summaryrefslogtreecommitdiff
path: root/sys/arch/i386
AgeCommit message (Collapse)Author
2009-01-23Add schsio(4), a driver for SMSC SCH311x LPC Super I/O ICs providingMichael Knudsen
environment sensors. These are found at least on VIA Epia SN18000G boards. Not enabled at 0x162e and 0x164e because probing these may not be safe according to kettenis (guess where my schsio is). Much help from kettenis. Input from const and fgs. ok kettenis
2009-01-20Install ACPI S3 resume trampoline code in a lowmem page. First partMike Larkin
of ACPI S3 suspend/resume support. This is for i386. Help/comments from art, toby, marco, jordan, kurt ok marco@, kurt@
2009-01-18enable atphy/ageKevin Lo
ok kettenis@
2009-01-17gdt_compact() was removed years ago, so kill the code and commentsPhilip Guenthe
that worked around its presence ok'ed by miod@
2009-01-16add next-server option to exampleOleg Safiullin
ok jmc@ stevesk@
2009-01-14int cpuspeed is global. ok markus@Alexander Yurchenko
2009-01-13Trivial lvalue abuse fixes, no binary changes.Alexander Yurchenko
ok millert@ tedu@
2009-01-13Fix lvalue abuse by removing a redundant cast.Alexander Yurchenko
ok kettenis@
2009-01-13Turn mpbios into a real device, such that people can disable it on crappyMark Kettenis
BIOSes. ok deraadt@ (for the amd64 bit, or the i386 bit, can't remember which)
2009-01-11Fix a sed bug in the makefile's depend target.Paul Irofti
What happened was that the output of mkdep was fed to a sed expression that trimmed a bit more than required and also failed to work when attempting to do make depend with pcc. Example: genassym_c.o: /tmp/genassym.whatever ../../../../../sys/param.h \ was changed to: assym.h: \ but what was intended was: assym.h: ../../../../../sys/param.h \ For the pcc -M output things were a bit different and after the make depend the genassym entry would still remain and make would fail. This affected all platforms except amd64 and sgi. Okay miod@.
2009-01-11Add additional values for the Atom FSB_FREQ bus clock MSR nowJonathan Gray
Intel actually document it publically.
2009-01-09Make sure we move an interrupt vector to the "lowest" priority level ifMark Kettenis
necessary. Fixes a nasty bug with establishing interrupts "late". Also make sure we skip the handler being removed when calculating the new "lowest" and "highest" priority level when disestablishing interrupts. Remove dmesg spam when freeing vectors. tested by many, ok marco@, beck@
2009-01-07Enable urlphy in addition to url.Jonathan Gray
Noticed by dlg on system that comes with url(4).
2009-01-07The E asm constraint refers to an immediate floating operandJonathan Gray
and isn't what we want here. ok weingart@ drahn@
2009-01-06- add and enable ipgphy(4) everywhere there's a stge(4) entry.Jasper Lievisse Adriaanse
(if the stge(4) entry is commented, so is the ipgphy(4) entry then). this allows ipgphy0 to attach to my stge0, which has a IC+ 1000A chip discussed with and ok jsg@, ok dlg@
2009-01-04Add support for CS5535 IDE adapted from the gcscide code in NetBSDJonathan Gray
originally written by Juan Romero Pardines. Thanks to Lars Kotthoff for testing on a HeroLogic HL-463 system. Because the CS5535 doesn't provide SMM emulated PCI access to set DMA/PIO modes or a real PCI IDE controller we have to resort to making MSR calls. As this is very MD specific, hide most of uglyness away in i386 MD pciide code. ok grange@
2009-01-03run(4), a driver for Ralink RT2700U/RT2800U/RT3000U USB 802.11agn devices.Damien Bergamini
the driver supports legacy 802.11abg modes only (no 802.11n yet.) WPA works, with hardware encryption/decryption. committed over an RT2800U (RT2870+RT2820). enabled on i386 and amd64 GENERIC only for now. thanks to Paul Lin at Ralink for letting us ship the firmware under an appropriate license. "man page looks ok" jmc@ "you should commit it" jsg@, ok deraadt@
2008-12-29fix compiling w/o apm(4)Oleg Safiullin
ok deraadt
2008-12-28No longer look for RSDP in memory marked as "ACPI Reclaim" in the BIOS memoryMark Kettenis
map. The standard doesn't tell us to look there, Linux and NetBSD don't look there either. It seems that some BIOSes use this memory as a scratch area to build the final ACPI tables which means we can accidentally find what seems like a valid RSDP there. Fixes PR 5847. ok krw@, marco@
2008-12-28Remove confusion between "raw" and "swizzled" interrupt pins. Remove someMark Kettenis
#ifdef __i386__ code that can go now things have been cleaned up.
2008-12-26use the exact same policy for choosing apic on all 3 types of kernel.Theo de Raadt
ok kettenis
2008-12-26If we have usable acpi, but chose to use amp instead for various reasons,Theo de Raadt
then mpbios should not be used because it is probably full of lies. ok kettenis, though he hates the variable name used.
2008-12-22Get rid of many arbitrary differences between the i386 and amd64 mpbios code.Mark Kettenis
Mostly KNF issues, some reorganisation of the code. ok dlg@, deraadt@, krw@
2008-12-18Don't set the global bit PG_G for kernel pmap low memory mappings. Use aKurt Miller
new function pmap_prealloc_lowmem_ptp() to setup kernel pmap ptp 0 without the PG_G bit set. This fixes the remaining reaper -> pmap_page_remove panics. With much diagnostic help from Art and Theo. ok deraadt@
2008-12-18use atomic operations to update ptes in pmap_unwire(). okay weingart@Kurt Miller
2008-12-11Kill the $ARCH prefix for isa_dma_cookie. With this change, the i386 andOwain Ainsworth
amd64 isa dma code is identical save for some formatting, and a slight difference in bus_dmamem_alloc. "Die x86_!" krw@.
2008-12-10When booting from a CD on i386/amd64, assume the root filesystemKenneth R Westerback
is on the CD rather than always asking for its location. Behaviour pointed out at OpenCON by Stephan Rickauer. Tested by Stephan. nick@ has no objections. ok deraadt@
2008-12-10both x86 platforms isa bus_dma implementations handle bus_dmamap_syncOwain Ainsworth
incorrectly. The spec (manpage) states that using two PRE or two POST ops together is entirely valid, but mixing pre and post is invalid. The way this was handled before with a switch statement meant that only individual commands actually would be recognised, so move to just checking the commands indidually using "if (op & $command)". Additionally, add a DIAGNOSTIC check and panic for the mixing of pre and post operations (this is done on several other architectures already). tested by several people; thanks! ok dlg@, kettenis@, "the diff made sense" deraadt.
2008-12-07Make acpiprt(4) check whether the current interrupt routing is "possible" andMark Kettenis
pick a new one from the list of possible routings if it isn't or if a pin is currently not routed. Delay re-routing interrupts until we establish a handler for it. This prevents us from messing with unused interrupt pins which may have fatal consequences (some machines spontaniously reboot). The heuristics for picking an interrupt from the list of possibe ones probably needs some tweaking still, but this makes several NVIDIA-based boards work much better than before. tested by many, ok marco@
2008-12-06revert all changes related to the mpsafe intr handler. i screwed up the commitTed Unangst
and even then it didn't work. we have higher standards than this.
2008-12-06some more trouble where i failed to account for oga's diffTed Unangst
2008-12-04*sigh* unbreak the tree.Owain Ainsworth
Some old diff snuck into this change.
2008-12-04Fix "fp_save ipi didn't" panic, and move i386/amd64 closer in the process.Tobias Weingartner
Positive test results by a handful of people. Ok kettenis@
2008-12-04Move the bus_dma generic functions out of the far-too-podgy machdep.c andOwain Ainsworth
into bus_dma.c (a-la amd64). Discussed with art and toby in coimbra. Ok mikeb.
2008-12-03Remove the x86 and i386 prefixes to the bus_dma types. It's really quiteOwain Ainsworth
pointless and just makes the code different for no reason. This moves i386 and amd64 bus_dma to being a lot closer to identical. suggestion to just remove the prefix instead of merge them from deraadt@. no objections art@, kettenis@, ok weingart@
2008-12-01terse is good, more readable; please commit deraadt@Todd T. Fries
2008-11-27Enable gpioow and gpioiic for runtime attachment, keep the previousMarc Balmer
synopsis as a comment (since it still works if needed). ok drahn
2008-11-26Enable Bluetooth (btd and btctl are not hooked into the build yet)Uwe Stuehler
2008-11-25we are heading quickly to a world where isp(4) will not fit, so let usTheo de Raadt
just accept the situation with a little sigh
2008-11-24Import btsco(4) from NetBSD to work on itUwe Stuehler
2008-11-24oops.Artur Grabowski
something snuck in here that wasn't supposed to be. noticed by yuo@
2008-11-24 - Remove the special case for one page pmap_remove, it's useless.Artur Grabowski
- Move the code that drops a ptp into a function and use that in both pmap_do_remove and pmap_page_remove. - In pmap_do_remove, instead of flushing the whole range from the tlb at the same time, shoot one PDE at a time. It will allow us to skip empty PDEs and reduces the amount of work one IPI has to do (although we'll get more IPIs). ok toby@ "I won't protest" deraadt@
2008-11-24die die dieTheo de Raadt
2008-11-24In pmap_write_protect() clear and set the protection bits atomically andKurt Miller
leave the rest alone. Also don't read *spte twice and compare results. feedback drahn@ okay art@ weingart@
2008-11-24In pmap_clear_attrs() use i386_atomic_clearbits_l() to only clear theKurt Miller
bits necessary. Also move up the pmap_map_ptes() call to before the diagnostic check. okay art@, weingart@
2008-11-23no more need to cope with pccomTheo de Raadt
2008-11-23pccom can finally die; ok kettenis dlg drahn, tested by okanTheo de Raadt
2008-11-22Move the drm drivers over from:Owain Ainsworth
vga1 at pci0 inteldrm0 at vga1 to vga1 at pci0 inteldrm0 at vga1 drm0 at inteldrm0 i.e. a similar scheme to audio(4) where the interface attaches on top of the wildly different drivers. This helps to clean up the code a lot (more is coming) and help me start to move drm to being essentially bus independent, which will help in the future.
2008-11-22Allow wired memory to be accounted on i386. This automatically changesMike Belopuhov
the policy for the mlock(2) on this architecture: all users are allowed to call mlock, while the limit is imposed by the current RLIMIT_MEMLOCK value. "makes sense" art, ok hshoexer (who was running with the same change for about 10 months).
2008-11-22The last parts of cpu_unidle. i386, amd64 and sparc64Artur Grabowski
In short, make cpu_unidle do what signotify used to do and make signotify use cpu_unidle. Also, include a cpu_unidle in need_resched, it won't change much right now but will be needed in the future.
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-15 04:18:52 +0000