summaryrefslogtreecommitdiff
path: root/sys/arch/alpha
AgeCommit message (Collapse)Author
2005-11-03use splvm() here instead of splimp(). in currently unused MULTIPROCESSOR code.Brad Smith
ok martin@
2005-10-30Remove unused pcb field (this should have been part of the fetch(9)/store(9)Miod Vallat
removal ages ago).
2005-10-28no more Mach-macrosMartin Reindl
2005-10-21remove two unused and deprecated Mach macros, alpha_round_page() andMartin Reindl
alpha_trunc_page() looks ok to Miod
2005-10-17Rename ncpus to alpha_ncpus so that MI ncpus matches the number ofMiod Vallat
attached processors; deraadt@ ok
2005-10-02enable ath(4) on alpha.Reyk Floeter
ok martin@ deraadt@
2005-09-27Don't try to read the DOS label from the device if spoofonly is 1.Kenneth R Westerback
This changes the architectures which use code identical to the amd64 disksubr.c. macppc seems different so it will be separate. The remaining archs don't try to read a DOS label. ok deraadt@
2005-09-15In syscall(), do not recompute the struct proc * after the syscall hasMiod Vallat
succeeded. This used to be necessary for fork(), when returning in the child, but we return in the child in child_return() which does TRT. ok art@
2005-09-15Change child_return() to record a proper ktrace record for vfork childMiod Vallat
processes. ok art@ uwe@ (some time ago)
2005-09-10beef up and correct the example sectionMartin Reindl
2005-08-29setpu stack for primary boot at 0x30000 after the load address.Michael Shalayeff
this allows netbooting uncompressed kernels on newer alphas. this shall be addressed futher after the release. deraadt@ ok
2005-08-10suggested by art, go for it deraadt, ok miod@Todd T. Fries
remove ECOFF support
2005-08-06If a process has trashed its stack, use sigexit() instead of psignal() toMiod Vallat
terminate it with prejudice, as done in native sendsig(). Plus you get a memory leak in COMPAT_HPUX on m68k fixed for free. review art@ test martin@
2005-08-03enable safte and ses on all archs that appear to have real scsi bussesDavid Gwynne
ok deraadt@ beck@ marco@
2005-08-01knfTheo de Raadt
2005-07-31Get rid of unused SINCR and SSIZE constants.Miod Vallat
2005-07-31unfortunately de(4) no longer fitsTheo de Raadt
people who add to the kernel without thinking of the growth caused this
2005-07-24Be more permissive when checking the interrupt vector for devices tied toMiod Vallat
the ISA interrupts; from NetBSD.
2005-07-14More netccitt and netns removals; OK henning, brad, mickeyUwe Stuehler
2005-06-20add ral(4)Martin Reindl
ok deraadt@
2005-06-17uipaq/ucom support; written by isis@www.cs.umd.eduTheo de Raadt
2005-06-17Override cpu_switch() with a faster version if we can use BWX instructions.Miod Vallat
From RusticBSD, ok deraadt@
2005-06-17Allow ev56 opcodes in the .s files by default (gas on alpha does not yet haveMiod Vallat
.ev56, .ev4, etc, in-file modifiers).
2005-05-30even shorterTheo de Raadt
2005-05-30shrink alpha boot blocks another little bitTheo de Raadt
2005-05-27add atapiscsi@umass to RAMDISKS as appropriate, ok krw@ deraadt@Dale Rahn
2005-05-26Add atapiscsi in parallel with scsibus@usb, this only changes systemDale Rahn
that dont have atapiscsi@wdc, fixes zaurus, ok krw@ todd@
2005-05-25comment typoJason Wright
2005-05-24fix links targetUwe Stuehler
2005-05-24move sk(4) and epic(4) out of the untested sectionBrad Smith
2005-05-17add epic(4) under the untested section.Brad Smith
ok martin@
2005-05-10Add epic(4) driver which supersedes tx(4).Brad Smith
This buys us.. - bus_dma - big endian support - multicast - support for another EPIC chipset From NetBSD ok mickey@
2005-05-09fixes and updates for root device detection from NetBSD:Martin Reindl
- remove unneeded checks and variables - use strncasecmp() to check b->protocol because the case is dependant on the firmware revision - check logical unit numbers - consistently check PCI function numbers - on the 6600 models check which PCI hierarchy is looked at - makes root device detection on DS20 (6600) work testing by claudio, xsa and miod; ok miod@
2005-05-09- add alpha optimized in4_cksumBrad Smith
- de register and use ANSI prototypes From NetBSD ok miod@ martin@
2005-05-06generate vim-compatible tag files, from netbsdPedro Martelletto
2005-05-01To allow for MD override of in4_cksum()...Brad Smith
ok deraadt@
2005-05-01spellingDavid Krause
2005-05-01spacing; ok miod@ deraadt@David Krause
2005-04-30start the job of making things look more alikeTheo de Raadt
2005-04-30nice docs at top of GENERICTheo de Raadt
2005-04-28shrink extra verbose crudTheo de Raadt
2005-04-28shut upTheo de Raadt
2005-04-21count fpu lazy context switches; deraadt@ okMichael Shalayeff
2005-04-21correct idiom for snprintf failure handling; ok beckTheo de Raadt
2005-04-20Enable ural(4), appears to work fine.Jonathan Gray
2005-04-19nothing uses spllowersoftclock() anymoreMichael Shalayeff
2005-04-14internal _BSD_WCTRANS_T_, _BSD_MSTATE_T_, _BSD_WCTYPE_T_ types.Marc Espie
2005-04-11use MD #define to choose stackgap size per-architecture. on sparc, specialTheo de Raadt
case sun4c/sun4 -- because address space is more constrained
2005-03-30first approximation: spoof UDF filesystem disklabels when we see them.Theo de Raadt
it is likely a slight bug or two will sneak in with this, so everyone please keep an eye out on your disklabels
2005-03-08Cosmetics...Thierry Deval
Mainly remove space between locators parens, replace spaces with tabs where appropriate and consistently align dependencies.