summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Collapse)Author
2007-05-28Move the MSIZE, MCLSHIFT, MCLBYTES and the MCLOFSETThordur I. Bjornsson
mbuf constants from MD param.h to MI param.h. Besides being the same on every arch, things will most probly break if any arch has different values then the others. The NMBCLUSTERS constants needs to be MD though; ok miod@,krw@,claudio@
2007-05-28Flush pmap from tlb before freeing it. Makes Core2Duo boxes moreKenneth R Westerback
stable, but is not a fully-understood or final fix. From and ok art@
2007-05-28Fix comparison between signed and unsigned integer. Makes the randomMark Kettenis
panics during autoconf disappear on the v215. ok dlg@, otto@
2007-05-28pcn(4) works hereMartin Reindl
2007-05-28Remove #wt0 lines.Kenneth R Westerback
2007-05-28Remove the Archive/Wangtek cartidge tape driver, wt(4). All tapes areKenneth R Westerback
st(4). ok tom@ deraadt@ jason@
2007-05-27be more agressive on invalidation; perform invalidates on load* and unloadJason Wright
2007-05-27Locore UP can use the same code as MULTIPROCESSOR to access the GDTTom Cosgrove
once we're up and running; similar to what was just committed for machdep.c and pmap.c. ok art@ deraadt@
2007-05-27We don't need to special-case access to the GDT for UP or MP, sinceTom Cosgrove
ci_gdt is set correctly even when ! MULTIPROCESSOR. ok art@ toby@
2007-05-27remove two TODO items:Jason Wright
- map the GART page table uncached - disable table walk probes
2007-05-27add pmap_(un)map_nc_direct() macro's to match the non-nc versionsJason Wright
2007-05-27pagemove() is no longer used.Miod Vallat
2007-05-27Unbreak cdboot and pxeboot machine memory +/-; bad mpf@.Tom Cosgrove
ok deraadt@
2007-05-27We can now access curcpu() members through %fs instead of doing itArtur Grabowski
with the complicated and expensive macros. tom@ ok
2007-05-27Make sure the minaddr hint for uvm_km_suballoc() calls is always initialized.Miod Vallat
ok pedro@
2007-05-27Dynamically allocate irq's for pci devices on elroy(4).Mark Kettenis
ok martin@
2007-05-27implement WSDISPLAYIO_GETSUPPORTEDDEPTHS. ok miod@ uwe@Matthieu Herrb
2007-05-27shrink the kernel for code that shouldn't be used anyway, ok kettenis@Dale Rahn
2007-05-27Move powerpc to vm_page_md, 'throw it in' kettenis@Dale Rahn
2007-05-27Enable ips(4).Alexander Yurchenko
2007-05-27Print reserved memory size correctly.Miod Vallat
2007-05-27USB cleanup leftover. Diff from jsg@ tested by me.Marcus Glocker
2007-05-27Always use XKPHYS addresses to perform cache operations now, for consistency.Miod Vallat
2007-05-27Use the direct maps (cached on uncached) for bus_space mappings.Artur Grabowski
jason@ ok
2007-05-27- Redo the way we set up the direct map. Map the first 4GB of itArtur Grabowski
in locore so that we can use the direct map in pmap_bootstrap when setting up the initial page tables. - Introduce a second direct map (I love large address spaces) with uncached pages. jason@ ok
2007-05-27fix another use of MAXNR for rijndaelTed Unangst
2007-05-27back out bio, breaks dlg's sparc64 v215Todd T. Fries
prompted by/ok dlg@ deraadt@
2007-05-27When deciding whether to allocate a bounce buffer, we need one ifTom Cosgrove
we are going to read exactly at 1 MB (i.e. fix an off-by-one that is already correct in the amd64 version of this file). ok toby@
2007-05-26last arch specific GENERIC with bio, remove itTodd T. Fries
ok drahn@
2007-05-26Move cpu_info/curcpu to be mapped by the %fs segment. Extra input andTobias Weingartner
debugging by tom@, art@, kettenis@, and others. Testing by many others. ok art@, kettenis@, tom@
2007-05-26Remove nbuf count (was missed by pedro)Dale Rahn
2007-05-26tht worksDavid Gwynne
2007-05-26missed this, remove bio here for move to sys/conf/GENERICTodd T. Fries
ok drahn@
2007-05-26Dynamic buffer cache. Initial diff from mickey@, okay art@ beck@ toby@Pedro Martelletto
deraadt@ dlg@.
2007-05-26prodded by marco, enable bio everywhereTodd T. Fries
grudgingly ok deraadt@
2007-05-26Add re(4). From brad at comstyle dot com.Mark Kettenis
2007-05-26tyopsMiod Vallat
2007-05-26If booted from ADPT,2940U2B, root is on scsi.Dale Rahn
2007-05-26More comment typos from Diego Casati. Including winners like funtion, allmost,Kenneth R Westerback
oustside, seqencer, toghether, nessissary, etc.
2007-05-25"interupt" -> "interrupt" in various comments. Mostly from Diego Casati.Kenneth R Westerback
2007-05-25Edge cases can trigger a TLB miss exception instead of an invalid TLBMiod Vallat
exception on early R5000 revisions. Despite this bug being supposedly fixed in R5000 revision 2 onwards, it nevertheless occurs quite frequently on matthieu's revision 2.1 R5000. Servicing the TLB miss exception would cause a duplicate TLB to be inserted, which causes the processor operation to become unpredictable (but lethal to the kernel, ten times out of nine). More details about the problem can be found in: http://www.linux-mips.org/archives/linux-mips/2000-02/msg00040.html We work around the issue by checking for an existing TLB entry, and handling this as an invalid TLB exception (as it was intended to be), in this case. Unfortunately this causes a measurable 1% slowdown on ``safe'' processors, so we'll work on providing different tlb handler flavours in the near future to recover from this.
2007-05-25Define rw_cas. Trivially simple on amd64 with atomic_cas_ul.Artur Grabowski
miod@ ok
2007-05-25Unifdef TLB_TRACE and TLBTRACE, no functional change.Miod Vallat
2007-05-25"boundries" -> "boundaries" in various comments. Started by Diego Casati.Kenneth R Westerback
2007-05-25'Proto type' -> 'Prototype' when discussing functions in comments.Kenneth R Westerback
Inspired by similar diffs from Diego Casati in other files.
2007-05-25Comment typos and tidying. From Diego Casati.Kenneth R Westerback
2007-05-25Change the old slow and complicated TLB shootdown code to new, fast andArtur Grabowski
simple. This is basically the same code as on i386 and basically the same performance improvements. This change also includes code to delay the freeing of ptps until they have been properly shot. in snaps for a week, no problems reported.
2007-05-25Replace the overdesigned and overcomplicated tlb shootdown code withArtur Grabowski
very simple and dumb fast tlb IPI handlers that have in the order of the same amount of instructions as the old code had function calls. All TLB shootdowns are reorganized so that we always shoot the, without looking at PG_U and when we're shooting a range (primarily in pmap_remove), we shoot the range when there are 32 or less pages in it, otherwise we just nuke the whole TLB (this might need tweaking if someone is interested in micro-optimization). The IPIs are not handled through the normal interrupt vectoring code, they are not blockable and they only shoot one page or a range of pages or the whole tlb. This gives a 15% reduction in system time on my dual-core laptop during a kernel compile and an 18% reduction in real time on a quad machine doing bulk ports build. Tested by many, in snaps for a week, no slowdowns reported (although not everyone is seeing such huge wins).
2007-05-25rename cpu_dumpconf() to dumpconf(); no need to be different from other ↵Theo de Raadt
architectures
2007-05-23Add code to spin up secondary cpu's; still work in progress.Mark Kettenis
prompted by deraadt@ a few times already.