src - OpenBSD base system

Age	Commit message (Collapse)	Author
2013-08-13	Let ray(4) rot in the attic. OK mpi, jsg, kettenis, henning, millert.	Mike Belopuhov
	Mickey approves these changes.
2013-08-12	Upcoming radeondrm(4) support will grow the kernel. This growth causes the	Mark Kettenis
	linker to merge ithe .text and .data sections, even though there is no reason why it should do this. Leave more space between these sections to keep the linker from doing this.
2013-08-12	radeondrm now attaches directly to pci.	Martin Pieuchot

2013-08-12	Match the framebuffer device against its OpenFirmware node in order	Martin Pieuchot
	to share the console detection logic with sparc64 in radeondrm(4). ok miod@
2013-08-12	Add a port of the TTM and Radeon DRM code from Linux 3.8.13.	Jonathan Gray
	Includes kernel modesetting, framebuffer console and support for newer hardware. Firmware needs to be present for acceleration and in some cases modesetting to work. It can be installed via fw_update or manually via pkg_add. With lots of help from kettenis@ some macppc bits from mpi@ and some ttm refcount/queue bits from FreeBSD. Thanks to M:Tier and the OpenBSD Foundation for sponsoring this work.
2013-08-10	On 'original' LUNA-88K, NVRAM contents and Timekeeper registers are	Kenji Aoyama
	mapped on the most significant byte of each 32bit word. (i.e. 4-bytes stride) Also, add small 'Wrong year set by UniOS-Mach after Y2K' hack. ok miod@
2013-08-10	Move the device table inside the softc.	Martin Pieuchot

2013-08-10	Initialize first_addr much earlier, in locore; this will help adding	Miod Vallat
	symbols support in the not-so-distant future. ok aoyama@
2013-08-10	Only clear the bss after we have picked a cpu as master, and on this cpu only;	Miod Vallat
	fixes a potential race where a secondary cpu resets carefully computed variables. ok aoyama@
2013-08-09	Add urtwn(4) for upgrades	Martin Pieuchot

2013-08-08	Add PMAP_NOCACHE as an alias for PMAP_NC for use in semi-MI code.	Mark Kettenis

2013-08-08	/dev/drmN for sparc64	Mark Kettenis

2013-08-08	Most network drivers include netinet/in_var.h, but apparently they	Alexander Bluhm
	don't have to. Remove these include lines from vax drivers. test lteo@
2013-08-08	Remove duplicate loop.	Raphael Graf
	ok patrick@
2013-08-07	Most network drivers include netinet/in_var.h, but apparently they	Alexander Bluhm
	don't have to. Remove these include lines from macppc drivers. test and OK mpi@
2013-08-07	We uniformly define size_t to be unsigned long and ssize_t to be long. Make	Mark Kettenis
	sure that SIZE_MAX and SSIZE_MAX are defined as constants with a matching type on all are architectures. ok millert@, matthew@
2013-08-07	Using phys_map for bus_space_map(9) is retarded. Use kernel_map just like	Mark Kettenis
	everybody else. ok mpi@, beck@, jsg@
2013-08-07	Managed device mappings should be uncached by default.	Mark Kettenis
	ok mpi@
2013-08-07	Mimic what's done on sparc64 and parse the OpenFirmware device tree to	Martin Pieuchot
	enumerate the pci devices. This allow us to not store the interrupt vector in the interrupt line register and to not try to attach pci devices disabled by firmware. With inputs from and ok kettenis@
2013-08-06	add Cortex A15 R4	Jonathan Gray
	ok patrick@
2013-08-06	The Cortex bus should be useful for Cortex-A7, too.	Patrick Wildt
	ok rapha@ jsg@
2013-08-06	Attach amptimer only on Cortex-A9, as all Cortex-A8 doesn't have a generic	Patrick Wildt
	timer and the newer Cortexes have another generic timer. ok rapha@ jsg@
2013-08-06	The atomic_setbits_int() and atomic_clearbits_int() functions touch the	Mark Kettenis
	condition code register so mark it s clobbered. Fixes uvm_page_physload() and who knows what else. ok jsg@, mpi@
2013-08-05	Add rsu(4).	Mark Kettenis

2013-08-04	de-lint hp300	Martin Reindl
	ok miod
2013-08-03	Don't flush the cache on page inflate for hibernate on amd64.	Mike Larkin
	Makes the x200 and w500 and probably other machines go a bit faster during resume from hibernate. This had been in my tree since t2k13.
2013-08-02	Drop the dependecy on objcopy.	Raphael Graf
	ok florian@
2013-07-30	Or in the CPUID_NXE bit from ci->ci_feature_eflags into ci->ci_feature_flags	Mark Kettenis
	to mimic what is done in locore.S. Otherwise we lose the CPUID_NXE bit. ok matthew@
2013-07-27	Revert wrong chunk introduced in 1.50, causing console keyboards not to attach	Miod Vallat
	as console; fixes non-wsmux kernels such as bsd.rd; noticed by sebastia@
2013-07-17	Use `long' instead of `int' in structure layouts which may get assigned from	Miod Vallat
	(or to) register_t, but need to be usable without register_t in scope, so that the compiler won't warn about these assignments after the recent _types.h changes; aoyama@
2013-07-17	Fix setregs() to match prototype. Exposed by recent _types.h change, found the	Miod Vallat
	hard way by aoyama@
2013-07-17	don't forget to insert the io to the list on the Jensen	Martin Reindl
	found my Maxime Villard, diff from NetBSD ok miod
2013-07-16	Enable jfb-style acceleration on XVR-600.	Miod Vallat

2013-07-16	Add support for BUS_INTR_ESTABLISH_MPSAFE. This makes the previous commit	Mark Kettenis
	to vcons(4) actually have an effect.
2013-07-16	The interrupt handler for vcons(4) is defenitely safe to run without holding	Mark Kettenis
	the kernel lock, so use BUS_INTR_ESTABLISH_MPSAFE.
2013-07-16	Add support for BUS_INTR_ESTABLISH_MPSAFE.	Mark Kettenis

2013-07-16	no longer needed, we use uenv.txt	Florian Obser
	OK bmercer@
2013-07-16	Don't use a non initialized variable and fix RCS Ids while here.	Martin Pieuchot
	Issue reported by Maxime Villard, diff from martin@
2013-07-15	Use getpropstringA(), not getpropstring(), to get the card's name. This ensures	Miod Vallat
	its value has not been clobbered when ifb_accel_identify() is invoked, and restores textmode acceleration on Expert3D{,-Lite} and XVR-1200. Broken since 1.17, sigh
2013-07-15	Use bus_space_set_region_4 instead of looping over bus_space_write_4,	Raphael Graf
	and remove some includes. ok patrick@
2013-07-14	To prevent lock ordering problems with the kernel lock, we need to make sure	Mark Kettenis
	we block all interrupts that can grab the kernel lock. The simplest way to achieve this is to make sure mutexes always raise the ipl to the highest level that has interrupts that grab the kernel lock. This will allow us to have "mpsafe" interrupt handlers at lower priority levels. No change for non-MULTIPROCESSOR kernels.
2013-07-14	IPL_VM is "mpsafe" now as well.	Mark Kettenis

2013-07-13	unify register_t to long on all platforms	Theo de Raadt
	ok miod kettenis
2013-07-13	Rework multiline string literal to appease gcc3.	Miod Vallat

2013-07-12	Actually assign return value to the variable that's used to check it.	Mark Kettenis
	Pointed out by Maxime Villard.
2013-07-12	replace fls functions with calls to the libkern versions	Jonathan Gray
	found the hard way by deraadt@
2013-07-11	Enable bus_space_set_region_4 for armv7.	Raphael Graf
	ok patrick@
2013-07-10	To prevent lock ordering problems with the kernel lock, we need to make sure	Mark Kettenis
	we block all interrupts that can grab the kernel lock. The simplest way to achieve this is to make sure mutexes always raise the ipl to the highest level that has interrupts that grab the kernel lock. This will allow us to have "mpsafe" interrupt handlers at lower priority levels. No change for non-MULTIPROCESSOR kernels. ok matthew@
2013-07-07	Fix pmap_extract() to actually only return TRUE for mappings which have the	Miod Vallat
	valid bit set; can't believe this went unnoticed for so long. This fixes uvm_km_pgremove_intrsafe panics when pool_put()'ing items in pools for larger-than-a-page-items at shutdown time.
2013-07-07	move to 5.4-beta	Theo de Raadt