Marcel Moolenaar [Sun, 10 Oct 1999 08:38:33 +0000 (08:38 +0000)]
\begin{quote}
setjmp() gets the jmp_buf pointer from the wrong place (the place
where the return address is) in the shlib case, and uses it (only)
to fetch the current signal mask to address (return_address + 28).
This address is normally read-only (I hope), so the sigprocmask()
call has no effect except to return an error code.
\end{quote}
Brian Feldman [Sat, 9 Oct 1999 20:47:59 +0000 (20:47 +0000)]
This implements the RLIMIT_SBSIZE ("sbsize") administrative limits for
userland. Currently, it can be enforced by login and csh. More
shells supporting sbsize are welcome.
Søren Schmidt [Sat, 9 Oct 1999 19:57:13 +0000 (19:57 +0000)]
Add support for the HPT366 chip, this is used on the Abit boards and
their HotRod controller and on SIIG PCI ultra DMA controller. These
changes also made lots of the Promise code go away, its all much more
generic this way.
Get rid of atapi_immed_cmd, instead use the queue to move atapi commands
from interrupt context if nessesary, the entire atapi layer has
gotten an overhaul.
Lots of fixes to utililize the new features in subr_disk.c etc, and
get rid of the last biots of softc arrays in the drivers, the
only one left is atadevices which cannot easily go away (yet).
Use our own malloc names, its a lot easier to track memory usage this way.
- Let physio take read/write compatible args and have it use uio->uio_rw
to determine the direction.
- physread/physwrite are now #defines for physio
- Remove the inversly named minphys(), dev->si_iosize_max takes over.
- Physio() always uses pbufs.
- Fix the check for non page-aligned transfers, now only unaligned
transfers larger than (MAXPHYS - PAGE_SIZE) get fragmented (only
interesting for tapes using max blocksize).
John Hay [Sat, 9 Oct 1999 14:49:56 +0000 (14:49 +0000)]
Update the PPSAPI to draft-mogul-pps-api-05.txt which is the latest.
NOTE: This will break building ntpd until ntpd has been upgraded to also
support draft 05. People that want to build ntpd in the meantime can
get patches from me.
Made mount more userfriendly (bad slashes are now filtered out)
and we remove in mount_nfs trailing slashes if there are any.
Fixed mount_xxx binarys to resolve with realpath(3)
the mountpoint.
Translate the deprecated nfs-syntax with '@' to ':' .
The ':' syntax has now precedence, but '@' still works.
Notify the user that the '@' syntax should not be used.
Bruce Evans [Sat, 9 Oct 1999 11:31:50 +0000 (11:31 +0000)]
Hard-configure z_off_t as long. Soft configuration using HAVE_UNISTD_H
just breaks the prototypes for gzseek() and gztell() when an application
defines HAVE_UNISTD_H before including <zlib.h>. z_off_t was always
long for compiling zlib, but was sometimes off_t for compiling
applications, e.g., Ethereal 0.7.5.
This "fix" preserves bug for bug binary compatibility. z_off_t should
be off_t for everything, but zlib doesn't support off_t being longer
than long, so using the correct type without fixing zlib's internals
would at best break binary compatibility. This "fix" also make the
namespace problems for HAVE_UNISTD_H no worse than hundreds of other
namespace prooblems in zconf.h. I'll wait for a new release of zlib
for proper fixes.
Matthew N. Dodd [Sat, 9 Oct 1999 04:02:02 +0000 (04:02 +0000)]
- Restore correct operation of bt_mca.
- Work around a problem not yet solved in the tree (but solved in mine.)
device_get_ivars() should never be cast to a struct resource_list *
The solution, under review, involves the creation of a
device_get_resource_list() function. More later.
Warner Losh [Sat, 9 Oct 1999 03:48:18 +0000 (03:48 +0000)]
Call DEVICE_IDENIFY in bus_generic_driver_added to allow devices to
add nodes to the tree. Also, default to bus_generic_driver_added for
the BUS_DRIVER_ADDED method.
Wes Peters [Sat, 9 Oct 1999 01:35:58 +0000 (01:35 +0000)]
Correct the description of the timeout argument. I've examined
the code, which seems to implement the POSIX requirements, and
have described the behavior here. Basically, it behaves the same
as select(2).
John Polstra [Sat, 9 Oct 1999 00:25:29 +0000 (00:25 +0000)]
Fix sigvec(). When the sigset_t changes came in, it was altered
to call osigaction(). But that's wrong because it causes the
handler to receive a struct osigcontext instead of the expected
struct sigcontext. Use sigaction() instead, copying the compatible
portion of the signal mask.
Fix a panic(8) implementation:
hexdump -C < /dev/drum
by simply refusing to do I/O from userland.
a panic. I'm not sure we even need /dev/drum anymore, it seems
to have been broken for a long time thi
Doug White [Fri, 8 Oct 1999 04:00:21 +0000 (04:00 +0000)]
Add a new floppy type, install. This type builds a disk that can
slice, partition, newfs, and install FreeBSD from a tarball on a remote server.
Handy for doing mass-installs for server farms.
Nick Hibma [Thu, 7 Oct 1999 19:26:38 +0000 (19:26 +0000)]
Major synchronisation with the NetBSD USB stack:
- Some cleanup and improvements in the uhci and ohci drivers
- Support for plugging and unplugging devices improved
- Now available is bulk transport over OHCI controllers
- Resume and suspend have been temporarily been disabled again. Proper
support for it is available in the uhci.c and ohci.c files but I have
not yet spent the brain cycles to use it.
- OpenBSD now uses the USB stack as well
- Add FreeBSD tags
Nick Hibma [Thu, 7 Oct 1999 18:53:08 +0000 (18:53 +0000)]
Remove last bits and pieces of NetBSD porting.
Remove the use of ukbd_disconnect and replace it with direct
calls to disable.
Change printf to DPRINTF
Remove the use of usbd_lock, change it to splusb()/splx()
Simplification of the signal trampoline and other cleanups.
o Remove unused defines from genassym.c that were needed
by the trampoline.
o Add load_gs_param function to support.s that catches
a fault when %gs is loaded with an invalid descriptor.
The function returns EFAULT in that case.
o Remove struct trapframe from mcontext_t and replace it
with the list of registers.
o Modify sendsig and sigreturn accordingly.
Brian Somers [Thu, 7 Oct 1999 07:32:12 +0000 (07:32 +0000)]
Don't go from datalink state opening directly to lcp, go via
carrier instead so that we can set up our carrier detect
timer and eventually notice when we lose carrier.
Honour the script.run value when coming out of carrier state.
Mike Smith [Thu, 7 Oct 1999 02:23:12 +0000 (02:23 +0000)]
This is a driver for the AMI MegaRAID family of controllers. It all of
the AMI PCI controllers using the 8LD firmware interface (40LD firmware
will be supported as soon as I have hardware to test with).
These controllers are rebadged by Dell as the PERC, as well as by HP
and possibly other vendors.
Mike Smith [Thu, 7 Oct 1999 02:20:32 +0000 (02:20 +0000)]
This is a driver for the Mylex DAC960 family of integrated RAID
controllers. It currently supports the P, PL, PD and PU variants,
with more to be supported shortly.