summaryrefslogtreecommitdiffstats
path: root/include/asm-mips64
Commit message (Collapse)AuthorAgeFilesLines
* Delete pte/pmd_free_kernel (unused), fix kptbl[] declaration, removeKanoj Sarcar2000-06-301-7/+1
| | | | silly check in pmd_alloc_kernel.
* Implement a functional vmalloc(). THe vmalloc range address translationsKanoj Sarcar2000-06-302-46/+18
| | | | | | | | are stashed in an array of page tables, starting from kptbl[]. The fast tlbmiss handler quickly checks to see if the faulting address is in the vmalloc range, and if so, it uses the translations in the kptbl to update the tlbs. Still to do: tlb invalid faults in the vmalloc range needs to be handled properly.
* Tlb miss handling updates: need to fill in the processor id only onceKanoj Sarcar2000-06-291-4/+15
| | | | during bootup.
* Add in the CONFIG_NUMA option for IP27s (distinct from DISCONTIGMEM whichKanoj Sarcar2000-06-271-0/+2
| | | | | might also be used by platforms with memory holes, but uniform access times). CONFIG_NUMA special code will probably make its way into generic kernel.
* Explain pgd/pmd allocation failure strategies. Create a global bad-pmd-table,Kanoj Sarcar2000-06-271-16/+9
| | | | | | | distinct from invalid-pmd-table for use during pmd allocation failure (so as to differentiate pgd_bad and pgd_none). The bad page table must have invalid pte entries to catch any user references to the range of virtual addresses it covers.
* Delete pmd_present/pgd_present. Explain pgd/pmd/pte allocation andKanoj Sarcar2000-06-261-7/+11
| | | | | initialization. Make sure to allocate only one page for the page table (prevent memory leaks), since only one page is freed up.
* Merge with Linux 2.4.0-test2.Ralf Baechle2000-06-251-0/+5
|
* Fix the number of open file resource limit. With the old limit, portmapKanoj Sarcar2000-06-201-1/+1
| | | | would core dump as soon as someone (rpc.mountd) wanted to talk to it.
* Fix struct stat64 in the 32-bit kernel and struct stat in the 64-bitRalf Baechle2000-06-201-14/+10
| | | | | kernel to match each other and the the glibc definition. The glibc part of this change has been sent to Andreas.
* Make it compile again.Ralf Baechle2000-06-201-2/+0
|
* R3000 cache handling. flush_icache_page now actually flushesRalf Baechle2000-06-202-7/+6
| | | | something.
* Merge with 2.4.0-test1-ac21 + pile of MIPS cleanups to make mergingRalf Baechle2000-06-191-25/+26
| | | | | possible. Chainsawed RM200 kernel to compile again. Jazz machine status unknown.
* Merge with 2.4.0-test1.Ralf Baechle2000-06-162-71/+110
|
* When doing kernel text replication, make sure we do not scribble onKanoj Sarcar2000-06-131-0/+1
| | | | | top of the replicated kernel text thinking it is unused memory. This lets us get to multiuser on a replicated kernel text system.
* Have the slave bootstrap code actually look at the text and data serverKanoj Sarcar2000-06-091-2/+2
| | | | | nasids that the master nasid found for it. Sanitize the replication procedures.
* Add in the hooks to do kernel text replication. Now start debugging thisKanoj Sarcar2000-06-091-0/+2
| | | | code.
* Each node needs to have a structure that the master nasid fills up toKanoj Sarcar2000-06-092-0/+31
| | | | | tell it where to get the kernel data from (ie, which nasid holds the kernel data). Remove debug cruft from head.S.
* Need to have the nasid of the boot/master node for future uses.Kanoj Sarcar2000-06-091-0/+2
|
* SMP Mapped kernel fixes to go along with the new layout. Use properlyKanoj Sarcar2000-06-081-0/+3
| | | | | massaged bootstrap address so that the PRM launches the slaves into legal code.
* Mapped kernel changes: create elf sections as close as possible toKanoj Sarcar2000-06-081-0/+15
| | | | | IRIX, to get the same kind of loader behavior. Read comments in mapped_kernel.h for more details.
* Get the mapped kernel working for SMP kernels. While launching slaves,Kanoj Sarcar2000-06-062-1/+25
| | | | | pass the proper address for the slave bootstrap routine based on where the kernel is compiled at.
* Get some type definitions into their rightful places for wider usage.Kanoj Sarcar2000-06-062-0/+3
|
* Make __kernel_caddr_t32 32-bitUlf Carlsson2000-06-061-1/+1
|
* Most of the kernel arrays/data structures and low level proceduresKanoj Sarcar2000-06-022-3/+3
| | | | | now accept the kernel assigned cpuid (instead of the prom assigned id) as input/index. The only exception is the early boot up code.
* Fix some type definitions, add a new field in the per-processor dataKanoj Sarcar2000-06-022-1/+2
| | | | | area to store the PROM assigned cpuid that is used by most of the IP27 kernel and prom data strucutures.
* Make __kernel_clock_t 64 bit.Ralf Baechle2000-05-261-1/+1
|
* This should resolve most of the offset.h issues which popped up lately, I hope.Harald Koerfgen2000-05-262-5/+6
| | | | | | | | | * #include <asm/offset.h> #ifdef'd: init/main.c and $(arch)/tools/offset.c compile without include/asm-$(ARCH)/offset.h. * include/asm-$(ARCH)/offset.h is now depending on $(TOPDIR)/.config: offset.h is beeing rebuilt when .config changes. Please don't count on that, makedep doesn't do it's job very well in this case and objects depending on offset.h aren't neccesarily rebuilt :(
* First cut of TLB handlers in assembler. I'm not usingUlf Carlsson2000-05-251-2/+8
| | | | | | | | | the context register as it should be used, but let's look into that later. If there is a problem with the code it will crash after right after freeing unused kernel memery. I have this code tested on both UP and SMP though.
* Further stresstesting revealed weaknesses in my first proposal.Harald Koerfgen2000-05-241-0/+1
| | | | This one survives "make distclean" and "make dep clean vmlinux".
* no real solution for the offset.h-chicken-egg-problem, but this does all I wnatHarald Koerfgen2000-05-241-0/+0
|
* HousekeepingHarald Koerfgen2000-05-232-99/+1
|
* Update defconfig files.Ralf Baechle2000-05-231-0/+47
|
* Implement a more dynamic method of associating IRQs with PCI devices.Kanoj Sarcar2000-05-191-2/+1
| | | | | Instead of encoding bus/slot numbers in the IRQ, have seperate arrays to store that information.
* irq.h: With the current scheme of encoding the bus number into theKanoj Sarcar2000-05-182-5/+5
| | | | | | | pci irq, we need to bump up NR_IRQs to handle more than 4 PCI busses. io.h: Port numbers are really "unsigned long", since they need to contain the nasids too. ip27-pci.c: Encode the nasid into the port number for isp1020 for pio.
* Some cleaning of the support for machines with HZ other than aRalf Baechle2000-05-161-0/+2
| | | | | architecture specific standard value. Not entirely at the point where we can send this to Linus.
* Fix SOCK_ constant definitions for all architectures. Patch sent toRalf Baechle2000-05-161-15/+14
| | | | | Linus. Our diffs relativ to Linus' latest and greatest is getting fairly small now.
* Remove references to dead flag ASYNC_IOC3.Ralf Baechle2000-05-161-1/+1
|
* Move the intercpu intr irq numbers to the lowest possible. AssignKanoj Sarcar2000-05-151-8/+5
| | | | | | | all PCI irqs above that. Fix pci_map_irq to not just have 2 bits for slot number, else irq numbers will not be unique. Include the bus number as part of the irq for now, so that scsi controllers on various PCI busses get unique irq numbers.
* Fix copyright message. Linus flamed me on these messages so pleaseRalf Baechle2000-05-153-48/+31
| | | | | make sure for the future that the (C) notice is ok. ``unpublished proprietary information of Silicon Graphics'' isn't good ...
* Merge with Linux 2.3.99-pre7 and various other bits.Ralf Baechle2000-05-124-1/+6
|
* Rudimentary nmi support to be able to do simple debugging on SMPKanoj Sarcar2000-05-121-0/+131
| | | | machines.
* SMP FPU management: similar to what sparc does, no lazy fpu contextKanoj Sarcar2000-05-112-0/+25
| | | | | | switching in SMP mode, use PF_USEDFPU to determine whether a program used the fpu in the last time quantum and so needs the fpu context to be saved during context switch.
* Some infrastructure for supporting multiple pci busses on origin200/2000.Leo Dagum2000-05-113-2/+14
| | | | | | | nasid and widget id get incoded in pci_dev->irq. pcibr_setup() now does some rudimentary probing for bridge widgets hanging off its xbow. We're not seeing interrupts from remote devices at the cpu, so mscsi card's are turned off for now.
* Fix the way we use the watchlo/watchhi registers to get to the taskKanoj Sarcar2000-05-041-1/+3
| | | | | struct of the currently executing thread: handle sign extension on the watchlo register.
* Fix NUMA compile problem.Kanoj Sarcar2000-05-031-2/+0
|
* Merge with 2.3.99-pre6.Ralf Baechle2000-04-281-2/+2
|
* sigprocmask needs to be 32-bitized.Kanoj Sarcar2000-04-271-0/+1
|
* We use the ST0_CU0 bit to determine whether we are coming intoKanoj Sarcar2000-04-252-17/+14
| | | | | | | | | | kernel mode for the first time from user mode (on intr/exception/ syscall etc). If so, we need to set the sp to point to the kernel stack. On UP kernels, the kernel stack pointer is stored in the global variable "kernelsp". For SMP kernel, the physical address of the current task structure is stuffed into the watchlo/watchhi registers, so on first entry into the kernel, we need to munge this value properly to setup the sp register.
* Fix the init_new_context code not to zap the percpu tlbpid array whenKanoj Sarcar2000-04-241-19/+12
| | | | | | doing a clone-vm operation. Also, the intercpu tlbflush code now properly does its job by flushing the tlbpid only on the current processor, and not on all.
* Change all instances of __SMP__ to CONFIG_SMP and includeRalf Baechle2000-04-237-11/+21
| | | | <linux/config.h> where necessary.
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-14 07:02:07 +0000