summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Cleanup.Ralf Baechle2000-01-2716-318/+181
|
* Cleanup.Ralf Baechle2000-01-272-2/+2
|
* Update.Ralf Baechle2000-01-272-14/+30
|
* Fix affiliation. From Geert Uytterhoeven.Ralf Baechle2000-01-271-1/+3
|
* Update.Ralf Baechle2000-01-272-8/+4
|
* Merge with Linux 2.3.23. The new bootmem stuff has broken variousRalf Baechle2000-01-27599-17064/+41654
| | | | | platforms. At this time I've only verified that IP22 support compiles and IP27 actually works.
* Allow scsi intrs to be serviced.Kanoj Sarcar2000-01-261-9/+30
|
* Start using the 2.3.40 qlogic driver with a couple of minor hacks toKanoj Sarcar2000-01-261-93/+169
| | | | work with 2.3.22.
* 1. Minor changes to start the scsi intrs going (untested).Kanoj Sarcar2000-01-262-2/+8
| | | | 2. Fix the irq range checking in request_irq/free_irq.
* 1. Setup the qlogic chips so that the driver decides to use io-mappedKanoj Sarcar2000-01-261-2/+31
| | | | | | | | | | style access instead of memory-mapped style access. Older qlogic drivers only support io-mapped style access, the 2.3.40 driver supports both; it is just easier to get things working with io-mapped access. 2. To achieve above, get CF0_WRITE_PCI_CFG to actually program the registers ... properly. 3. Hack CF0_WRITE_PCI_CFG so that it still only does real work in the above case ... else things hangs. To be investigated.
* Support for the NEC DDB Vrc-5074 evaluation board. Patches byRalf Baechle2000-01-2626-21/+1761
| | | | Geert Uytterhoeven (geert@linux-m68k.org).
* Similar to a bunch of other architectures, extend the range of possibleKanoj Sarcar2000-01-251-0/+1
| | | | | ioport values to prevent gratuitous check_region failures from qlogic driver.
* Remove junk files.Ralf Baechle2000-01-252-0/+0
|
* SN0ism: inw/outw used by scsi driver requires address swizzling.Kanoj Sarcar2000-01-251-7/+7
| | | | | inb/outb used by serial driver, doesn't need swizzling. May also need to update the outw_p/inw_p/outsw/insw if we have problems with them.
* mips_io_port_base needs to be set up for the asm/io.h macros to doKanoj Sarcar2000-01-251-1/+5
| | | | their work.
* Remove .align 13, it's no longer needed.Ralf Baechle2000-01-251-2/+0
|
* Add memory clobbers.Ralf Baechle2000-01-252-10/+18
|
* After discussion with Kanoj I set the value of ISA_DMA_THRESHOLDRalf Baechle2000-01-251-1/+1
| | | | back to 0xffffff.
* Add in scatterlist.h that is needed to compile qlogicisp.c. Mostly theKanoj Sarcar2000-01-251-0/+22
| | | | mips32 stuff, except that we do not support ISA devices on SN0.
* Fix the fix.Ralf Baechle2000-01-241-0/+3
|
* For system calls, need to do a stack switch, as well as save theKanoj Sarcar2000-01-241-1/+3
| | | | s-registers, at, sp, ra. Use SAVE_SOME to save these registers.
* Let's start the run on SMP. Here are the spinlocks.Ralf Baechle2000-01-232-4/+274
|
* The should have been remove by the 2.3.22 merge.Ralf Baechle2000-01-234-0/+0
|
* Fix typo.Ralf Baechle2000-01-231-2/+2
|
* Add missing entry.Ralf Baechle2000-01-221-1/+3
|
* Merge with Linux 2.3.22.Ralf Baechle2000-01-21173-11887/+14219
|
* Add {get,put}_user()-style functions that are protected against DBE.Ralf Baechle2000-01-209-37/+451
| | | | | | | | | Unlike the IRIX functions they only have one instruction overhead per access, don't involve subroutine calls or any spinlocks, so are implicitly threadsafe. Add PCI infrastructure for SN0 + the necessary fixups for IOC3 brokeness.
* Fix typo.Ralf Baechle2000-01-201-1/+1
|
* Fix corruption of $1 in exception handler for R4000 / R4400 SC and MCRalf Baechle2000-01-201-7/+8
| | | | versions.
* Disable 32-bit compatibility for now until I actually checkin the codeRalf Baechle2000-01-192-8/+4
| | | | into CVS.
* Define STD_TO_HZ() to make it compile.Ralf Baechle2000-01-191-1/+3
|
* Remove junk.Ralf Baechle2000-01-191-0/+0
|
* - IOC3 driver now will panic when encountering a RX/TX PCI DMA error.Ralf Baechle2000-01-1768-358/+2744
| | | | | | | | | - IOC3 driver does no longer use GFP_DMA which given the _very_ small number of available GFP_DMA pages might have deadlocked the system. - First cut of Origin support. Last minute change: Do no longer use ARC memory / MD hub memory configuration information but klconfig.h stuff. Simpler, faster, shorter. - Zillions of MIPS64 fixes.
* Add missing recalc_sigpending. From David Wragg (dpw@doc.ic.ac.uk).Ralf Baechle2000-01-171-1/+2
|
* Passing of structures in ABI64 sucks, so we no instead pass a pointerRalf Baechle2000-01-171-2/+2
| | | | to struct pt_regs to exception handlers.
* Optimize alignment of received buffers such that we don't take insaneRalf Baechle2000-01-171-6/+5
| | | | | piles of address error exceptions due to bad alignment any longer. This results in a considerable performance boost.
* ARC fixes that are necessary because on Origin we juggle with bothRalf Baechle2000-01-171-12/+20
| | | | | XKPHYS and CKSEG0 addresses. free_init_mem() now actually frees memory and doesn't corrupt memory.
* Syscall handler fixes.Ralf Baechle2000-01-163-13/+11
|
* Don't bother Origins with stupid floppy controller oub()'s.Ralf Baechle2000-01-161-4/+4
|
* - Bug report from Bill Earl (wje@sgi.com):Ralf Baechle2000-01-162-4/+4
| | | | | | Don't use the top 32kb of address space of 32-bit compatibility processes as certain instruction sequences might generate code outside the 32-bit compat address space.
* Fix warning.Ralf Baechle2000-01-161-2/+4
| | | | (IRIX people: the same typing bug also exists in IRIX.)
* Fix memset which was misstreating certain argument cases.Ralf Baechle2000-01-162-18/+23
|
* 64-bit fixes for dump_tlb.c debug helper.Ralf Baechle2000-01-161-19/+18
|
* 64-bit fixes for kernel_thread().Ralf Baechle2000-01-161-1/+1
|
* - Disabling promiscuous mode was broken.Ralf Baechle2000-01-161-23/+35
| | | | | - Use hash filtering for multicast packets. - Fix for small packets.
* Initialize current_pgd such that vmalloc() can be used before the firstRalf Baechle2000-01-162-7/+14
| | | | fork().
* R100000 cache operations.Ralf Baechle2000-01-162-0/+391
|
* Document configuration option for the SGI IP27 aka Origin 200 / 2000 /Ralf Baechle2000-01-161-0/+5
| | | | Onyx 2.
* Remove debugging hack which was setting the lowest octet of theRalf Baechle2000-01-151-1/+0
| | | | | Ethernet address to 0x42 in order to make the Linux Ethernet address different from the IRIX one.
* IOC3 Ethernet driver and a ton of de-irix-ized IRIX header files withRalf Baechle2000-01-1323-0/+7401
| | | | | hardware related definitions. For more Origin bits stay tuned on this channel ...
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-10 13:47:30 +0000