summaryrefslogtreecommitdiff
path: root/sys/arch
AgeCommit message (Collapse)Author
2007-10-13Enable interrupts in secondary processors before invoking cpu_switchto(),Miod Vallat
rather the expecting it to do this for us.
2007-10-13It is no longer necessary to fiddle with spl in cpu_idle_{enter,leave} nowMiod Vallat
that proc_trampoline has been fixed.
2007-10-13Be sure to spl0() in proc_trampoline, so that kernel threads start at IPL_NONE.Miod Vallat
2007-10-13There is no need to fiddle with spl in cpu_idle_{enter,leave}, actually.Miod Vallat
2007-10-13Remove leftovers art forgot to prune...Miod Vallat
2007-10-13Fix cpu_exit() comments to be more closer to reality.Miod Vallat
2007-10-13Do not splhigh() before invoking sched_exit(), sched_exit() will do it better.Miod Vallat
2007-10-13Various typos in comments; Joel SingMiod Vallat
2007-10-11enable puc(4) for GENERIC on armishStuart Henderson
ok deraadt@
2007-10-10prepare for the futureTheo de Raadt
2007-10-10GENERIC.MP dirs will show up here some timeTheo de Raadt
2007-10-10Disable PSL_POW code for now in cpu_idle_cycle, because it does not yetTheo de Raadt
do the cache flush / instruction cache disable / tlb invalidate dance. For some reason in a switchto world this suddenly matters or machines become really slow running NFS... huh? figured out with kettenis
2007-10-10Make context switching much more MI:Artur Grabowski
- Move the functionality of choosing a process from cpu_switch into a much simpler function: cpu_switchto. Instead of having the locore code walk the run queues, let the MI code choose the process we want to run and only implement the context switching itself in MD code. - Let MD context switching run without worrying about spls or locks. - Instead of having the idle loop implemented with special contexts in MD code, implement one idle proc for each cpu. make the idle loop MI with MD hooks. - Change the proc lists from the old style vax queues to TAILQs. - Change the sleep queue from vax queues to TAILQs. This makes wakeup() go from O(n^2) to O(n) there will be some MD fallout, but it will be fixed shortly. There's also a few cleanups to be done after this. deraadt@, kettenis@ ok
2007-10-10malloc+bzero -> malloc+M_ZERO. Use 'malloc(sizeof(*p) ...' idiom.Kenneth R Westerback
This completes my bzero/memset(,0,) -> M_ZERO sweep of the tree. ok kettenis@
2007-10-10Change argsize from size_t to short, since it only stores a shortRay Lai
value. Fixes a format string error. Pointed out by deraadt, OK miod.
2007-10-08Align the suspend saved data to a cache line boundary, apparently preventsMiod Vallat
issues upon resume.
2007-10-08A few trailing bzero/memset -> M_ZERO occurrences, cast removal andKenneth R Westerback
size(*p) usage. This should be the last of the simple malloc() changes.
2007-10-08A few trailing bzero/memset -> M_ZERO occurrences, cast removal andKenneth R Westerback
size(*p) usage.
2007-10-08isa0 at amdpcib? got skipped, sighTheo de Raadt
2007-10-08enable spdmem where it mattersTheo de Raadt
2007-10-08More simple memset(,0,) -> M_ZERO changes. In this batch move toKenneth R Westerback
size(*p) as the first malloc() parameter where p is declared locally and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c even though there is no obvious bzero or memset nearby.
2007-10-08Don't skip EEPROMs in ofw i2c scan,Jonathan Gray
match on sparc64 SPD name in spdmem. Tested on a V120. ok deraadt@
2007-10-07Add a driver, amdpcib(4), for the AMD 8111 series LPC bridge and HPET writtenMarc Balmer
by mickey, man page by me. Help, suggestions by Theo and jmc. Enable this driver and glxpcib(4) in all RAMDISK kernel, but glxpcib(4) does not provide the watchdog timer in the RAMDISK kernels. Prodded by deraadt. E
2007-10-07Add support for the PCI-ISA bridge, timecounter, and watchdog timerMarc Balmer
found in the CS5536 companion chip to the AMD Geode LX-800 processor. glxpcib(4) driver written by mickey, with some tweaks by me. Manpage nits by jmc. Tested on the Soekris net5501 (by me) and the Alix 2a2 (by earxtacy@dixinet.com) ok kettenis
2007-10-06Simpliest memset(,0,) -> M_ZERO changes. One (caddr *) cast removal,Kenneth R Westerback
otherwise just adding M_ZERO to malloc() and removing the immediately adjacent memset(,0,).
2007-10-06Some archs used memset() rather than bzero(). So duplicate diffKenneth R Westerback
previously applied to other archs deleting a memset() this time. e.g. - if ((mapstore = malloc(mapsize, M_DEVBUF, - (flags & BUS_DMA_NOWAIT) ? M_NOWAIT : M_WAITOK)) == NULL) + if ((mapstore = malloc(mapsize, M_DEVBUF, (flags & BUS_DMA_NOWAIT) ? + (M_NOWAIT | M_ZERO) : (M_WAITOK | M_ZERO))) == NULL) return (ENOMEM); - memset(mapstore, 0, mapsize);
2007-10-06Provide <machine/lock.h> on hppa64 too.Mark Kettenis
2007-10-06Simple single-processor only mutex implementation; cloned from hppa.Mark Kettenis
2007-10-06Remove unused variable.Mark Kettenis
2007-10-06Oops. Forgot to do FREE -> free when I did MALLOC -> malloc.Kenneth R Westerback
2007-10-03MALLOC+bzero -> malloc+M_ZERO.Kenneth R Westerback
In ip_esp.c all allocated memory is now zero'd in the "malloc(sizeof(*tc) + alen ..." case. The +alen memory was not initialized by the bzero() call. Noticed by chl@. "Looks good" art@ "seems ok" chl@
2007-10-03Add pcb_flags member to 'struct pcb', and set PCB_SAVECTX on contexts savedMark Kettenis
by savectx(). ok art@, miod@
2007-10-02Fiddle Sun traditional disklabel recognition to eliminate union andKenneth R Westerback
associated field access defines. This reverts most changes to dev/sun/disklabel.h without eliminating new functionality. Fix comments to line up and fit on 80 column line, making both tsi@ and I feel better. Simplify logic in disksubr.c. No semantic change. Tested & ok tsi@.
2007-10-02Apply (with slight variants) this elimination of bzero() with M_ZERO:Kenneth R Westerback
- if ((mapstore = malloc(mapsize, M_DEVBUF, - (flags & BUS_DMA_NOWAIT) ? M_NOWAIT : M_WAITOK)) == NULL) + if ((mapstore = malloc(mapsize, M_DEVBUF, (flags & BUS_DMA_NOWAIT) ? + (M_NOWAIT | M_ZERO) : (M_WAITOK | M_ZERO))) == NULL) return (ENOMEM); - bzero(mapstore, mapsize);
2007-10-01More easy bzero() -> M_ZERO. Use 'p = malloc(sizeof(*p) ...' whereKenneth R Westerback
obvious.
2007-10-01more btoc -> atopMartin Reindl
2007-09-30Move intrpending array into 'struct cpu_info'.Mark Kettenis
2007-09-30On i386, only call amd64_errata() on amd64 processors (family 0xf),Gordon Willem Klok
tested by krw@. This diff by itself is insufficient to ensure that we are only running on an amd64: todd@'s machine is a mobile Athlon XP that has a cpuid family of 0xf but is not an amd64 processor so more must be done but this should prevent the vast majority of cases. ok tom@
2007-09-30Enable Broadcom wireless, bwi(4).Jonathan Gray
If people testing could send mail to mglocker and myself we'd appreciate it. ok mglocker@
2007-09-29At the requst of deraadt@, backpout the temporary hack for pcc.Otto Moerbeek
Paraphrasing, "make it clear this should be fixed in pcc asap".
2007-09-24Ooops, i386 entry was enabled by mistake. Disable it. Sorry.Marcus Glocker
2007-09-24Add bwi PCI and Cardbus GENERIC entries for i386 and amd64, but stillMarcus Glocker
disabled for now. From brad@
2007-09-22Add kernel locking.Mark Kettenis
2007-09-22cf_context_switch takes an u_int argument.Mark Kettenis
ok drahn@
2007-09-22M_ZERO -> bzero.Kenneth R Westerback
ok art@
2007-09-22replace even more ctob and btoc with ptoa and atop respectively plusMartin Reindl
uvm_extern.h where needed
2007-09-19On sparc & sparc64, change `mount -a` to recognise Linux ext2 partitionsMarc Aurele La France
by interpreting more fields out of a standard Sun disk label. ok krw@
2007-09-18arc4random_bytes() is the preferred interface for generating nonces;Damien Miller
"looks ok" markus@
2007-09-18As a temporary workaround, implement va_start not usingOtto Moerbeek
__builtin_next_arg for !GCC until pcc supports it. ok kjell@ miod@
2007-09-18Fix typo (`Advancded').Michael Knudsen
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-12 17:40:27 +0000