Daniel Eischen [Fri, 18 Jul 2003 02:46:30 +0000 (02:46 +0000)]
Add a preemption point when a mutex or condition variable is
handed-off/signaled to a higher priority thread. Note that when
there are idle KSEs that could run the higher priority thread,
we still add the preemption point because it seems to take the
kernel a while to schedule an idle KSE. The drawbacks are that
threads will be swapped more often between CPUs (KSEs) and
that there will be an extra userland context switch (the idle
KSE is still woken and will probably resume the preempted
thread). We'll revisit this if and when idle CPU/KSE wakeup
times improve.
Inspired by: Petri Helenius <pete@he.iki.fi>
Reviewed by: davidxu
Ian Dowse [Fri, 18 Jul 2003 01:10:16 +0000 (01:10 +0000)]
When the file system to unmount is specified by device name instead
of by mount point, umount had to take care not to unmount the wrong
file system if another file system was covering the requested one.
Now that the file system to unmount is specified to the kernel using
the filesystem ID, this confusion cannot occur, so remove the code
that checked for it.
Robert Drehmel [Fri, 18 Jul 2003 00:06:10 +0000 (00:06 +0000)]
Make the 80x60, 132x25, 132x43, 132x50, and 132x60 VESA text modes
work when using a graphics chipset which identifies itself as
`VIA CLE266', used in some VIA EPIA boards. Two values need to be
patched in the VESA mode information structure: the widths of the modes
mentioned above are encoded in a format which was unknown to the VESA
module (and to my copy of the VBE spec.) whereas the window memory
segment values seem to be just incorrect.
Changes to allow top to decide whether or not to show multiple threads per
process. Option -H enables it and it is toggled at the interactive
screen by 'H'.
Robert Drehmel [Thu, 17 Jul 2003 23:49:10 +0000 (23:49 +0000)]
To avoid a kernel panic provoked by a NULL pointer dereference,
do not clear the `sb_sel' member of the sockbuf structure
while invalidating the receive sockbuf in sorflush(), called
from soshutdown().
The panic was reproduceable from user land by attaching a knote
with EVFILT_READ filters to a socket, disabling further reads
from it using shutdown(2), and then closing it. knote_remove()
was called to remove all knotes from the socket file descriptor
by detaching each using its associated filterops' detach call-
back function, sordetach() in this case, which tried to remove
itself from the invalidated sockbuf's klist (sb_sel.si_note).
David Xu [Thu, 17 Jul 2003 22:52:55 +0000 (22:52 +0000)]
Fix sigwait to conform to POSIX.
When a signal is being delivered to process, first find a sigwait
thread to deliver, POSIX's argument is speed of delivering signal
to sigwait thread is faster than other ways. A signal in its wait
set will cause sigwait to return the signal number, a signal not
in its wait set but in not blocked by the thread also causes sigwait
to return, but sigwait returns EINTR, sigwait is oneshot operation,
only one signal can be delivered to its wait set, when a signal is
delivered to the sigwait thread, the thread's sigwait state is canceled.
David Xu [Thu, 17 Jul 2003 22:45:33 +0000 (22:45 +0000)]
o Refine kse_thr_interrupt to allow it to handle different commands.
o Remove TDF_NOSIGPOST.
o Add a member td_waitset to proc structure, it will be used for sigwait.
Robert Drehmel [Thu, 17 Jul 2003 22:44:41 +0000 (22:44 +0000)]
Correct six return statements which returned zero instead of
an appropriate error number after a failure condition.
In particular, three of the changed statements return ESRCH for a
failed pfind(), and in also three places a non-zero return
from p_cansee() will be passed back,
Make sure that SWP_NPAGES always has the same value in all source
files, so that SWAP_META_PAGES does not vary either.
swap_pager.c ended up with a value of 16, everybody else 8. Go with
the 16 for now.
This should only have any effect in the "kill processes because we
are out of swap" scenario, where it will make some sort of estimate
of something more precise.
Mark Murray [Thu, 17 Jul 2003 19:37:56 +0000 (19:37 +0000)]
Style(9) cleanup. There was no consistent style in this driver, and
The next round of commits will be to fix up locking in it. This lot
is to at least give a consistent base to work off.
Mike Makonnen [Thu, 17 Jul 2003 11:06:40 +0000 (11:06 +0000)]
Fix umtx locking, for libthr, in the kernel.
1. There was a race condition between a thread unlocking
a umtx and the thread contesting it. If the unlocking
thread won the race it may try to wakeup a thread that
was not yet in msleep(). The contesting thread would then
go to sleep to await a wakeup that would never come. It's
not possible to close the race by using a lock because
calls to casuptr() may have to fault a page in from swap.
Instead, the race was closed by introducing a flag that
the unlocking thread will set when waking up a thread.
The contesting thread will check for this flag before
going to sleep. For now the flag is kept in td_flags,
but it may be better to use some other member or create
a new one because of the possible performance/contention
issues of having to own sched_lock. Thanks to jhb for
pointing me in the right direction on this one.
2. Once a umtx was contested all future locks and unlocks
were happening in the kernel, regardless of whether it
was contested or not. To prevent this from happening,
when a thread locks a umtx it checks the queue for that
umtx and unsets the contested bit if there are no other
threads waiting on it. Again, this is slightly more
complicated than it needs to be because we can't hold
a lock across casuptr(). So, the thread has to check
the queue again after unseting the bit, and reset the
contested bit if it finds that another thread has put
itself on the queue in the mean time.
3. Remove the if... block for unlocking an uncontested
umtx, and replace it with a KASSERT. The _only_ time
a thread should be unlocking a umtx in the kernel is
if it is contested.
Fix the ski loader, broken by the gcc upgrade. Update the linker
script to match the one for the EFI loader and rewrite __start()
in assembly to have gp defined without getting in the way of the
compiler.
Have the linker script look more like the default linker script
on ia64. This fixes the breakage caused by the gcc upgrade that
resulted in a broken executable.
Mark Murray [Wed, 16 Jul 2003 20:59:15 +0000 (20:59 +0000)]
Very big makeover in the way telnet, telnetd and libtelnet are built.
Previously, there were two copies of telnet; a non-crypto version
that lived in the usual places, and a crypto version that lived in
crypto/telnet/. The latter was built in a broken manner somewhat akin
to other "contribified" sources. This meant that there were 4 telnets
competing with each other at build time - KerberosIV, Kerberos5,
plain-old-secure and base. KerberosIV is no longer in the running, but
the other three took it in turns to jump all over each other during a
"make buildworld".
As the crypto issue has been clarified, and crypto _calls_ are not
a problem, crypto/telnet has been repo-copied to contrib/telnet,
and with this commit, all telnets are now "contribified". The contrib
path was chosen to not destroy history in the repository, and differs
from other contrib/ entries in that it may be worked on as "normal"
BSD code. There is no dangerous crypto in these sources, only a
very weak system less strong than enigma(1).
Kerberos5 telnet and Secure telnet are now selected by using the usual
macros in /etc/make.conf, and the build process is unsurprising and
less treacherous.
Change the style of the english used to print accounting enabled
and disabled. This means no period at the end and changing
"Process accounting <foo>" to "Accounting <foo>".
Bill Paul [Wed, 16 Jul 2003 00:09:56 +0000 (00:09 +0000)]
Add support for the BCM5705 and its ilk. Changes:
- 5705 doesn't support jumbo frames
- Statistics must be read from registers
- RX return ring must be capped at 512 entries
- Omit initialization of certain device blocks
- Acknowledge link change interrupts by setting the 'link changed'
bit in the status register (used to have no effect)
- Remember to toggle the MI completion bit too
- Set the mbuf low watermark differently (on-chip memory buffers,
not BSD mbufs)
- Don't enable Ethernet@WireSpeed feature for certain 5705 chip revs
- Add additional PCI IDs for 5705 and 5782 parts
- Add a forgotten 5704 PCI ID
Most changes ripped kicking and screaming from the Broadcom linux driver.
Thanks to Paul Saab for sanity testing. (My lack of sanity has been
confirmed.)
add support for interrupt counting on sparc64. This copies part of the
code from i386. The code has a slight bogon that interrupts are counted
twice. Once on the ithread dispatch and once on the dispatch for the vector
vmstat -i and systat -vm now contains interrupt counts.
Luigi Rizzo [Tue, 15 Jul 2003 23:08:44 +0000 (23:08 +0000)]
Userland side of:
Allow set 31 to be used for rules other than 65535.
Set 31 is still special because rules belonging to it are not deleted
by the "ipfw flush" command, but must be deleted explicitly with
"ipfw delete set 31" or by individual rule numbers.
This implement a flexible form of "persistent rules" which you might
want to have available even after an "ipfw flush".
Note that this change does not violate POLA, because you could not
use set 31 in a ruleset before this change.
Luigi Rizzo [Tue, 15 Jul 2003 23:07:34 +0000 (23:07 +0000)]
Allow set 31 to be used for rules other than 65535.
Set 31 is still special because rules belonging to it are not deleted
by the "ipfw flush" command, but must be deleted explicitly with
"ipfw delete set 31" or by individual rule numbers.
This implement a flexible form of "persistent rules" which you might
want to have available even after an "ipfw flush".
Note that this change does not violate POLA, because you could not
use set 31 in a ruleset before this change.
sbin/ipfw changes to allow manipulation of set 31 will follow shortly.
Change the msleep part of EcWaitEvent to be a separate loop, fixing a
problem that for some very slow ECs (~6 ms occasionally) causes a
timeout. Also finish resource cleanup in the error case in attach.
Fix the ACPI_DEBUG build for the non-module case. Move the #define into
acfreebsd.h and remove it from the Makefile. Now ACPI_DEBUG implies
ACPI_DISASSEMBLER.
David Malone [Tue, 15 Jul 2003 16:46:45 +0000 (16:46 +0000)]
Get the connections per minute calculation right. By good fortune
(or possibly testing) the previous formula worked for the default
constants compiled into inetd, but if you recompiled with different
values of CHTSIZE and CHTGRAN the calculation might not have worked.
PR: 54354
Submitted by: Claus Assmann <ca@sendmail.org>
Submitted by: Jose Marcio Martins da Cruz <Jose-Marcio.Martins@ensmp.fr>
MFC after: 5 days
Instead of returning an error call the ioctl() handler of the interface
when we get request that we cannot handle ourself. This allows userland
to reach the ATM interfaces for ioctls.
Be careful to call bus_dmamap_load with BUS_DMA_NOWAIT so that the
callback will never be deferred. ATM needs to prevent cell and packet
ordering. Also use the default mutex and lock functions (those that
panic) for the tag creation.
This is a driver for IDT77252 based ATM interfaces. It has been tested
with a ProATM-155 and an IDT evaluation board and should also work
with a ProATM-25 (it seems to work at least, I cannot really measure
what the card emits). The driver has been tested on i386 and sparc64,
but should work an other archs also. It supports UBR, CBR, ABR and VBR;
AAL0, AAL5 and AALraw. As an additional feature VCI/VPI 0/0 can be
opened for receiving in AALraw mode and receives all cells not claimed
by other open VCs (even cells with invalid GFC, VPI and VCI fields and
OAM cells).
Thanks to Christian Bucari from ProSum for lending two cards and answering
my questions.
Implement an utility function that can be used by device drivers to
implement the ATMIOCGVCCS ioctls. This routine handles changing
VCC tables (which can occure because we cannot hold the driver mutex
while allocating memory) with a loop and a re-allocation, should the
table not fit in the allocated memory.
Add a facility for devices, specifically network interfaces, that require
large to huge amounts of small or medium sized receive buffers. The problem
with these situations is that they eat up the available DMA address space
very quickly when using mbufs or even mbuf clusters. Additionally this
facility provides a direct mapping between 32-bit integers and these buffers.
This is needed for devices originally designed for 32-bit systems. Ususally
the virtual address of the buffer is used as a handle to find the buffer as
soon as it is returned by the card. This does not work for 64-bit machines
and hence this mapping is needed.
More changes to use __FBSDID() for setting rcsids, and fix the
format of 'sccsid' lines so they consistently match style(9).
A minor Makefile change is needed so lptest.c can find lp.cdefs.h.
Reviewed by: discussed with bde and obrien
MFC after: 15 days
More changes to use __FBSDID() for setting rcsids, and fix the
format of 'sccsid' lines so they consistently match style(9)
Also the 'sccsid' line is formatted to match style(9), and
a 'From:' is removed so the sccsid returns to what it was back
in the days of '-r CSRG' (1996).
Reviewed by: discussed with bde and obrien
MFC after: 15 days
More changes to use __FBSDID() for setting rcsids. For the filter
programs, minor Makefile changes are needed to find lp.cdefs.h.
For lpf.c, the 'sccsid' line is formatted to match style(9), and
a 'From:' is removed so the sccsid returns to what it was back
in the days of '-r CSRG' (1994).
Reviewed by: discussed with bde and obrien
MFC after: 15 days