summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2007-10-30don't use intermediate target (a.out) in multiple targetsOtto Moerbeek
2007-10-30avoid intermediate files with same name (ie. a.out) for parallel makeTheo de Raadt
ok miod
2007-10-29remove extra backslash at the end of RB_PROTOTYPE, report fromDamien Miller
Jan.Pechanec AT Sun.COM; ok deraadt@
2007-10-29run depend so that SFILES depend on assym.h, which allows them to buildMarc Espie
with make -j on fast SMP boxes. noticed by robert@, input by miod, kettenis, okay toby@, robert@, and tested to work if people build kernels correctly...
2007-10-29add missing dependency. Taken from upstream gcc.Marc Espie
okay kettenis@ (`if it builds' and it does)
2007-10-29UltraSPARC CPUs (and other SPARC V9 implementations) don't have aMark Kettenis
floating-point deferred-trap queue. Remove redundant code inherited from sparc that deals with this. Also remove the code dealing with saving and restoring the FPU state from unaligned memory; we always allocate properly aligned memory for storing the FPU state.
2007-10-29Make sure the dma_cachectl*() functions actually do their work on allMiod Vallat
affected processors if option MULTIPROCESSOR. It's amazing bsd.mp could boot multiuser without this.
2007-10-29When a secondary cpu gets its interrupt pin stuck, be sure to savectxMiod Vallat
and put the process it was running back on the run queue (unless this was the idle proc).
2007-10-29For fd passing, at externalize time we are dealing with int[] not pointer[],Theo de Raadt
so we need to correct the mbuf length using CMSG_LEN() not CMSG_SPACE() test case was sparc64 sending 2 (or more) descriptors ok millert thib
2007-10-29Pull from master repo:Stefan Kempf
Fix sorting bug in addcase(). Spotted by Gregory McGarry. ok gmcgarry, ragge@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29More tweaking of option processing, this time around the creation ofKenneth R Westerback
the priority list. Add create_priority_list() and use it to consistantly merge mandatory, supplied and default option priority lists into a reliable and complete list of options sorted by priority. No intended functional change other than ensuring all options available are returned. "looks good" henning@ "OK, fair enough" millert@
2007-10-29From master repo:Anders Magnusson
> calls argument numbers are in registers, not bytes. > If a jump is to the return label, emit ret insn instead. > Add ADDROF table entry. > Remove some unused entries. > Do a Z call for emitted GOTOs. > Only create constants for function names, they are called by name anyway.
2007-10-29Updates from master repo:Anders Magnusson
> Allow ADDROF to fall through to table, for PIC code generation. > Tighten check on GOTO argument and only complain that the label is out-of-range > if we are actually jumping to label (ICON).
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
replace an hard coded value with M_WAITOK ok krw@
2007-10-29regenMike Belopuhov
2007-10-29Add missing Marvell 88SE61xx IDs.Mike Belopuhov
ok dlg
2007-10-29MALLOC/FREE -> malloc/freeCharles Longeau
ok krw@
2007-10-29unmap on error and correct some output in those cases.Federico G. Schwindt
martin@ deraadt@ ok.
2007-10-29remove reference to old ICE1712 name in brackets from .NdAlexandre Ratchov
from jmc@
2007-10-29Touch of KNF and style(9), spaces and comments.Thordur I. Bjornsson
make pretty. ok krw@
2007-10-29fix the setup of Tx descriptors. the frame type and "no ack" bits wereReyk Floeter
initialized in ar5k_ar521*_setup_tx_desc() but cleared in ar5k_ar521*_fill_tx_desc() by setting the segment length incorrectly. From ath5k via Nick Kossifidis (mickflemm at gmail dot com)* ---snip--- The missing no_ack in broadcast frames caused them to be retried up to the retry_limit(1+4=5 transmissions by default). ---snap--- * it was explicitly mentioned that it is ISC-licensed now
2007-10-29regenDavid Gwynne
2007-10-29the areca with the 1201 pci id is known as the 1200 everywhere. itDavid Gwynne
does the same things as a real 1200, but its based around a different chip. fix the string in pcidevs to say "ARC-1200 rev B" like the writing on the board does.
2007-10-29regen, finally.David Gwynne
2007-10-293 new areca controllersDavid Gwynne
2007-10-29clean up after previous macro removal;Jason McIntyre
2007-10-29Add simple regress test for LocalCommand; ok djm@Darren Tucker
2007-10-29Make LocalCommand work for Protocol 1 too; ok djm@Darren Tucker
2007-10-29ProxyCommand and LocalCommand use the user's shell, not /bin/sh; ok djm@Darren Tucker
2007-10-29Send config block back to slave for invalid users too so optionsDarren Tucker
set by a Match block (eg Banner) behave the same for non-existent users. Found by and ok djm@
2007-10-29Plug tiny mem leaks in ControlPath and ProxyCommand option processing; ok djm@Darren Tucker
2007-10-29get rid of the unused cond_predicate_t typedef.David Gwynne
ok miod@ deraadt@ krw@
2007-10-29Allow build without -DGSSAPI; ok deraadt@Darren Tucker
2007-10-28syncTheo de Raadt
2007-10-28add new envy(4) man pageAlexandre Ratchov
ok jmc@, ok deraadt@
2007-10-28Enable pcie snoop for SB600 and MCP51. Fixes a variety of noise,Deanna Phillips
looping and distortion problems. This is a temporary solution until we get something like the BUS_DMA_NOCACHE flag to bus_dmamem_map(9), which came with the driver when ported from NetBSD but means nothing here. It's likely that other NVIDIA and ATI chipsets will need this, but I'd rather have some testers before adding them in. From ariff@freebsd, ok dlg.
2007-10-28syncTheo de Raadt
2007-10-28This is a horrible kluge: invoke sched_init_cpu for seconday processorsMiod Vallat
before they are started (and not skipping gaps for machine setups with holes in cpu slots). Since we start secondary cpus very late in the boot process, and sched_init_cpu() has to be invoked before proc0 execve's init, I don't think there is a better way to do this. This lets MVME188 systems with more than one processor boot multiuser.
2007-10-28Start secondary processors synchronously. This only wins us a nicer dmesgMiod Vallat
output, and nothing else.
2007-10-28Do not flag a processor as ``alive'' until it really is ready to accept IPIs.Miod Vallat
2007-10-28more id tags.Federico G. Schwindt
2007-10-28id tags.Federico G. Schwindt
2007-10-28stop putting xfs (and only xfs) rcsid into the kernel; ok miodTheo de Raadt
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-11 14:51:03 +0000