summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2005-10-15regenBrad Smith
2005-10-15Tweak names and add entries to the RCC (ServerWorks/Broadcom) PCI ids.Brad Smith
Based on Linux's pci.ids and sata_svw.c
2005-10-15sort PCI idsBrad Smith
2005-10-15remove `workaround' for supported data rates now that core is fixed.Federico G. Schwindt
ifconfig -m shows the correct rates now. tested by joris@, deraadt@ go ahead.
2005-10-15correctly parse the supported data rates. needed by wi@usb.Federico G. Schwindt
tested by jsg@ and joris@, deraadt@ go ahead.
2005-10-15spl values are int; evol@online.ptt.ruTheo de Raadt
2005-10-14Eliminate uneeded variable 'i' and use under utilized 'data_track'Kenneth R Westerback
instead. ok deraadt@
2005-10-14no functional change here, just moving the stripping of theBrad Smith
FCS into the SIS_RXBYTES() macro. As done in the FreeBSD sis driver.
2005-10-14only print warning message in bge_stop_block() if debug is enabled.Brad Smith
2005-10-14sprinkle some ANSI and KNF and remove an empty useless function.Brad Smith
2005-10-14Fixup some mnemonics, code was correct, stupid assembler doesn't typeDale Rahn
check closely...
2005-10-14avoid silly static variables that even caused nesting issues,Brad Smith
not to mention reentrancy concerns. From christos NetBSD ok deraadt@
2005-10-14bye bye more compile time knobs, use the sysctl to enable v6 forwarding.Brad Smith
ok deraadt@
2005-10-14shuffle these structs up to the top.Brad Smith
2005-10-13pump up the high water mark on the dirhash pool to avoid page allocation ↵Michael Shalayeff
throttling; pedro@ ok
2005-10-13Merge <machine/cpu_number.h> into <machine/cpu.h>, preparing for intrusiveMiod Vallat
changes.
2005-10-13Hide more definitions from userland and from locore.Miod Vallat
2005-10-13Put the register clobbering constraints on the BUG trap instruction,Miod Vallat
not the first ldcr; hopefully gcc did not dare optimizing too much around here, so this should not change anything in practice. test martin@
2005-10-13be sure to always set fifolen to at least 1 and if the code probes a 0 byteFederico G. Schwindt
fifo (ie., a 16450 misdetected as 16550a) the HW_FIFO bit is disabled. reported and ok by miod@. work with him.
2005-10-13In realitexpire(), do not re-add the itimer timeout if the process is inAaron Campbell
the midst of exiting. This solves a race condition that causes freed memory to be left referenced in the master kernel timeout worklist, leading to a uvm_fault. The same race condition was already fixed earlier in r1.53 of kern_clock.c for the process virtual and profile timeout structs. deraadt@ ok, testing by krw@
2005-10-13Over the years, different UARTs have appeared in the market.Federico G. Schwindt
Unfortunately most vendors implement hidden bits/features and ships with buggy buffers. This is our first attempt to fix this. Test the fifo size by putting the fifo into loopback mode and reading back what we wrote. It seems to be very much accurate and has correctly identified several buggy UARTs. Commented out for now. Tested by several people with different hardware, many thanks. Idea from Theo, code by myself.
2005-10-13first thing found by the fifo tester; pxa2x0 uarts are actually 64 bytesFederico G. Schwindt
deep. ok by uwe@ who confirmed this against specs.
2005-10-13sync multicast code with the FreeBSD sis driver.Brad Smith
2005-10-12have sis strip off the Ethernet FCS before passing it to bpf.Brad Smith
ok fgsch@
2005-10-12Define IPL_SCHED and spslsched(), matching the statclock level.Miod Vallat
2005-10-12regenBrad Smith
2005-10-12add Radeon X600 (RV380), Radeon 9550 and Radeon Mobility M300 M22Brad Smith
2005-10-12Kill unused global variable.Miod Vallat
2005-10-12Move sigcode to the m88k-agnostic location. No functional change.Miod Vallat
2005-10-12Remove commented out dead wrong FPU handling code in setregs().Miod Vallat
2005-10-12regenMichael Shalayeff
2005-10-12add few other ancient cpu types we have not seen yetMichael Shalayeff
2005-10-12Stop mapping the u area at fixed UADDR in addition to its actual va.Miod Vallat
While there, attempt to clean and comment stack usage in the kernel. No functional change. From the m88k SMP tree; help&test martin@
2005-10-12General scheduler cleanup:Miod Vallat
- at the end of switch_exit(), jump in a better location inside cpu_switch() - skip unnecessary curproc == NULL tests, and remove other dead code - simplify the idle loop - optimize whichqs bit operations - and add comments at strategic places help&test martin@
2005-10-12unconflict (one day legacy) wavelan ioctl's from the vlan ioctl'sTheo de Raadt
so that vlan's can be done on wavelan (if you are crazy enough to want to). spotted by fgs, but my fix, ok a few people
2005-10-12no need to call mfree() since m is always NULL; from Andrey MatveevMarkus Friedl
2005-10-12also handle network device bootpaths. also tested by drahn and kettenisTheo de Raadt
2005-10-11have gem strip off the Ethernet FCS before passing it to bpf.Brad Smith
ok krw@, also tested by Peter Hessler on macppc
2005-10-11Dont reset the registers if the same speed is programmed. helps reduceDale Rahn
the possible race where the frequency fails to set when changing speeds. hopefully this will workaround the problem reported on misc@
2005-10-11use [bl]etohXX instead of ntoh[sl], makes the code more consistent andHans-Joerg Hoexer
better readable. actually no binary change. ok krw@
2005-10-11Eliminate use of M_HASFCS.Brad Smith
From thorpej NetBSD ok jsg@
2005-10-11Change some byte shifting to _4btol() and swap32() calls.Kenneth R Westerback
ok hshoexer@
2005-10-11regenJonathan Gray
2005-10-11Consistently refer to Intel chipsets, ie 915G becomes 82915G.Jonathan Gray
2005-10-11regenJonathan Gray
2005-10-11A handful of Intel and Acer Labs devices from submitted dmesgs.Jonathan Gray
2005-10-11ansi and a sprinkle of knf. i'll do anything to avoid doing assignments.David Gwynne
2005-10-11make all usb1 controllers sleep until all the usb2 controllers have probedDavid Gwynne
and handed over any usb1 devices to the companion. without this usb1 devices didnt appear till after root was mounted, which is frustrating if you want to use a usb keyboard to enter the root device. tested by kettenis@ and drahn@ ok drahn@ go for it deraadt@
2005-10-11TC le(4) needs to be accessed in sparse address space for alpha.Martin Reindl
Should fix TC ethernet cards as noted by some people. from NetBSD ok miod@
2005-10-10Some KNF. malloc/free instead of MALLOC/FREE since it's a variableKenneth R Westerback
sized allocation. Remove some dead code.