack-protector switch for GCC > 4.0. Ubuntu distros
apparently use stack smashing protection by default. Kudos to P.Perez
|
2007-12-03 11:26:42 +01:00 |
Jan Stoess
|
3e8444f46d
|
- Only update pdir_cache if it has not been set before
|
2007-12-03 11:26:16 +01:00 |
Jan Stoess
|
adf3a4a1fa
|
- Properly save and restore MRs when saving or restoring tcb resources
- Fix compabitility mode word size mask check
|
2007-12-03 11:25:46 +01:00 |
Jan Stoess
|
4761e538f2
|
- Enable -fno-stack-protector switch for GCC > 4.0. Ubuntu distros
apparently use stack smashing protection by default. Kudos to P.Perez
|
2007-12-03 11:24:40 +01:00 |
Jan Stoess
|
f07bfc5a20
|
- Newer Versions of grub-0.97 interpret the MULTIBOOT_AOUT_KLUDGE bit even
if the OS image is an ELF object. Remove that flag therefore. Kudos to L.
Vivier and P.Perez.
|
2007-12-03 09:05:15 +01:00 |
Jan Stoess
|
2673e96e6f
|
- Newer Versions of grub-0.97 interpret the MULTIBOOT_AOUT_KLUDGE bit even
if the OS image is an ELF object. Remove that flag therefore. Kudos to L.
Vivier and P.Perez.
|
2007-12-03 09:05:07 +01:00 |
Jan Stoess
|
268f326596
|
- Bugfix: pass frame as 1st argument
|
2007-12-01 13:59:58 +01:00 |
Jan Stoess
|
3b540840dc
|
- printing bugfix
|
2007-12-01 01:27:40 +01:00 |
Jan Stoess
|
d8f8efac11
|
- Bugfix: make sure that current->send_head is null, not to->send_head
- Bugfix: when being switched to from C-code, only save valid MRs
|
2007-12-01 01:27:22 +01:00 |
Jan Stoess
|
a660c27f8b
|
- Print SMP pagetable resource
|
2007-12-01 01:25:44 +01:00 |
Jan Stoess
|
520e1b2cd1
|
-When testing receive cancel, use a IPC version that potentially tests the
fastpath
|
2007-12-01 00:00:16 +01:00 |
Jan Stoess
|
a2abcc2a1b
|
- Bugfixed fastpath IPC: when switching to the fast path from C (e.g.,
return_from_ipc(), switch_to()) registers and MRs weren't saved properly
- Added not yet completely tested asm version of tcb resource handling
|
2007-11-30 23:57:57 +01:00 |
Jan Stoess
|
ec0dec1bbe
|
- When switching threads, mark dest->pdir_cache as clobbered
- Use a safer variant of putting tcb pointer into r11 when returning from IPC
|
2007-11-30 23:55:08 +01:00 |
Jan Stoess
|
b0c501b6d7
|
- Use macro when saving MRs
|
2007-11-30 23:53:40 +01:00 |
Jan Stoess
|
e06676547c
|
- Move retrieving kernel space to top
- Bugfix: x86_last_ip has size word_t
|
2007-11-30 22:48:09 +01:00 |
Jan Stoess
|
e161d019f7
|
- Bugfix: x86_last_ip has size word_t
|
2007-11-30 22:47:40 +01:00 |
Jan Stoess
|
ac85cb5468
|
- Add option to enter kdb on breakpoint
|
2007-11-30 22:47:26 +01:00 |
Jan Stoess
|
1895f4f73d
|
- Bugfix: on cancelled receive IPCs, we get activated via return_from_ipc(),
which doesn't put the TCB into esi. Refetch it therefore. Kudos to Aleksey
Mandrykin
|
2007-11-30 09:02:00 +01:00 |
Jan Stoess
|
236c3ada1d
|
- On transfer error, set our state to running (may have been locked_running)
on errorneous string copy IPC
|
2007-11-30 09:00:07 +01:00 |
Jan Stoess
|
b4170dff5a
|
- Bugfix reversed boolean arguments to setup_ipc_threads
|
2007-11-29 18:00:35 +01:00 |