| Rev |
Age |
Author |
Path |
Log message |
Diff |
| 4343 |
5484 d 19 h |
svoboda |
/branches/dynload/ |
Merged trunk changes up to r3925 to dynload. |
|
| 4339 |
5484 d 20 h |
svoboda |
/branches/dynload/ |
Merge trunk changes up to r3802 to dynload. |
|
| 3674 |
5629 d 18 h |
svoboda |
/branches/dynload/ |
Merge latest trunk changes to dynload. |
|
| 3149 |
5802 d 3 h |
svoboda |
/branches/dynload/ |
Merge trunk changes up to r3071 into dynload (broken) |
|
| 2927 |
5834 d 6 h |
decky |
/branches/dynload/ |
added dynload branch |
|
| 2141 |
6248 d 17 h |
jermar |
/trunk/ |
The Ultimate Solution To Illegal Virtual Aliases.
It is better to avoid them completely than to fight them.
Switch the sparc64 port to 16K pages. The TLBs and TSBs
continue to operate with 8K pages only. Page tables and
other generic parts operate with 16K pages.
Because the MMU doesn't support 16K directly, each 16K
page is emulated by a pair of 8K pages. With 16K pages,
illegal aliases cannot be created in 16K D-cache. |
|
| 2134 |
6249 d 21 h |
jermar |
/trunk/kernel/ |
Reworked handling of illegal virtual aliases caused by frame reuse.
We moved the incomplete handling from backend's frame method to
backend's page_fault method. The page_fault method is the one that
can create an illegal alias if it writes the userspace frame using
kernel address with a different page color than the page to which is
this frame mapped in userspace. When we detect this, we do D-cache
shootdown on all processors (!!!).
If we add code that accesses userspace memory from kernel address
space, we will have to check for illegal virtual aliases at all such
places.
I tested this on a 4-way simulated E6500 and a real-world Ultra 5,
which has unfortunatelly only one processor.
This solves ticket #26. |
|
| 2089 |
6307 d 3 h |
decky |
/trunk/ |
huge type system cleanup
remove cyclical type dependencies across multiple header files
many minor coding style fixes |
|
| 2071 |
6317 d 20 h |
jermar |
/trunk/ |
(c) versus (C) |
|
| 2054 |
6345 d 18 h |
jermar |
/trunk/kernel/ |
Fix important comment in kernel/arch/sparc64/src/proc/scheduler.c.
Improve framebuffer code.
Formatting and indentation fixes. |
|
| 2015 |
6356 d 20 h |
jermar |
/trunk/ |
Rework support for virtually indexed cache.
Instead of repeatedly flushing the data cache, which was a huge overkill, refuse to create an illegal address alias
in the kernel (again) and allocate appropriate page color in userspace instead. Extend the detection also to
SYS_PHYSMEM_MAP syscall.
Add support for tracking physical memory areas mappable by SYS_PHYSMEM_MAP.
Lots of coding style changes. |
|
| 2009 |
6361 d 19 h |
jermar |
/trunk/kernel/ |
Initial support for handling illegal virtual aliases on sparc64. |
|
| 1923 |
6410 d 20 h |
jermar |
/trunk/ |
Replace the old IRQ dispatcher and IPC notifier with new implementation.
Note that all architectures except for sparc64 are now broken
and don't even compile. |
|
| 1922 |
6412 d 4 h |
jermar |
/trunk/kernel/ |
More IRQ stuff.
Modify the IRQ hash table functions to support lookup based on inr and devno.
Add method member to irq_t. |
|
| 1920 |
6413 d 5 h |
jermar |
/trunk/kernel/ |
Move the new IRQ redirector to ddi/.
Add function for assigning unique device numbers.
Change sparc64/drivers/kbd.c to assign devno to keyboard. |
|
| 1919 |
6413 d 20 h |
jermar |
/trunk/kernel/ |
Prototypical implementation of new IRQ redirector in sparc64.
The new code can support shared IRQs in kernel (and multiple IRQs per device).
Userspace support is yet to be written.
The only architecture that uses this code is actually sparc64 only. |
|
| 1909 |
6424 d 19 h |
jermar |
/trunk/ |
Extend the OFW device tree node with a void pointer. The boot loader is
supposed to reset all pointers in all nodes. The kernel uses the pointer
to recognize that a particular device has already been visited and initialized.
Read interrupt mapping from the OFW device tree for FHC devices (z8530) and
EBUS devices (ns16550). In case of FHC devices, remove hardwired values from
the code and use only values read from the tree. FHC initialization is started
when a FHC device wants to map its interrupt. In case of EBUS devices, map
the interrupt to INO. Interrupt enabling in the interrupt controller for that
specific interrupt is not implemented yet. |
|
| 1904 |
6429 d 1 h |
jermar |
/trunk/kernel/ |
IPI/cross-call support for sparc64.
SMP on sparc64 is now fully supported. |
|
| 1896 |
6434 d 19 h |
jermar |
/trunk/ |
Convert sparc64 to detect keyboard and determine
its physical address by walking the memory representation
of the OpenFirmware device tree.
Add bus-specific functions that know how to apply the
"ranges" property to one component of the "reg" property.
Buses supported so far include FHC, EBUS and PCI. |
|
| 1875 |
6449 d 20 h |
jermar |
/trunk/ |
sparc64 work.
- Changes to enable userspace keyboard drivers.
- Fix z8530 initialization (i.e. clear any pending Tx interrupts).
- Experimental support for framebuffers with inverted colors. |
|