49611 Commits

Author SHA1 Message Date
Maxim Konovalov
4bc37f9836 o Turn net.inet.ip.check_interface sysctl off by default.
When net.inet.ip.check_interface was MFCed to RELENG_4 3+ years ago in
rev. 1.130.2.17 ip_input.c it was 1 by default but shortly changed to
0 (accidently?) in rev. 1.130.2.20 in RELENG_4 only.  Among with the
fact this knob is not documented it breaks POLA especially in bridge
environment.

OK'ed by:	andre
Reviewed by:	-current
2004-09-24 12:18:40 +00:00
Pawel Jakub Dawidek
12f5309d62 Document vfs.root.mountdelay value.
Reminded by:	ru
2004-09-24 09:33:48 +00:00
Pawel Jakub Dawidek
d0257d9c10 Rename 'mount_root_delay' tunable to 'vfs.root.mountdelay', which fits
a bit better to our current naming scheme.

Discussed with:	ru
2004-09-24 09:19:03 +00:00
Poul-Henning Kamp
891822a853 XXX mark two places where we do not hold a threadcount on the dev when
frobbing the cdevsw.

In both cases we examine only the cdevsw and it is a good question if we
weren't better off copying those properties into the cdev in the first
place.  This question will be revisited.
2004-09-24 08:32:36 +00:00
Poul-Henning Kamp
b2deb1d205 Remove the cdevsw() function which is now unused. 2004-09-24 08:30:57 +00:00
Poul-Henning Kamp
6f077de596 Hold threadcount while throbbing cdevsw in our underlying driver.
This is a bit heavyhanded, and will be simplified once the tty code
learns to properly deal with disappearing hw and drivers.
2004-09-24 08:26:03 +00:00
Poul-Henning Kamp
969d098b0f Hold thread reference while we frob cdevsw. 2004-09-24 08:12:41 +00:00
Poul-Henning Kamp
9bd188b936 Hold proper thread count while frobbing drivers ioctl. 2004-09-24 07:24:02 +00:00
Poul-Henning Kamp
8f7bea8b99 Hold threadcount reference when we call into the underlying console
driver.
2004-09-24 07:16:56 +00:00
Poul-Henning Kamp
fe0b82752b Eliminate devsw() call, we are not dereferencing the pointer. 2004-09-24 07:11:02 +00:00
Poul-Henning Kamp
bd8a0d70f4 Remove devsw() call missed in last commit. 2004-09-24 07:08:33 +00:00
Poul-Henning Kamp
af8b19780a Remove SI_ISDISK, I found a better solution. 2004-09-24 06:55:32 +00:00
Poul-Henning Kamp
a7830346e2 Assert topology is held in g_dev_getprovider().
Don't call devsw().  It is not necessary, and we do not need to hold dev_lock
to compare the devsw pointer to our own since we do not dereference it.
2004-09-24 06:43:20 +00:00
Poul-Henning Kamp
f69f5fbd42 Hold thread reference while frobbing cdevsw. 2004-09-24 06:37:00 +00:00
Poul-Henning Kamp
8ff33adb8c Hold threadref while we throb cdevsw in devtoname() 2004-09-24 06:29:23 +00:00
Poul-Henning Kamp
38f878d739 Use vn_isdisk() to check if vnode is a disk.
(repeat, CVS core dumped on me)
2004-09-24 06:23:31 +00:00
Poul-Henning Kamp
233b81be1c use vn_isdisk() to see if vnode is a disk. 2004-09-24 06:21:43 +00:00
Poul-Henning Kamp
6e8d420249 Hold dev_lock and check for NULL devsw pointer when we service FIODTYPE ioctl. 2004-09-24 06:16:48 +00:00
Poul-Henning Kamp
70526ca6a5 Hold dev_lock and check for NULL devsw pointer when we determine
if a vnode is a disk.
2004-09-24 06:16:08 +00:00
Poul-Henning Kamp
751fdd08fe Use dev_re[fl]thread() to maintain a ref on the device driver while
we call the ->d_mmap function.
2004-09-24 05:59:11 +00:00
Poul-Henning Kamp
5ef8cac184 Use def_re[fl]thread().
Retire various old compatibility helpers.
2004-09-24 05:58:06 +00:00
Poul-Henning Kamp
6230ce6aa9 use dev_re[fl]thread() rather than home rolled versions. 2004-09-24 05:55:03 +00:00
Poul-Henning Kamp
2c15afd888 Introduce dev_re[lf]thread() functions.
dev_refthread() will return the cdevsw pointer or NULL.  If the
return value is non-NULL a threadcount is held which much be released
with dev_relthread().  If the returned cdevsw is NULL no threadcount
is held on the device.
2004-09-24 05:54:32 +00:00
Marcel Moolenaar
759e38dbbe Remove. This file was repocopied to ../ski. 2004-09-24 04:49:32 +00:00
Marcel Moolenaar
8da69dadbe Replace misuse of NULL with 0UL. 2004-09-24 04:40:28 +00:00
Marcel Moolenaar
9ba97f7322 Remove the dependency on the Intel EFI headers in sys/boot/efi. Instead
use <machine/efi.h> for the necessary definitions. This makes the EFI
code in sys/boot/efi totally unused, except for pure EFI loaders. As
such, maintenance and porting (to IA-32) of the EFI code is made as easy
as possible.
2004-09-24 04:35:07 +00:00
Marcel Moolenaar
7886b298bf Unhook libski and skiload and hook up ski. 2004-09-24 04:21:14 +00:00
Marcel Moolenaar
cf1dd2e903 Post repocopy build fixes. 2004-09-24 04:06:22 +00:00
Peter Wemm
beafe28690 Implement the KDENABIO etc ioctl's for amd64 as well. 2004-09-24 01:19:11 +00:00
Peter Wemm
d88cc59371 Implement the syscons mouse pointer characters for amd64. It was only
implemented for i386 and alpha.
2004-09-24 01:18:22 +00:00
Peter Wemm
6fdf763cef Like on i386, use the definition of struct bios_smap from machine/pc/bios.h
again.
2004-09-24 01:11:11 +00:00
Peter Wemm
2169193596 Converge towards i386. I originally resisted creating <machine/pc/bios.h>
because it was mostly irrelevant - except for the silly BIOS_PADDRTOVADDR
etc macros.  Along the way of working around this, I missed a few things.

* Make syscons properly inherit the bios capslock/shiftlock/etc state like
  i386 does.  Note that we cannot inherit the bios key repeat rate because
  that requires a bios call (which is impossible for us).
* Give syscons the ability to beep on amd64.  Oops.

While here, make bios.c compile and add it to files.amd64.
2004-09-24 01:08:34 +00:00
Peter Wemm
c3277f936c Severely strip down the repocopied i386/bios.c and bios.h files. It turns
out that bios_sigsearch() etc is useful for finding tables in roms.
2004-09-24 00:42:36 +00:00
John Baldwin
6111dcd2ef A modest collection of various and sundry style, spelling, and whitespace
fixes.

Submitted by:	bde (mostly)
2004-09-24 00:38:15 +00:00
Olivier Houchard
e0370a187c On arm, set the default elf brand to FreeBSD, until the binutils do it for us. 2004-09-23 23:29:24 +00:00
Olivier Houchard
b19d97eb9c Re-route interrupts on arm as well. 2004-09-23 22:58:43 +00:00
Olivier Houchard
2ae0fcd220 Do not use the IO-mapping to issue the reset on the 82546 on arm. For some
reason, it results in corrupted descriptors.
2004-09-23 22:57:53 +00:00
Olivier Houchard
4b7d15c6dc Add the config file for the IQ31244 board. 2004-09-23 22:55:00 +00:00
Olivier Houchard
f67baa4d6b Use the new KERNVIRTADDR and PHYSADDR options.
Add KDB.
2004-09-23 22:53:50 +00:00
Olivier Houchard
107d4bd621 Use -O2 for the arm kernel as well 2004-09-23 22:53:22 +00:00
Olivier Houchard
c60e48e2b6 Use the address provided in the conf file, instead of hardcode 0xc0000000. 2004-09-23 22:52:55 +00:00
Olivier Houchard
c2f29b3e64 Add new options :
PHYSADDR : Address of the physical memory
KERNPHYSADDR : Physical address where the kernel starts
KERNVIRTADDR : Virtual address of the kernel
STARTUP_PAGETABLE_ADDR : Where to put the page table at bootstrap
+ Xscale specific options
2004-09-23 22:52:25 +00:00
Olivier Houchard
332d16a165 Add the possibility to specify the kernel virtual address and the kernel
physical address in the kernel config file, as it varies from CPU to CPU.
2004-09-23 22:47:57 +00:00
Olivier Houchard
6052fa47a9 Import partial support for the IQ31244 eval board (i80321 CPU). IQ80321 might
work out of the box too, but I have no hardware to test.
It works well enough to go multiuser. Network works, SATA does not, as I have
no drive to test.
Thanks to Intel for sending such a board.

Obtained from:  NetBSD
2004-09-23 22:45:36 +00:00
Olivier Houchard
3ce77c124e Add Xscale common headers. 2004-09-23 22:36:13 +00:00
Olivier Houchard
906ce37658 Big cleanup: get ride of the whole spl level logic, as FreeBSD doesn't use
it anymore.
2004-09-23 22:33:38 +00:00
Olivier Houchard
8413603da8 Now that we have pmap_growkernel(), use more KVA. 2004-09-23 22:32:33 +00:00
Olivier Houchard
1f5f31b4ec Remove the empty definition of struct osigcontext, as it will never be used. 2004-09-23 22:31:49 +00:00
Olivier Houchard
7ea7271711 Remove the pcb32_cstate field of struct pcb. 2004-09-23 22:31:08 +00:00
Olivier Houchard
f04d49ad11 Declare sigcode and szsigcode. 2004-09-23 22:30:05 +00:00