summaryrefslogtreecommitdiffstats
path: root/sys/ia64
Commit message (Expand)AuthorAgeFilesLines
* Mega busdma API commit.scottl2003-07-013-12/+79
* - Export pmap_enter_quick() to the MI VM. This will permit thealc2003-06-291-8/+10
* - Remove the calls to pmap_install() from pmap_object_init_pt(); they arealc2003-06-291-9/+9
* Implement cpu_set_upcall_kse(). Elementary testing shows that thismarcel2003-06-281-0/+26
* Add a machine depended function thread_siginfo, SA signal codedavidxu2003-06-281-0/+19
* Do the first and mostly mechanical step of adding mutex support to thescottl2003-06-271-0/+5
* Add TLS related relocation.marcel2003-06-191-0/+17
* Fix a performance bug in all of the various implementations ofalc2003-06-181-1/+1
* Rename P_THREADED to P_SA. P_SA means a process is using schedulerdavidxu2003-06-151-1/+1
* Migrate the thread stack management functions from the machine-dependentalc2003-06-142-47/+1
* Move the *_new_altkstack() and *_dispose_altkstack() functions out of thealc2003-06-141-26/+0
* Remove kernel event tracing. The overhead is significant when runningmarcel2003-06-141-6/+0
* Make sure pcpu->pc_pcb is pointing to a 16-byte aligned address. Themarcel2003-06-121-2/+10
* GC unused cpu_wait() functionpeter2003-06-111-7/+0
* Note that scbus is required for SCSI, not just "required" in general.jmallett2003-06-081-1/+1
* pmap_find_vhpt() has been observed to return a NULL pointer whenmarcel2003-06-071-10/+11
* If we get a fault in the gateway page, which would happen if we trymarcel2003-06-071-2/+5
* Use TRAPF_USERMODE() to replace an equivalent check in trap(). Whilemarcel2003-06-061-3/+2
* Have TRAPF_USERMODE() take into account that the gateway page is notmarcel2003-06-061-2/+8
* Fix the dreaded double counting that was present on alpha as well andmarcel2003-06-041-17/+21
* Change the second (and last) argument of cpu_set_upcall(). Previouslymarcel2003-06-041-6/+6
* Improve set_mcontext:marcel2003-06-011-0/+5
* Improve on cpu_set_upcall:marcel2003-06-011-13/+13
* Implement cpu_thread_setup(). This is mostly the same as on i386,marcel2003-06-011-0/+8
* Now that we have the signal trampolines in the gateway page and themarcel2003-05-311-1/+7
* Implement cpu_set_upcall(). Required by libthr and used bymarcel2003-05-311-0/+22
* Implement set_mcontext() and get_mcontext(). Just as for sendsig() andmarcel2003-05-311-4/+14
* Make the regset pointers const pointers for the context restore functions.marcel2003-05-311-3/+3
* Some ia32 related finetuning for the EPC syscall path:marcel2003-05-312-50/+64
* Make sure we have all the dirty registers in user frames on themarcel2003-05-311-2/+12
* Rename BUS_DMAMEM_NOSYNC to BUS_DMA_COHERENT.hmp2003-05-301-1/+1
* Move the sysctls of the misalignment handler to where they belongmarcel2003-05-293-26/+18
* Fix what I think is a cut-n-paste bug: use OID_AUTO for themarcel2003-05-291-1/+1
* A flushrs must be the first in an instruction group.marcel2003-05-272-0/+2
* Bring back bus_dmasync_op_t. It is now a typedef to an int, though thescottl2003-05-272-2/+3
* Have the unwinder allocate memory with M_NOWAIT. The unwinder ismarcel2003-05-271-1/+1
* Fix fu{byte|word*} and su{byte|word*}:marcel2003-05-273-298/+902
* Revision 1.99 of this file changed the allocation request frommarcel2003-05-261-1/+1
* Now that we define user mode as any IP address that isn't in themarcel2003-05-251-2/+1
* Fix a source of instability specific to an EPC userland. We returnmarcel2003-05-242-14/+8
* Be more careful how we restore interrupts. Don't rewrite most of themarcel2003-05-241-3/+4
* Consistently us the same metric to differentiate between kernel modemarcel2003-05-242-3/+2
* Unconditionally restore ar.k7 (memory stack) and ar.k6 (register stack)marcel2003-05-242-4/+4
* Fix an alpha inheritance bug:marcel2003-05-244-68/+28
* Fix a (new) source of instability:marcel2003-05-232-16/+14
* In swapctx(), put the RSE in enforced lazy mode before we flush themarcel2003-05-232-4/+4
* o Fix a definite bogon: the dirty bity fault, instruction accessmarcel2003-05-202-70/+56
* Rename the "IA64 ITC" counter to "ITC" counter. We don't call themarcel2003-05-201-1/+1
* Prevent corruption of the VHPT collision chain by protecting it withmarcel2003-05-201-45/+43
* sys/sys/limits.h:kan2003-05-192-3/+3
OpenPOWER on IntegriCloud