Rev |
Age |
Author |
Path |
Log message |
Diff |
8329 |
1268 d 6 h |
rgimad |
/kernel/trunk/ |
is_region_userspace: refactoring and fix |
|
8270 |
1275 d 13 h |
dunkaist |
/kernel/trunk/ |
kernel: Protect OS threads from terminating by sf18.2. |
|
8249 |
1279 d 12 h |
rgimad |
/kernel/trunk/ |
fix return value from .addr_error in sysfn 26.2 , 9 |
|
8248 |
1279 d 13 h |
rgimad |
/kernel/trunk/ |
fix prior |
|
8247 |
1279 d 13 h |
rgimad |
/kernel/trunk/ |
SysFn 18.13 security fix: now apps cannot destroy kernel memory via passing illegal address in ecx
Update docs on SysFn 18.13 |
|
8246 |
1280 d 7 h |
rgimad |
/kernel/trunk/ |
SysFn 9 security fix: now apps cannot destroy kernel memory via passing illegal address in ebx
Update docs on SysFn 9 |
|
8216 |
1286 d 6 h |
rgimad |
/kernel/trunk/ |
fix is_region_userspace (worked incorrectly for e.g 0xFFFFFFFF) |
|
8176 |
1292 d 4 h |
dunkaist |
/kernel/trunk/ |
Fix r8139: delay_ms was 10x slower when using HPET. |
|
8160 |
1296 d 10 h |
rgimad |
/kernel/trunk/ |
added is_region_userspace, fix sysfn 26.2 address security checks |
|
8158 |
1297 d 3 h |
rgimad |
/kernel/trunk/ |
SysFn 26.2 security fix: now apps cannot destroy kernel memory via passing illegal address to this sysfn |
|
8139 |
1298 d 6 h |
dunkaist |
/kernel/trunk/ |
kernel: Make delay_ms use HPET when available.
I got a laptop that has timer on channel 1 of PIT that doesn't tick. |
|
8119 |
1307 d 7 h |
dunkaist |
/kernel/trunk/ |
fix r8111: Map acpi_apic_base page before access. |
|
8111 |
1313 d 12 h |
dunkaist |
/kernel/trunk/ |
kernel: Split acpi.inc and hpet.inc from init.inc.
Also, replace some hardcoded constants with macros. |
|
8093 |
1322 d 23 h |
dunkaist |
/kernel/trunk/ |
Assert sizeof.APPDATA=256, use it instead of hardcoded 256 value. |
|
8092 |
1328 d 23 h |
dunkaist |
/ |
uefi4kos: Read config and kernel files from the disk.
* Now there is a config file with boot options, example included.
* Config, kernel and ramdisk are loaded from the disk, not compiled in.
* DEVICES.DAT file is also optionally loaded from the disk.
* Also, move the loader to /kernel/trunk/bootloader/uefi4kos. |
|
8091 |
1330 d 9 h |
dunkaist |
/ |
Add extended_primary_loader to autobuild.
* Build fat1x, fat32, cdfs, after_win loaders.
* Set use_lba=1 by default in fat32 loader.
* Build kolibri.raw: flash / hdd image.
* Add example config.ini file.
* Add an option (config file and screen) to not load ramdisk image.
* Add an option (config file only) to set /sys path.
Now you can boot without a ramdisk from a storage that is supported
by the kernel itself, i.e. without loadable drivers. Thus you can not
load the system without a ramdisk from SATA/AHCI and USB drives. |
|
8089 |
1332 d 13 h |
dunkaist |
/kernel/trunk/ |
kernel: don't use ramdisk if asked.
The blue screen is not aware of this option. |
|
8087 |
1332 d 14 h |
dunkaist |
/kernel/trunk/ |
Initialize APPDATA.{fd,bk}_ev fields for kernel threads.
This prevents crash in event.inc:RemoveEventTo because of
zero pointer dereference. |
|
8086 |
1332 d 14 h |
dunkaist |
/kernel/trunk/ |
Fix random boot issue: 'Failed to start first app'
Access to BOOT.dev variable was sometimes done with obsolete
offset from non-flat-kernel ages. If you are unlucky, the ramdisk
is not loaded and launcher app is not started. |
|
7965 |
1466 d 12 h |
hidnplayr |
/kernel/trunk/ |
Improved readability, no semantic changes. |
|