summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2009-08-13Remove unrelated bit from last commit which breaks at least 2 arches.Bret Lambert
Bad blambert@, no biscuit.
2009-08-13- ansify function declarations, no binary changeJasper Lievisse Adriaanse
"fine" thib@
2009-08-13Forgot to add break to the switch cases. Silly.Paul Irofti
Okay deraadt@.
2009-08-13make scsi_done set ITSDONE on the xs. this means hba drivers dont have toDavid Gwynne
do it (but doesnt preclude them using it internally). discussed with krw@ miod@ and deraadt@
2009-08-13More changes to make the parsing of those evil messages a bit easier.Claudio Jeker
2009-08-13rwlock assertion functions, currently unusedBret Lambert
ok art@
2009-08-13Add wake/suspend support for pckbd(4).Paul Irofti
Patch initially from mlarkin@. KNF and refactoring by me. Suggestions and okay deraadt@.
2009-08-13Rewrite part of the promiscuous/multicast handling; tested by myself,Stuart Henderson
naddy@ and several tech@ readers. From Brad.
2009-08-13Enable guard pages and slow recycling of kva.Ariane van der Steldt
Debugging aid, hiding in "option KVA_GUARDPAGES" in config.
2009-08-13Add more stuff change some fields to make the handling a bit easier.Claudio Jeker
OK dlg@
2009-08-13add ioctls to allow userland to initiator a probe or detach of devices itDavid Gwynne
is providing to the kernel. this uses the req api i just added, otherwise attaches and detaches would be run from the context of the process issuing the ioctls, which then will not be able to answer them since theyre busy running the attach/detach in the kernel.
2009-08-13provide an api for an interrupt (or something like it) to notify theDavid Gwynne
midlayer that a scsi device has appeared or dissapeared. the midlayer will queue an event and run it in the system workq (which has process context) to handle the attach or detach.
2009-08-13repair vscsi lineTheo de Raadt
2009-08-13toby do you even care enough to try to do right?Theo de Raadt
2009-08-13typoMiod Vallat
2009-08-13rewrite the promisc mode / multicast handling code for the sparc hme(4)Peter Hessler
driver and is based off of the MI hme(4) driver. tested with ping6 -w and tcpdump from brad OK naddy@, sthen@
2009-08-13In pmap_pte_flush flush the icache before the dcache, as is done elsewhere.Theo de Raadt
We cannot find gaurantees in the documents that any icache filling from the cache is terminated by a icache flush or idtlb invalidation, so this is probably safer, and unlikely to be unsafer. ok kettenis
2009-08-13- remove super-obvious comments from vnodeop_entries[]Jasper Lievisse Adriaanse
- remove #ifdef someos blocks, makes this a tad easier to read agreed by art@ and thib@
2009-08-13fix a typo its supposed to be show all vnodes not show all vnodes,Thordur I. Bjornsson
ddb does keyword matching so i missed this. spotted by jasper@
2009-08-13Cleanup differences between i386 and amd64.Mark Kettenis
2009-08-13PAGEFASTRECYCLE is an option we have been using for a while to encourageTheo de Raadt
the kernel to reuse freed pages as quickly as possible, and it has been finding bugs (some of which we have already fixed) ok kettenis
2009-08-13Replace the error strings that were being passed around with much simplerTheo de Raadt
errnos. Note that the error strings are being ignored, since we long ago decided to not spam the console, and there is no other nice way to use the errors (without changing the ioctls to pass it back) The errno is now useful, since we can pass b_error from failing IO up, and the drive can decide how to use that ok miod
2009-08-13Replace the error strings that were being passed around with much simplerTheo de Raadt
errnos. Note that the error strings are being ignored, since we long ago decided to not spam the console, and there is no other nice way to use the errors (without changing the ioctls to pass it back) The errno is now useful, since we can pass b_error from failing IO up, and the drive can decide how to use that ok miod
2009-08-13Teach nfsm_fhtom about nfsm_info structsBret Lambert
ok thib@
2009-08-13- remove the ef(4) chunk, it's unused now and the ioc ethernet driverJasper Lievisse Adriaanse
won't be called ef(4) anyway. ok miod@
2009-08-13i forgot a commentDavid Gwynne
2009-08-13wire vscsi up to a cdevDavid Gwynne
for claudio@ ok deraadt@
2009-08-13change version 3 post-operation attribute building function to notBret Lambert
use mbufs directly, but the shiny new nfsm_info struct ok thib@
2009-08-13- remove super-obvious comments from $fs_vnodeop_entries[]Jasper Lievisse Adriaanse
prodded by and ok thib@ agreed by art@ and blambert@
2009-08-13- consistify cfdriver for the ethernet drivers (0 -> NULL)Jasper Lievisse Adriaanse
ok dlg@
2009-08-13add a show all vnodes command, use dlg's nice pool_walk() to accomplishThordur I. Bjornsson
this. ok beck@, dlg@
2009-08-13Start using a linking script for this kernel. ThisTobias Weingartner
should help in future using large pages for text/etc. Also, since we do not use the .eh frame stuff, we can nuke them, saving some bytes... Ok kettenis@, "more control over linking is a good thing, but I can't comment further" art@.
2009-08-13A new(er) mtx_enter_try().Tobias Weingartner
Ok oga@, "the time is now" deraadt@.
2009-08-13Make the i386 APIC code closer to the amd64 version. Many whitespace fixes,Mark Kettenis
some additional code. ok toby@, oga@
2009-08-13- missing printf("\n"); after calling lockmgr_printinfo()Jasper Lievisse Adriaanse
"do it" thib@
2009-08-13set ITSDONE on the xs before calling scsi_done()David Gwynne
2009-08-13- set DV_IFNET, instead of DV_NULLJasper Lievisse Adriaanse
ok kettenis@
2009-08-12Add wake support to acpihpet(4).Paul Irofti
Patch initially from mlarkin@. KNF and refactoring by me. Okay deraadt@.
2009-08-12Optimize Syndrome calculation, 30% speedup on writesJordan Hargrave
ok marco@
2009-08-12just bzero the pkthdr instead of setting each and every member of thatHenning Brauer
struct to 0/NULL. no performance impact but way less error prone on addition of new pkthdr field (as just ran into with a theo diff). ok theo
2009-08-12only pull in the source file if it is needed, so that all the othersTheo de Raadt
we use compile brilliant cnst, please learn to test better
2009-08-12revert my change to m_cluncount which tries to prevent the systemDavid Gwynne
running out of mbufs for rx rings. if the system low watermark is lower than a rx rings low watermark, we'll never send a packet up the stack, we'll always recycle it. found by thib@ on a bge sadface
2009-08-12Let this work with C99 compilers that do not define __GNUC__Jonathan Gray
ok niklas@
2009-08-12Non-equivalent writable aliases (aka illegal aliases) are bad. Since PA-RISCMark Kettenis
doesn't have a usuable uncachable bit, whack all other mappings of a page if we are about to create a non-equivalent writable alias by entering a mapping for it. ok art@, miod@
2009-08-12Mark sensor as invalid if the i2c transaction to read its value fails.Mark Kettenis
ok deraadt@
2009-08-12ix currently relies on a gcc extension that removes a commaJonathan Gray
if no arguments are passed to a variadic macro. Create a seperate non variadic macro for this fixed argument case. No binary change. ok reyk@
2009-08-12Namecache revamp.Bob Beck
This eliminates the large single namecache hash table, and implements the name cache as a global lru of entires, and a redblack tree in each vnode. It makes cache_purge actually purge the namecache entries associated with a vnode when a vnode is recycled (very important for later on actually being able to resize the vnode pool) This commit does #if 0 out a bunch of procmap code that was already broken before this change, but needs to be redone completely. Tested by many, including in thib's nfs test setup. ok oga@,art@,thib@,miod@
2009-08-12Backing store for ega and vga virtual console is not allocated until theMiod Vallat
second vc is created. However, it was allocated using the geometry of the second vc to allocate backing store for the first. Be sure to use the proper values in case geometries differ.
2009-08-12dlg deferred calling interfaces' if_start routine so we call them less,Henning Brauer
which does pay out, performance wise. one of the conditions to call the interfaces' if_start routine immediately was "send queue is full". on a very busy (hammered) machine this will itroduce too much latency since we spend almost all cpu time in interrupt handlers and softnet, so the softint actually doing the if_start gets called to seldom and the queue full check is what triggers the actual transmit. change the logic to call if's if_start routing immediately when there are at least 8 packets (or in case if maxlen being smaller than 8, maxlen) 8 chose because it shows best performance in my test setup here. ok dlg
2009-08-12New lisa(4) driver for ST LIS331DL accelerometer in HP 2133 Mini-Note PC; ↵Constantine A. Murenin
ok deraadt