src - OpenBSD base system

Age	Commit message (Collapse)	Author
2013-01-01	Remove libsa exec.c, not used by anything anymore; all boot blocks now use	Miod Vallat
	loadfile or a derivative of.
2013-01-01	Fix gap value when formatting 1.44MB 3"1/2 floppy disks; from NetBSD	Miod Vallat

2012-12-31	Remove emulation errno mapping code from platforms that no longer have	Philip Guenther
	non-native emulations. ok miod@
2012-12-22	We already zero out the trapframe, so there is no reason to set individual	Mark Kettenis
	members to zero. Gets rid of a bogus comment as a bonus.
2012-12-08	Make ldomd(8) to control the availability of virtual disks to guest domains.	Mark Kettenis
	This is done by opening the corresponding /dev/vdspN device file. The virtual disk will remain available until that device is closed, which happens automatically when ldomd(8) exits.
2012-12-08	Descriptor rings can be larger than a single page, so add the offset instead	Mark Kettenis
	of or-ing it in.
2012-12-07	Use a timeout to do a delayed handshake if the receive queue of the LDC	Mark Kettenis
	channel transitions into the "up" state.
2012-12-05	Remove excessive sys/cdefs.h inclusion	Theo de Raadt
	ok guenther millert kettenis
2012-12-04	Attach the 2nd core of SPARC64-VI and SPARC64-VII CPUs.	Mark Kettenis

2012-12-03	Close disk images if the client disconnects. This makes it possible to	Mark Kettenis
	replace a disk image and have it visible to the guest when it reboots instead of having to reboot the primary domain.
2012-12-02	Implement hardware flow control and enable it by default (by setting CRTSCTS).	Mark Kettenis
	ok miod@, deraadt@
2012-12-02	Add "support" for VD_OP_GET_WCE and VD_OP_SET_WCE. These operations don't	Mark Kettenis
	really make sense for virtual disks backed by disk images, so we return ENOTSUP just like Solaris does. But advertising support for these operations works around a bug in the Linux kernel.
2012-12-02	Determine whether we're currently on the alternative signal stack	Philip Guenthe
	dynamically, by comparing the stack pointer against the altstack base and size, so that you get the correct answer if you longjmp out of the signal handler, as tested by regress/sys/kern/stackjmp/. Also, fix alt stack handling on vax, where it was completely broken. Testing and corrections by miod@, krw@, tobiasu@, pirofti@
2012-12-01	Silently ignore VD_OP_GET_DEVID as Solaris uses it even though we don't	Mark Kettenis
	advertise it.
2012-11-24	Do not assign the "host" MAC address to virtual switch ports. Don't print the	Mark Kettenis
	address for those ports either.
2012-11-23	Change the way we encode devhandle and devino in the sysino such that we can	Mark Kettenis
	handle up to 256 different interrupt vectors on cbus(4). This is enough to handle 128 channels.
2012-11-21	Atempting to eject a CD-ROM makes the SCSI midlayer generate a READ_TOC	Mark Kettenis
	command. Silently fail that command instead of spamming the console.
2012-11-21	Try to negotiate version 1.1 of the vDisk protocol such that we can get	Mark Kettenis
	the vdisk media type and fake a CD-ROM drive.
2012-11-07	Enable %tick access for userland on sun4u systems (sun4v systems already have	Mark Kettenis
	this enabled). ok pirofti@, mikeb@
2012-11-06	Fix comment.	Mark Kettenis

2012-11-05	unifdef -D __HAVE_TIMECOUNTER	Miod Vallat

2012-11-01	When possible, send multiple characters per packet. Also, stop sending when	Mark Kettenis
	the transmit queue is full. If the transmit queue is full, set a timeout to get things going again.
2012-10-29	Set the type of a processed MSI data record to 0, and check that the type	Mark Kettenis
	is non-zero before handling them, like we do in pyro(4). Seems to fix issues with MSIs on sun4v, so enable them again.
2012-10-29	On UltraSPARC T1/T2, block the current strand while spinning in the hope	Mark Kettenis
	other strands can do some useful work. Idea stolen from Linux. Results in a small, but measurable speedup doing a kernel build and reduces the system time by almost 10%.
2012-10-27	Also attach to channels name "ldom-<guestname>" (but not "ldom-primary").	Mark Kettenis

2012-10-27	Make polling work.	Mark Kettenis

2012-10-26	Turns out that for devices that attach to cbus(4), it is better to disable	Mark Kettenis
	interrupts up-front and explicitly enabling them later than the other way around.
2012-10-22	Stop passing around PS_STRINGS in %g1. The ELF ABI reserves this register	Mark Kettenis
	for passing around a pointer to a cleanup function and we'd like to use it for that purpose in the near future. ok miod@
2012-10-21	Fix thinko in previous.	Mark Kettenis

2012-10-21	Make sure that there is enough room in the transmit queue to send a VIO	Mark Kettenis
	message. Sleep if necessary.
2012-10-21	I want msdos filesystem support on the big bsd.rd	Theo de Raadt

2012-10-21	Clear a pending channel reset when a device is opened.	Mark Kettenis

2012-10-20	off-by-one in device number check; found the hard way and reported by J Sisson.	Miod Vallat

2012-10-17	Swap arguments to wdog_register() since it is nicer, and prepare	Theo de Raadt
	wdog_shutdown() for external usage.
2012-10-17	Enable vldc(4) and vldcp(4).	Mark Kettenis

2012-10-17	Disable debug code.	Mark Kettenis

2012-10-17	Rename hvctl(4) into vldcp(4) since this driver now handles other services	Mark Kettenis
	as well.
2012-10-16	Add support for multiple channels.	Mark Kettenis

2012-10-14	Fix typo in debug print.	Mark Kettenis

2012-10-08	Revamp the sequences for suspend/hibernate -> resume so that the code	Theo de Raadt
	paths are reflexive. It is now possible to fail part-way through a suspend sequence, and recover along the resume code path. Split DVACT_SUSPEND by adding a new DVACT_POWERDOWN method is used after hibernate (and suspend too) to finish the job. Some drivers must be converted at the same time to use this instead of shutdown hooks (the others will follow at a later time) ok kettenis mlarkin
2012-09-27	enable smsc	Jonathan Gray

2012-09-11	Remove the 'OLF method' used for the transition from a.out to ELF and	Theo de Raadt
	for all the compat layers which are now gone. Linux compat still works because it always used another method in any case, and nothing looks at p_os anymore. ok jsing
2012-09-08	On SPARC64 VI/VII CPUs, use the lseep instruction in the idle loop to force a	Mark Kettenis
	thread switch in the hope the other thread can do some useful work.
2012-08-31	remove some compat junk; ok kettenis	Theo de Raadt

2012-08-30	On SPARC64 VI/VII CPUs, use the sleep instruction while spinning to force a	Mark Kettenis
	thread switch in the hope the other thread can do some useful work.
2012-08-30	Slightly tweak the way we set up the code patch sections such that we can	Mark Kettenis
	use them to patch inline assembly in C code as well.
2012-08-30	Kill NOPIE and NOPIE_FLAGS in sparc64/stand. Binaries are identical according	Pascal Stumpf
	to deraadt@, ok kettenis@
2012-08-30	no md_root_setconf	Theo de Raadt

2012-08-29	The low-level guts to support MTP (Multi-Threaded Processing) on the	Mark Kettenis
	Fujitsu SPARC64-VI and SPARC64-VII CPUs. Since the two threads on each core share the TLBs of the core we cannot enter different mappings for the same virtual address. Instead we use a scratch register to store the per-cpu pointer. This is very similar to what we do on sun4v. For now we still only attach the first thread of each SPARC64-VI/VII core since we currently don't handle the VMT (Vertical Multi-Threading) of the SPARC64-VI very well.
2012-08-29	Make sure the interrupt stack for the boot processor is properly aligned to	Mark Kettenis
	a 64K boundary. This means we don't have to worry about virtual cache aliasing anymore since SPARC V9 CPUs have at most a 16K aliasing.