summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2013-12-28Back out the previous commit; rodata gets merged with text by the linker.Mark Kettenis
Having a seperate rodata segment won't work anyway without significant pmap changes.
2013-12-28create rodata PHDR; ok kettenisTheo de Raadt
2013-12-28Require an explicit PT_OPENBSD_RANDOMIZE program header in the kernel linkerMiod Vallat
script (note that the amd64 linker script is currently not used). Discussed with deraadt@
2013-12-28Further clean up and unification of the amd64 and i386 boot(8) makefiles.Joel Sing
2013-12-28Attempt to approximate what should happen on a suspend/resume cycle.Theo de Raadt
If the driver was doing some IO, we remove the timeouts, and force the fdc state machine into IOTIMEDOUT state with the final timeout count before a clean retry. In theory upon resume it should freak out quietly, and try the operation again. Noone has stepped forward to test this yet.
2013-12-28Sync activate code sequnces to if_msk.c as much as possible, in caseTheo de Raadt
one of these is ever found in a suspend/hibernate system.
2013-12-28The few network drivers that called their children's (ie. mii PHYTheo de Raadt
drivers) activate functions at DVACT_RESUME time do not need to do so, since their PHYs are repaired by IFF_UP. NOTE: if_msk is the one that previously relied on mii/eephy.c doing a crazy dance.
2013-12-28The few network drivers that called their children's (ie. mii PHYTheo de Raadt
drivers) activate functions at DVACT_RESUME time do not need to do so, since their PHYs are repaired by IFF_UP.
2013-12-28mii drivers no longer need activate functions. Repair of the PHYTheo de Raadt
configuration setting is done at resume time because all networks drivers which were previously up, do an IFF_UP operation which hits PHY_RESET. This was in snapshots for about 2 weeks.
2013-12-28Move the fairly heavy eephy_init sequence [which was only done atTheo de Raadt
attach() and activate() time] into the eephy_reset() routine. This means that a bit more work gets done at PHY_RESET time, but it means also means it gets done in all scenarios. Why? For the next commit... This was in snapshots for about 2 weeks.
2013-12-28change the stack protector guard into a long word (removing the old legacyTheo de Raadt
compat pointed out by miod), and place it inside the ELF .openbsd.randomdata segment. Inside main(), only re-initialize the guard if the bootblocks failed to initialize it for us.
2013-12-28We can initialize the srandom/random system earlier from arc4random,Theo de Raadt
and do not need microtime.
2013-12-28We can random_start() earlier (not that it does too much) and callTheo de Raadt
arc4random() much earlier. Leading to random pids for anything besides 0 and 1.
2013-12-28Put the entropy_pool[] into the ELF .openbsd.randomdata segment.Theo de Raadt
Also allow random_init() to be called later, by moving a few entropy control initializions into the lower-level _rs_seed() layer. tested by jsing, phessler and a few others
2013-12-28crank the versionTheo de Raadt
2013-12-28Try to load entropy data from disk:/etc/random.seed, and additionallyTheo de Raadt
use a MD-supplied random function. Then, insert this into the ELF openbsd.randomdata of the kernel, so that it has entropy right from the start. Some help from jsing for the softraid aspects. Also tested by phessler
2013-12-28If we fail to decrypt the softraid keys, return EPERM rather than fallingJoel Sing
through and failing when we attempt to read and validate the disklabel. Also return ENOTSUP rather than EPERM if an attempt is made to write to a softraid volume.
2013-12-28Use preprocessor symbols where possible.Theo de Raadt
2013-12-28Add initial implementations of early MD random, for use in the boot code.Joel Sing
This makes use of rdrand if the CPU supports it, otherwise we fall back to using rdtsc. Further development will happen in the tree. Tested by phessler@ ok deraadt@
2013-12-27remove lots from the Makefile, since this never depended on libsa.Theo de Raadt
2013-12-27oops, correct wordingTheo de Raadt
2013-12-27correct comment about why first 16 pages are skipped: SMI corruptionTheo de Raadt
is the real cause discussed with mlarkin
2013-12-26When running the ll/sc version of the mutex code (for MULTIPROCESSOR kernels),Miod Vallat
correctly handle sc failures. All other ll/sc constructs were doing this correctly but apparently noone had noticed mutex did not.
2013-12-26Back at t2k13, I wrote code to park APs in real mode before resuming aMike Larkin
hibernated image. We backed out the code because it was causing reboots on resume. Turns out the parking code had a bug that caused the CPU to jump to some bogus address (calculating a bad offset for the jump target), which was likely the source of the problem. This diff fixes the bad offset calculation (verified by looking at the resulting asm output). This will be the first step in attempting to resurrect the original idea (and eventually add i386 if/when it works). discussed with deraadt@
2013-12-26either dvp == vp or dvp != vp: zap extra test.Marc Espie
okay guenther@
2013-12-25Instead of deciding which iockbc port is the keyboard port, and which one is theMiod Vallat
mouse port, depending upon the system time, match what the prom is doing and actually probe for a keyboard on both ports, and decide the first port with a keyboard is the keyboard port. If no keyboard is found, but a mouse is found, decide the keyboard port is the empty one. If no device is found, then we can try and pick the defaults, depending upon the system we are running on, as this used to be the case (i.e. coping with Fuel having keyboard on port 1 and mouse on port 0 when connecting devices according to the chassis' markings). This is necessary because different IO9 board revisions on Tezro come with different wirings, and we can not tell these boards apart. Discussed with "nullnilaki" (nullnilaki on gmail) who is the lucky owner of an Onyx 350 with correct wiring and a Tezro with inverted wiring. Tested on Octane and Fuel with all combinations of devices connected (mouse only, keyboard only, keyboard and mouse) in both ports, glass and serial console. XXX We probably want to allow for more pckbd attachment flexibility on non-x86 XXX platforms eventually (at least where the PS/2 slots are really independent, XXX so that we can attach pckbd to any port and better cope with human error XXX when connecting devices.
2013-12-25zap defines that don't make any sense for usMarc Espie
okay tedu@
2013-12-25Add missing IPL_MPSAFE, as same as mvme88k.Kenji Aoyama
ok miod@
2013-12-24rearrange/correct timeout conditionals to work better.Ted Unangst
fixes negative timeout panics. tested by sthen.
2013-12-24If the FADT has its SMI_CMD set to zero, assume we're only ACPI-only hardwareMark Kettenis
and don't need to disable SMI ownership of the ACPI hardware registers. ok mlarkin@
2013-12-24get rid of if (timeout_pending()) timeout_del(). this is racy. anyDavid Gwynne
conditionals you did on timeout_pending can now be done on timeout_del now that it returns what it did. ok and a very good fix from kettenis@
2013-12-24rename local ticks to nticks to avoid aliasing global. ok krwTed Unangst
2013-12-24fix error checking oddity in msdosfs code, as noted by kettenis@Alexander Hall
ok mikeb@ espie@
2013-12-23remove junkTheo de Raadt
2013-12-23spellingTheo de Raadt
2013-12-23- fix gamma ramp (now white is real white instead of light grey)Miod Vallat
- speed-up the integer to single precision float routine by omitting the sign check and starting the log2 computation at 12 instead of 32, since all the values passed to this routine are actually on-screen coordinates.
2013-12-23sync specvops with ffs (fixing spec_open along the way)Ted Unangst
ok espie
2013-12-23Add an instruction barrier to bus_space_barrier(). From David Hill.Mark Kettenis
ok deraadt@
2013-12-23Move KASSERT a bit further down because it can apparently be triggered byMark Kettenis
trying to hardlink a directory. While this is not allowed on tmpfs, this allows us to reach the code that checks for this such that we can return the appropriate error. Cluestick from Pedro Martelletto. ok espie@
2013-12-23regenBrad Smith
2013-12-23Add Atheros AR9462 and AR9565Brad Smith
2013-12-23Revert previous commit. It makes resume hang on some MP systems runningMark Kettenis
GENERIC.MP.
2013-12-23provide a TASK_INITIALIZER() macro.David Gwynne
2013-12-22Try to keep the framebuffer console layout that was set up by the firmwareMark Kettenis
on sparc64.
2013-12-22Add support for BCM57786,which seems to be almost indistinguishable from theMark Kettenis
BCM57785. Tested by Mark Rowland.
2013-12-22Use the rasops_info struct as the accesscookie for wsdisplay. Simplifies theMark Kettenis
code as it allows us to use some of the rasops_xxx functions directly.
2013-12-22Bail out early if the PCI bus number is -1. This means the hardware isn'tMark Kettenis
there and the AML might reference stuff that isn't there.
2013-12-22Only attach cpu0 on single-processor kernels, to match the expected behaviour.Miod Vallat
2013-12-22Compute ncpusfound correctly. The logic was there but the count variable wasMiod Vallat
kept MD.
2013-12-22MP-safe rw_cas() implementation for rwlocks.Miod Vallat
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-07 08:13:55 +0000