summaryrefslogtreecommitdiff
path: root/sys/arch/amd64
AgeCommit message (Collapse)Author
2009-12-09this does not even compileTheo de Raadt
2009-12-09add uhts(4).Matthieu Herrb
2009-12-09sloppy oga, tsk tsk, you are scaring usTheo de Raadt
spotted by ckuethe and must be in immediately before my build gets there..
2009-12-09Remove the clean gdt bit and leave the idt part in.Paul Irofti
Fixes most laptops out there on resume. Okay deraadt@.
2009-12-09add two new MD only pmap apis to amd64 and i386 (not to be used in MIOwain Ainsworth
code): pmap_flush_cache(vaddr_t, vsize_t) and pmap_flush_page(paddr_t) to flush the cache for virtual addresses and physical pages respectively using the clflush instruction. These apis will shortly be used by the agp bus_dma functions to avoid doing a wbinvd on each dmamap_sync. ok kettenis@, some comments from miod@
2009-12-09add cpufunc functions for the clflush instruction and the mfenceOwain Ainsworth
instruction. ok kettenis@ as part of a larger diff.
2009-12-09Detect the cache line size for the clflush instruction when we identifyOwain Ainsworth
the cpu. ok kettenis@ as part of a larger diff.
2009-12-03Stop using the linker script. Something is arranging memory just soTheo de Raadt
resulting in kernels starting up the other cpu's incorrectly on processors with very long pipelines, perhaps some bug instruction prefetch dependency in the mptramp.
2009-12-01Don't claim to know about nehalem until we can be sureJonathan Gray
the the msr paths are sane. ok deraadt@
2009-12-01The MD ACPI PSS paths are using MSRs they shouldn't beJonathan Gray
so don't try to do EST on machines we don't know the bus_clock on for now. ok deraadt@
2009-11-30KNFPaul Irofti
2009-11-30Bump up array size to prevent overflow when probing memoryCan Erkin Acar
on machines reporting > 32 memory regions. ok kettenis@
2009-11-29disable the isa wdc's because amd64 machines are much more likely toTheo de Raadt
be legacy-free machines built by Dell monkeys who then (it seems) reuse the io address space for something else. Touch it, and the
2009-11-29disable the isa wdc's because amd64 machines are much more likely toTheo de Raadt
be legacy-free machines built by Dell monkeys who then (it seems) reuse the io address space for something else. Touch it, and the machine resets. Great....
2009-11-29At suspend time there is no need for a fancy cr3 dance to decide ifTheo de Raadt
pmap_activate() should be called to force the kthread into the kernel pmap. If it is i386 it is lazy pmap it may not be in the kernel kthread, so it must be called so that pmap_activate can decide. If it is amd64 without lazy pmap, it must not be called. A comment is added to remind us about this in the future in case these assumptions change. ok kettenis
2009-11-29Reload mtrr state on all CPUs after updates. Seems to speed up X on MPMark Kettenis
systems, at least with Intel graphics. ok marco@, deraadt@
2009-11-27Move MB_LEN_MAX into the machine-independent sys/limits.h header,Philip Guenthe
rather than defining it separately for each architecture. Also set it to 4, to accommodate for future UTF-8 support (rfc3629). Diff by stsp, committing to catch the libc major bump ok kettenis@, guenther@
2009-11-26Reset the wakeup vector to ensure that we don't resume on reboot.Mike Larkin
Temporarily disabled for debugging purposes, but can be turned on if needed. ok deraadt@
2009-11-26do not pull acpi_wakecode.o into RAMDISKSTheo de Raadt
2009-11-26wdc@isa ok deraadtJonathan Gray
2009-11-26knf relative to i386Theo de Raadt
2009-11-26re-init the fpu like on i386Theo de Raadt
2009-11-26Really want isa.h when using #ifdef NISA > 0Theo de Raadt
2009-11-26Fix a stack problem on amd64, now the resume functions get called properly.Mike Larkin
Tested mlarkin on thinkpad x60, pirofti on dell d620, and deraadt on thinkpad T61. ok deraadt@, pirofti@
2009-11-25Make sure we get a clean gdt from the BIOS.Paul Irofti
Some vendors screw us up on resume giving back a dirty gdt which prevents us to go into protected mode. This makes sure the gdt is clean, its the only way to do this and its the only way to be sure we're clean on resume. This fixes quite a few laptops that didn't resume but rebooted or did other screwy things because of a dirty gdt. Worked with mlarkin@ for quite a few houres last night. Tested by many on both amd64 and i386. Okay deraadt@.
2009-11-24Save FPU state before suspend.Mike Larkin
ok deraadt@, kettenis@
2009-11-24Poke CR3 one last time before resuming. Suggested by deraadt@.Mike Larkin
ok deraadt@
2009-11-24enable uthum(4), and rearrange some nearby goopTheo de Raadt
2009-11-23bios* devices need to call their children on suspend/resume.Mike Larkin
ok deraadt@
2009-11-23Remove ACPI_SLEEP_ENABLED checks.Paul Irofti
This enables by default the suspend/resume paths in the kernel. Okay deraadt@.
2009-11-23oops, call config_activate_childrenTheo de Raadt
2009-11-23pchb must walk children tooTheo de Raadt
2009-11-23Use config_activate_children() for our cf_activate functionTheo de Raadt
ok mlarkin pirofti
2009-11-22Missed 4 MSRs on suspend/resume in previous version. Tested by piroftiMike Larkin
and myself, ok pirofti@
2009-11-22Check for ISA before calling the setup function.Paul Irofti
Suggested by and okay mlarkin@.
2009-11-22Bring amd64 closer to resume, while at it kill dead code. Okay deraadt@.Paul Irofti
2009-11-22Match the i386 behaviour and only care about bus_clock if weJonathan Gray
don't have the tables from acpi. This way we don't have to have knowledge of the cpu model to do speedstep in the acpi case. ok kettenis@
2009-11-15athn(4) has a future in here.Theo de Raadt
2009-11-14athn(4), a driver for Atheros 802.11a/g/n devices.Damien Bergamini
written from scratch based on the vendor driver for Linux (ath9k). AR9285 and AR9287 parts are 100% untested. only basic functionnalities are enabled for now. committed over an AR9281. "commit" deraadt
2009-11-08update description for iwn(4)Damien Bergamini
2009-11-04Get rid of __HAVE_GENERIC_SOFT_INTERRUPTS now that all our platforms support it.Mark Kettenis
ok jsing@, miod@
2009-11-01Remove bogus #define __HAVE_GENERIC_SOFT_INTERRUPTS. No code change.Mark Kettenis
2009-10-29iha(4), just like on GENERIC; from bradTheo de Raadt
2009-10-28rcsid[] and sccsid[] and copyright[] are essentially unmaintained (andTheo de Raadt
unmaintainable). these days, people use source. these id's do not provide any benefit, and do hurt the small install media (the 33,000 line diff is essentially mechanical) ok with the idea millert, ok dms
2009-10-27Bah didn't man to commit the GENERIC part just yet.Marco Peereboom
Reminded by deraadt.
2009-10-27Add driver for MPI2 SAS HBAs. This covers LSI MPT2 and Dell H200 HBAs.Marco Peereboom
IO works fine but it will remain disabled for now. From James Giannoules dlg: go go go
2009-10-19Correct a target name so that we don't rebuild vers.o (and thenPhilip Guenthe
bsd) unless some other object has changed. Rebuild and reinstall in /usr/src/usr.sbin/config/ after updating! "I like it" deraadt@
2009-10-07add support for the temperature sensor of VIA Nano and C7-M CPUs.Kevin Lo
some improvements suggested by jsg@ "commit" deraadt@
2009-10-06Allow MD limits for allocating PCI io and mem space and specify limits onMark Kettenis
i386 and amd64 to make avoid legacy ISA I/O registers. Fixes issues with ThinkPad T-series Dock II units as reported by sobrado@ ok miod@
2009-09-30Fix comments to match reality.Tobias Weingartner
Thank you Vladimir Kirillov <proger () hackndev ! com>