Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | For system calls, need to do a stack switch, as well as save the | Kanoj Sarcar | 2000-01-24 | 1 | -1/+3 | |
| | | | | s-registers, at, sp, ra. Use SAVE_SOME to save these registers. | |||||
* | Fix typo. | Ralf Baechle | 2000-01-23 | 1 | -2/+2 | |
| | ||||||
* | Add missing entry. | Ralf Baechle | 2000-01-22 | 1 | -1/+3 | |
| | ||||||
* | Merge with Linux 2.3.22. | Ralf Baechle | 2000-01-21 | 36 | -453/+800 | |
| | ||||||
* | Add {get,put}_user()-style functions that are protected against DBE. | Ralf Baechle | 2000-01-20 | 8 | -37/+359 | |
| | | | | | | | | | 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 corruption of $1 in exception handler for R4000 / R4400 SC and MC | Ralf Baechle | 2000-01-20 | 1 | -7/+8 | |
| | | | | versions. | |||||
* | Disable 32-bit compatibility for now until I actually checkin the code | Ralf Baechle | 2000-01-19 | 2 | -8/+4 | |
| | | | | into CVS. | |||||
* | - IOC3 driver now will panic when encountering a RX/TX PCI DMA error. | Ralf Baechle | 2000-01-17 | 43 | -208/+2099 | |
| | | | | | | | | | - 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 Baechle | 2000-01-17 | 1 | -1/+2 | |
| | ||||||
* | Passing of structures in ABI64 sucks, so we no instead pass a pointer | Ralf Baechle | 2000-01-17 | 1 | -2/+2 | |
| | | | | to struct pt_regs to exception handlers. | |||||
* | ARC fixes that are necessary because on Origin we juggle with both | Ralf Baechle | 2000-01-17 | 1 | -12/+20 | |
| | | | | | XKPHYS and CKSEG0 addresses. free_init_mem() now actually frees memory and doesn't corrupt memory. | |||||
* | Syscall handler fixes. | Ralf Baechle | 2000-01-16 | 2 | -11/+9 | |
| | ||||||
* | Fix memset which was misstreating certain argument cases. | Ralf Baechle | 2000-01-16 | 1 | -15/+20 | |
| | ||||||
* | 64-bit fixes for dump_tlb.c debug helper. | Ralf Baechle | 2000-01-16 | 1 | -19/+18 | |
| | ||||||
* | 64-bit fixes for kernel_thread(). | Ralf Baechle | 2000-01-16 | 1 | -1/+1 | |
| | ||||||
* | Initialize current_pgd such that vmalloc() can be used before the first | Ralf Baechle | 2000-01-16 | 2 | -7/+14 | |
| | | | | fork(). | |||||
* | Fix Makefile. | Ralf Baechle | 2000-01-13 | 1 | -12/+5 | |
| | ||||||
* | o R3000 fixes | Harald Koerfgen | 1999-12-08 | 2 | -9/+6 | |
| | | | | o R4x00 little endiam semaphores | |||||
* | Resurrect serial console. | Ralf Baechle | 1999-12-08 | 1 | -4/+4 | |
| | ||||||
* | Fix compilation errors, add .cvsignore files. | Ralf Baechle | 1999-12-06 | 3 | -3/+10 | |
| | ||||||
* | Merge with Linux 2.3.21. | Ralf Baechle | 1999-12-04 | 223 | -7521/+15992 | |
| | ||||||
* | The ARC code hasn't changed in quite a while, so commit it ... | Ralf Baechle | 1999-11-19 | 13 | -49/+68 | |
| | | | | | | | | Until I I merge with even newer kernel releases from Linus this code limits the MIPS64 kernel to a maximum of 64mb RAM on Indys. If I wouldn't do so, on startup kernel internal structures would collide with PROM memory. Solving this for real doesn't to make sense for now as a shiny new bootmem allocator is available in newer kernels. | |||||
* | Move various headers into a SGI specific subdirectory. These files | Ralf Baechle | 1999-10-21 | 13 | -41/+34 | |
| | | | | | need to be at the same place for MIPS32 and MIPS64. Make the semaphore code actually compile. | |||||
* | Some minor bugfixes: | Harald Koerfgen | 1999-10-17 | 4 | -6/+20 | |
| | | | | | | | o Fix braino in lib/Makefile o Export __wbflush() and release_tc_card() for modules o DMA isn't limited on DECstations o remove elf2ecoff and addinitrd while "make clean" | |||||
* | o R3000 fixes | Harald Koerfgen | 1999-10-12 | 4 | -9/+196 | |
| | | | | | o moved to original 2.3.19 NCR53C9x driver o CONFIG_ULTRIX_PARTION is working again and now big endian proof | |||||
* | Merge with 2.3.19. | Ralf Baechle | 1999-10-09 | 731 | -23521/+45752 | |
| | ||||||
* | Don't count VCEI/VCED if the proc fs support isn't choosen | Gleb O. Raiko | 1999-10-07 | 1 | -3/+5 | |
| | ||||||
* | Add necessary #ifdefs to compile cleanly in case the proc fs and the vt ↵ | Gleb O. Raiko | 1999-10-07 | 1 | -2/+6 | |
| | | | | support aren't choosen | |||||
* | Merge with 2.3.10. | Ralf Baechle | 1999-09-28 | 100 | -2111/+2908 | |
| | ||||||
* | Fix bug in tlbl/tlbs handlers that assumed the faulted entry already in tlb | Gleb O. Raiko | 1999-09-27 | 1 | -22/+28 | |
| | | | | | | | because tlb refill must occur before and store the tlb entry. On r3k it's not always true: miss on kseg2 doesn't generate tlb refill. Now modules work. As a side effect, the kernel runs faster because a vmalloced area is distrubuted across entire tlb. | |||||
* | o made CVS compile again for R5000 | Harald Koerfgen | 1999-09-18 | 3 | -5/+15 | |
| | | | | | | o dz fixes by David Airlie o small zs fix (init is fast again) o #ifdef in softirq.h fixed | |||||
* | Same fix for mainline. | Ralf Baechle | 1999-08-24 | 1 | -2/+2 | |
| | ||||||
* | Look ma - a tank has hit the MIPS sources ... | Ralf Baechle | 1999-08-21 | 53 | -938/+3120 | |
| | ||||||
* | More MIPS64 chainsawing. | Ralf Baechle | 1999-08-20 | 45 | -91/+3325 | |
| | ||||||
* | MIPS 64, day 2. | Ralf Baechle | 1999-08-19 | 2 | -34/+21 | |
| | ||||||
* | - Drop support for 32/32 fp register model | Ralf Baechle | 1999-08-18 | 40 | -170/+4862 | |
| | | | | | | | | | | | - Fix siginfo structure definition - Ptrace fixes, one of them was an uninitialized pointer. - Header file fixes. - Fixes to linker scripts required for new binutils. - MIPS64. Chainsaw edition. Btw, Harald, why does the kernel no longer compile ... | |||||
* | Baget updates. | Ralf Baechle | 1999-08-17 | 4 | -55/+38 | |
| | ||||||
* | Made implicit assumptions about the availability of some CPU capabilities on ↵ | Harald Koerfgen | 1999-08-13 | 4 | -10/+36 | |
| | | | | certain ISA levels explicitely configurable | |||||
* | Updates in preperation of EISA support. | Andrew R. Baker | 1999-08-11 | 1 | -4/+4 | |
| | ||||||
* | My proposal for non-generic kernels: | Harald Koerfgen | 1999-08-09 | 22 | -351/+199 | |
| | | | | | | | | | o only code for the configured CPU is compiled and linked (saves ~100k for R3000 kernels!) o removed a lot of indirect function calls o removed Ralf's "cowboy patch" o added sanity check for DECstations (print warning if the kernel is configured for the wrong CPU) | |||||
* | Make it link. | Ulf Carlsson | 1999-08-04 | 1 | -1/+2 | |
| | ||||||
* | Make it compile. | Ulf Carlsson | 1999-08-04 | 1 | -4/+4 | |
| | ||||||
* | The remaining R3000 changes. From now on the CVS will be R3000 aware. R3000 ↵ | Harald Koerfgen | 1999-07-26 | 7 | -58/+210 | |
| | | | | Indigo anyone? :-) | |||||
* | Due to the MIPS calling sequence pread(2) / pwrite(2) effectivly are | Ralf Baechle | 1999-07-11 | 1 | -3/+3 | |
| | | | | | | | | getting called with 6 arguments from userland. This also means that _every_ damn fucking piece of software that attempts to do syscalls to these functions in the absence of glibc is borken, pciutils just being one example. | |||||
* | Merge with Linux 2.3.9. | Ralf Baechle | 1999-07-05 | 61 | -210/+283 | |
| | ||||||
* | Dito for mainstream. | Ralf Baechle | 1999-06-27 | 1 | -46/+48 | |
| | ||||||
* | Merge, cleanup, typo fixes. | Ralf Baechle | 1999-06-23 | 1 | -2/+2 | |
| | ||||||
* | Merge with Linux 2.3.8. | Ralf Baechle | 1999-06-22 | 15 | -297/+326 | |
| | ||||||
* | Merge with Linux 2.3.7. | Ralf Baechle | 1999-06-22 | 98 | -2550/+15174 | |
| | | | | | | WARNING: 2.3.7 is known to eat filesystems for breakfast and little children for lunch, so if you try this on your machine make backups first ... | |||||
* | changed #include from <asm/ide.h> to <linux/ide.h> to make it compile again | Thomas Bogendoerfer | 1999-06-22 | 1 | -2/+2 | |
| |