]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoA failure to allocate memory for auxiliary TCP data is now fatal.
ru [Thu, 19 Oct 2000 10:44:44 +0000 (10:44 +0000)]
A failure to allocate memory for auxiliary TCP data is now fatal.
This fixes a null pointer dereference problem that is unlikely to
happen in normal circumstances.

23 years agoNEWCARD/Cardbus -
jon [Thu, 19 Oct 2000 08:34:32 +0000 (08:34 +0000)]
NEWCARD/Cardbus -

This commit adds support for Xircom X3201 based cardbus cards.
Support for the TDK 78Q2120 MII is also added.
IBM Etherjet, Intel and Xircom cards uses these chips.

Note that as a result of this commit, some Intel/DEC 21143 based cardbus
cards will also attach, but not get link.  That is being looked at.

23 years agoMove init_clocks earlier in the system startup so that hardclock and clock
jhb [Thu, 19 Oct 2000 08:16:34 +0000 (08:16 +0000)]
Move init_clocks earlier in the system startup so that hardclock and clock
interrupts are started before the device probe.  This allows interrupt
threads to run during the device probe among other things.

23 years agoWrite the routed interrupt back to PCI configuration space.
msmith [Thu, 19 Oct 2000 08:07:23 +0000 (08:07 +0000)]
Write the routed interrupt back to PCI configuration space.

23 years agoCall the BIOS to route the selected interrupt. Correctly calculate the
msmith [Thu, 19 Oct 2000 08:06:50 +0000 (08:06 +0000)]
Call the BIOS to route the selected interrupt.  Correctly calculate the
interrupt from the PCI routing table (ffs returns 1 for the rightmost
bit, not 0).

23 years agoAdd PCI BIOS function codes for IRQ routing fetch and route.
msmith [Thu, 19 Oct 2000 08:02:46 +0000 (08:02 +0000)]
Add PCI BIOS function codes for IRQ routing fetch and route.

23 years agoo Introduce new VOP_ACCESS() flag VADMIN, allowing file systems to perform
rwatson [Thu, 19 Oct 2000 07:53:59 +0000 (07:53 +0000)]
o Introduce new VOP_ACCESS() flag VADMIN, allowing file systems to perform
  "administrative" authorization checks.  In most cases, the VADMIN test
  checks to make sure the credential effective uid is the same as the file
  owner.
o Modify vaccess() to set VADMIN as an available right if the uid is
  appropriate.
o Modify references to uid-based access control operations such that they
  now always invoke VOP_ACCESS() instead of using hard-coded policy checks.
o This allows alternative UFS policies to be implemented by replacing only
  ufs_access() (such as mandatory system policies).
o VOP_ACCESS() requires the caller to hold an exclusive vnode lock on the
  vnode: I believe that new invocations of VOP_ACCESS() are always called
  with the lock held.
o Some direct checks of the uid remain, largely associated with the QUOTA
  and SUIDDIR code.

Reviewed by: eivind
Obtained from: TrustedBSD Project

23 years agoAxe the idle_event eventhandler, and add a MD cpu_idle function used
jhb [Thu, 19 Oct 2000 07:47:16 +0000 (07:47 +0000)]
Axe the idle_event eventhandler, and add a MD cpu_idle function used
for things such as halting CPU's, idling CPU's, etc.

Discussed with: msmith

23 years agoAdd Daniel's name and fix release data
roger [Thu, 19 Oct 2000 07:34:18 +0000 (07:34 +0000)]
Add Daniel's name and fix release data

23 years agoUpdate to driver 2.17
roger [Thu, 19 Oct 2000 07:33:28 +0000 (07:33 +0000)]
Update to driver 2.17
Fixes bugs in devfs when unloading and reloading
Syncs with NetBSD changes

Submitted by: Alexander Langer <alex@big.endian.de>
Submitted by: Thomas Klausner <wiz@netbsd.org>
Submitted by: Daniel O'Connor" <doconnor@gsoft.com.au>

23 years agoObCommit from BSDCon: Update SMPng note and add an entry for devfs,
bmah [Wed, 18 Oct 2000 23:45:31 +0000 (23:45 +0000)]
ObCommit from BSDCon:  Update SMPng note and add an entry for devfs,
which hopefully reflect something of what grog and phk presented in their
talks.

23 years agoMake these PICOBSD's compile... well, almost. Still some work to
joe [Wed, 18 Oct 2000 22:02:33 +0000 (22:02 +0000)]
Make these PICOBSD's compile... well, almost.  Still some work to
do on PPP (atm problems.)

23 years agoReference Larry Wall's perlstyle man page, and recommend it's use where
joe [Wed, 18 Oct 2000 21:57:37 +0000 (21:57 +0000)]
Reference Larry Wall's perlstyle man page, and recommend it's use where
entries in this guide don't override it.

Requested by: markm

Comment on tabulation style.

23 years agoTrim index lines down to 79 characters, not 80.
des [Wed, 18 Oct 2000 20:16:26 +0000 (20:16 +0000)]
Trim index lines down to 79 characters, not 80.

Approved by: jkh

23 years agoMinor changes to the ATA RAID support code, remove some verbosity
sos [Wed, 18 Oct 2000 18:49:42 +0000 (18:49 +0000)]
Minor changes to the ATA RAID support code, remove some verbosity
and put some under bootverbose..

23 years agoUse 'unless' instead of 'if (! ... )' where it improves readability.
joe [Wed, 18 Oct 2000 18:01:35 +0000 (18:01 +0000)]
Use 'unless' instead of 'if (! ... )' where it improves readability.

23 years agoEVENTHANDLER_INVOKE() takes two arguments.
peter [Wed, 18 Oct 2000 17:56:06 +0000 (17:56 +0000)]
EVENTHANDLER_INVOKE() takes two arguments.

23 years agoOpening braces should be at the end of the controlling line. Else
joe [Wed, 18 Oct 2000 17:48:10 +0000 (17:48 +0000)]
Opening braces should be at the end of the controlling line.  Else
and elsif belong on the same line as the closing brace for the
previous if or elsif block.

Suggested by: des

23 years agoThis didn't compile. Fix typo: s/rmang_get_start/rman_get_start/
peter [Wed, 18 Oct 2000 17:45:29 +0000 (17:45 +0000)]
This didn't compile. Fix typo: s/rmang_get_start/rman_get_start/

23 years agoContract a statement that's too verbose.
joe [Wed, 18 Oct 2000 17:34:43 +0000 (17:34 +0000)]
Contract a statement that's too verbose.

23 years agoThere are enough examples of using 'my $var = ...' definitions, so remove
joe [Wed, 18 Oct 2000 17:32:21 +0000 (17:32 +0000)]
There are enough examples of using 'my $var = ...' definitions, so remove
one.

23 years agoAll variables should be commented.
joe [Wed, 18 Oct 2000 17:25:59 +0000 (17:25 +0000)]
All variables should be commented.
Local variables should be separated from function arguments by a
blank line.

Suggested by: des

23 years agoMake subroutine prototypes mandatory.
joe [Wed, 18 Oct 2000 17:21:54 +0000 (17:21 +0000)]
Make subroutine prototypes mandatory.
The main code should be wrapped in a MAIN:{ ... } declaration.

Suggested by: des

23 years agoFix the places that I missed last commit: 'vars' not 'var'.
joe [Wed, 18 Oct 2000 17:12:07 +0000 (17:12 +0000)]
Fix the places that I missed last commit: 'vars' not 'var'.

23 years agoConverted `da' and `wd' into `rda' and `rwd', respectively.
kato [Wed, 18 Oct 2000 15:28:10 +0000 (15:28 +0000)]
Converted `da' and `wd' into `rda' and `rwd', respectively.

Submitted by: MURAMATSU Atsushi <amura@ma3.seikyou.ne.jp>

23 years agoIf we do not byte-swap the ip_id in the first place, don't do it in
ru [Wed, 18 Oct 2000 11:36:09 +0000 (11:36 +0000)]
If we do not byte-swap the ip_id in the first place, don't do it in
the second.  NetBSD (from where I've taken this originally) needs
to fix this too.

23 years agoAttempt to fix the random read blocking. The old code slept at
peter [Wed, 18 Oct 2000 10:39:18 +0000 (10:39 +0000)]
Attempt to fix the random read blocking.  The old code slept at
priority "0" and without PCATCH, so it was uninterruptable.  And
even when it did wake up after entropy arrived, it exited after the
wakeup without actually reading the freshly arrived entropy.  I
sent this to Mark before but it seems he is in transit.
Mark: feel free to replace this if it gets in your way.

23 years agoAdd in a simple API for memory barriers to machine/bus.h:
jhb [Wed, 18 Oct 2000 10:30:12 +0000 (10:30 +0000)]
Add in a simple API for memory barriers to machine/bus.h:
- barrier_read() enforces a memory read barrier
- barrier_write() enforces a memory write barrier
- barrier_rw() enforces a memory read/write barrier

23 years agoGo ahead and disable the legacy mapping of the pcic registers on the
imp [Wed, 18 Oct 2000 09:55:39 +0000 (09:55 +0000)]
Go ahead and disable the legacy mapping of the pcic registers on the
RF5C47x.  Before the code would only do this on the RF5C46x bridges.

Submitted by: Jonathan Chen

23 years agoAdd ifpnp driver to list of i4b hardware drivers.
hm [Wed, 18 Oct 2000 09:16:30 +0000 (09:16 +0000)]
Add ifpnp driver to list of i4b hardware drivers.

23 years agoMerged from sys/i386/i386/machdep.c revision 1.416.
kato [Wed, 18 Oct 2000 09:05:09 +0000 (09:05 +0000)]
Merged from sys/i386/i386/machdep.c revision 1.416.

23 years agoAdd a debug flag bit for a new driver.
hm [Wed, 18 Oct 2000 08:39:24 +0000 (08:39 +0000)]
Add a debug flag bit for a new driver.

23 years agoDon't needlessly pass the diagnostic counter to the idle_event event
jhb [Wed, 18 Oct 2000 08:10:25 +0000 (08:10 +0000)]
Don't needlessly pass the diagnostic counter to the idle_event event
handlers.

23 years agoUse dd to better block the data being sent to /dev/random.
jwd [Wed, 18 Oct 2000 05:29:26 +0000 (05:29 +0000)]
Use dd to better block the data being sent to /dev/random.
Note: I am not redirecting stdout so that the amount of
      data being transferred can be seen (22k and 576k currently).

23 years agoAdd new bus method 'GET_RESOURCE_LIST' and appropriate generic
mdodd [Wed, 18 Oct 2000 05:15:40 +0000 (05:15 +0000)]
Add new bus method 'GET_RESOURCE_LIST' and appropriate generic
implementation.

Add bus_generic_rl_{get,set,delete,release,alloc}_resource() functions
which provide generic operations for devices using resource list style
resource management.

This should simplify a number of bus drivers.  Further commits to follow.

23 years agoInitial commit of NEWCARD cardbus side (that actually compiles and works)
jon [Wed, 18 Oct 2000 03:37:21 +0000 (03:37 +0000)]
Initial commit of NEWCARD cardbus side (that actually compiles and works)

Files:
dev/cardbus/cardbus.c
dev/cardbus/cardbusreg.h
dev/cardbus/cardbusvar.h
dev/cardbus/cardbus_cis.c
dev/cardbus/cardbus_cis.h
dev/pccbb/pccbb.c
dev/pccbb/pccbbreg.h
dev/pccbb/pccbbvar.h
dev/pccbb/pccbb_if.m

This should support:
  - cardbus controllers:
    * TI 113X
    * TI 12XX
    * TI 14XX
    * Ricoh 47X
    * Ricoh 46X
    * ToPIC 95
    * ToPIC 97
    * ToPIC 100
    * Cirrus Logic CLPD683x
  - cardbus cards
    * 3c575BT
    * 3c575CT
    * Xircom X3201 (includes IBM, Xircom and, Intel cards)
    [ 3com support already in kernel, Xircom will be committed real soon now]

This doesn't work with 16bit pccards under NEWCARD.

Enable in your config by having "device pccbb" and "device cardbus".
(A "device pccard" will attach a pccard bus, but it means you system have
a high chance of panicing when a 16bit card is inserted)

It should be fairly simple to make a driver attach to cardbus under
NEWCARD -- simply add an entry for attaching to cardbus on a new
DRIVER_MODULE and add new device IDs as necessary.  You should also make
sure the card can be detached nicely without the interrupt routine doing
something weird, like going into an infinite loop.  Usually that should
entail adding an additional check when a pci register or the bus space is
read to check if it equals 0xffffffff.

Any problems, please let me know.

Reviewed by: imp

23 years agoInitial commit of NEWCARD cardbus side (that actually compiles and works)
jon [Wed, 18 Oct 2000 03:25:13 +0000 (03:25 +0000)]
Initial commit of NEWCARD cardbus side (that actually compiles and works)

Files:
dev/cardbus/cardbus.c
dev/cardbus/cardbusreg.h
dev/cardbus/cardbusvar.h
dev/cardbus/cardbus_cis.c
dev/cardbus/cardbus_cis.h
dev/pccbb/pccbb.c
dev/pccbb/pccbbreg.h
dev/pccbb/pccbbvar.h
dev/pccbb/pccbb_if.m

This should support:
  - cardbus controllers:
    * TI 113X
    * TI 12XX
    * TI 14XX
    * Ricoh 47X
    * Ricoh 46X
    * ToPIC 95
    * ToPIC 97
    * ToPIC 100
    * Cirrus Logic CLPD683x
  - cardbus cards
    * 3c575BT
    * 3c575CT
    * Xircom X3201 (includes IBM, Xircom and, Intel cards)
    [ 3com support already in kernel, Xircom will be committed real soon now]

This doesn't work with 16bit pccards under NEWCARD.

Enable in your config by having "device pccbb" and "device cardbus".
(A "device pccard" will attach a pccard bus, but it means you system have
a high chance of panicing when a 16bit card is inserted)

It should be fairly simple to make a driver attach to cardbus under
NEWCARD -- simply add an entry for attaching to cardbus on a new
DRIVER_MODULE and add new device IDs as necessary.  You should also make
sure the card can be detached nicely without the interrupt routine doing
something weird, like going into an infinite loop.  Usually that should
entail adding an additional check when a pci register or the bus space is
read to check if it equals 0xffffffff.

Any problems, please let me know.

Reviewed by: imp

23 years agoMove auto doesn't work comment closer to config line
imp [Wed, 18 Oct 2000 03:14:37 +0000 (03:14 +0000)]
Move auto doesn't work comment closer to config line

23 years agoRemove some silly debug writes and code that we don't need
imp [Wed, 18 Oct 2000 02:22:39 +0000 (02:22 +0000)]
Remove some silly debug writes and code that we don't need

23 years agoo Remove a boatload of debugging printfs.
imp [Wed, 18 Oct 2000 02:21:00 +0000 (02:21 +0000)]
o Remove a boatload of debugging printfs.
o Report function number and config index on probe line
o Activate the resources (I hope) when RF_ACTIVE is set on those resources
  I'm allocating on behalf of my children.
o Always enable interrupts on multifunction cards in the multifunction
  register.

23 years agoUse appropriate resource management accessors instead of directly
mdodd [Wed, 18 Oct 2000 00:09:26 +0000 (00:09 +0000)]
Use appropriate resource management accessors instead of directly
referencing structure members.

Use rman_get_size() instead of end - start + 1.

23 years agoAdd rman_get_size(), which does what you would expect.
mdodd [Tue, 17 Oct 2000 23:45:28 +0000 (23:45 +0000)]
Add rman_get_size(), which does what you would expect.

Further commits to make use of this will follow.

23 years ago- Wrap the sanity checks for staying in the idle loop for absurdly long
jhb [Tue, 17 Oct 2000 23:12:37 +0000 (23:12 +0000)]
- Wrap the sanity checks for staying in the idle loop for absurdly long
  amounts of time in #ifdef DIAGNOSTIC
- Call vm_page_zero_idle() during the idle loop.

23 years ago- Catch up to moving headers, machine/ipl.h -> sys/ipl.h
jhb [Tue, 17 Oct 2000 23:10:23 +0000 (23:10 +0000)]
- Catch up to moving headers, machine/ipl.h -> sys/ipl.h
- Fix some whitespace bogons.

Submitted by: bde (2)

23 years agoWork much harder at obtaining a correct device node. If what devname(3)
des [Tue, 17 Oct 2000 22:43:41 +0000 (22:43 +0000)]
Work much harder at obtaining a correct device node. If what devname(3)
returns doesn't exist, or doesn't match the desired device, fall back to
scanning /dev for a matching node, and as a last resort, if that fails,
try to create the node ourselves as /dev/dump.

Add comments to several variables and functions.

Clean up syslog(3) usage; use %m instead of strerror(3).

Other minor cleanup.

23 years agoAdd a bunch of wireless cards from Japan. Shigeru YAMAMOTO-san
imp [Tue, 17 Oct 2000 22:31:27 +0000 (22:31 +0000)]
Add a bunch of wireless cards from Japan.  Shigeru YAMAMOTO-san
brought them to BSDcon and gratiously allowed me to test each one with
the wireless setup here.  In addition, support another CF card which
YAMAMOTO-san also brought.

Also add a Lexar CF card which didn't work with auto on many laptops.

23 years agoRemove random from boot floppy on advice of Mark Murray
jkh [Tue, 17 Oct 2000 22:16:15 +0000 (22:16 +0000)]
Remove random from boot floppy on advice of Mark Murray

23 years agoImplement resource alignment as discussed in arch@ a long time ago.
imp [Tue, 17 Oct 2000 22:08:03 +0000 (22:08 +0000)]
Implement resource alignment as discussed in arch@ a long time ago.
This was implemented by Shigeru YAMAMOTO-san and Jonathan Chen.  I've
cleaned them up somewhat and they seem to work well enough to boot
current (but given current's state it can be hard to tell).  Doug
Rabson also reviewed the design and signed off on it.

23 years agoClarify the location of the kernel and modules.
bmah [Tue, 17 Oct 2000 20:32:04 +0000 (20:32 +0000)]
Clarify the location of the kernel and modules.

Submitted by: dcs

23 years agoArgh! Fix passive mode selection (again)
des [Tue, 17 Oct 2000 19:01:18 +0000 (19:01 +0000)]
Argh! Fix passive mode selection (again)

23 years agoRoll minor revision- for once we'll use this because.... if revision >= 5.4,
mjacob [Tue, 17 Oct 2000 18:18:14 +0000 (18:18 +0000)]
Roll minor revision- for once we'll use this because.... if revision >= 5.4,
compile time will build in mutex locks, otherwise the old locking (splcam/splx
with a recursion counter) will be compiled in.

We still depend on config_intr_hook to tell us when it's okay to call
msleep instead of polling. It'd be real nice if we could do this early
enough to not hang up a machine struggling with a bad Fibre Channel loop,
but that's still to come.

23 years agoremove "SERVICING_INTERRUPT" nonsense
mjacob [Tue, 17 Oct 2000 18:15:30 +0000 (18:15 +0000)]
remove "SERVICING_INTERRUPT" nonsense

23 years agoFix the build break.
n_hibma [Tue, 17 Oct 2000 17:13:38 +0000 (17:13 +0000)]
Fix the build break.

The solution isn't perfect, instead of printing the first report, it
will the report no report id, but the proper solution is more complex
than this. We would need to iterate over the report descriptor and
figure out all the report id's before starting to print.

23 years agoCorrection: 'use vars ..' not 'use var'.
joe [Tue, 17 Oct 2000 15:50:22 +0000 (15:50 +0000)]
Correction: 'use vars ..' not 'use var'.

Submitted by: des

23 years agoBuild and install the perl.style(7) manual page.
joe [Tue, 17 Oct 2000 15:33:29 +0000 (15:33 +0000)]
Build and install the perl.style(7) manual page.

23 years agoReclarify variable definition blocks.
joe [Tue, 17 Oct 2000 15:32:57 +0000 (15:32 +0000)]
Reclarify variable definition blocks.
Ask programmers to use modules where possible instead of reinventing
  the wheel.
Use 'chomp' not 'chop' please.
Fixup some mdoc.

23 years agoSyslogd normally converts messages of facility kern to facility
dwmalone [Tue, 17 Oct 2000 10:19:42 +0000 (10:19 +0000)]
Syslogd normally converts messages of facility kern to facility
user unless they come directly from the kernel. Document this and
add a flag to syslogd which prevents this conversion.

Sort getopt args while I'm at it.

PR: 21788
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>

23 years agoFix spelling of comparison.
dwmalone [Tue, 17 Oct 2000 10:14:51 +0000 (10:14 +0000)]
Fix spelling of comparison.

23 years agoImplement write combining for crashdumps. This is useful when
ps [Tue, 17 Oct 2000 10:05:49 +0000 (10:05 +0000)]
Implement write combining for crashdumps.  This is useful when
write caching is disabled on both SCSI and IDE disks where large
memory dumps could take up to an hour to complete.

Taking an i386 scsi based system with 512MB of ram and timing (in
seconds) how long it took to complete a dump, the following results
were obtained:

Before: After:
WCE           TIME WCE           TIME
------------------ ------------------
1 141.820972 1  15.600111
0 797.265072 0  65.480465

Obtained from: Yahoo!
Reviewed by: peter

23 years agoFix error in sample code in the man page
roger [Tue, 17 Oct 2000 08:27:18 +0000 (08:27 +0000)]
Fix error in sample code in the man page

23 years agoAdd new bktr options
roger [Tue, 17 Oct 2000 07:59:56 +0000 (07:59 +0000)]
Add new bktr options

23 years agoTry and get libc_r to compile again on the alpha after deischen's commit
peter [Tue, 17 Oct 2000 06:31:40 +0000 (06:31 +0000)]
Try and get libc_r to compile again on the alpha after deischen's commit

23 years agoMore NEWCARD fixes. We now almost properly print the probe message.
imp [Tue, 17 Oct 2000 06:29:21 +0000 (06:29 +0000)]
More NEWCARD fixes.  We now almost properly print the probe message.
o Remember the resources we allocate for the config entry.
o When we get the resource, do an resource_list_add and do a
  resource_list_delete if we fail later in the resource list.
o In the pccard bus, we allocate the resources.  When a child asks for
  them, just return the resources that we allocated (thanks to Paul
  Richards and Mike Smith for the idea).

23 years agouse ``config auto'' for NCR WaveLAN/IEEE
sanpei [Tue, 17 Oct 2000 06:20:21 +0000 (06:20 +0000)]
use ``config auto'' for NCR WaveLAN/IEEE

23 years agoAn initial draft of a style guide for perl scripts in the tree.
joe [Tue, 17 Oct 2000 02:51:03 +0000 (02:51 +0000)]
An initial draft of a style guide for perl scripts in the tree.

23 years agoImplement simple write combining for newfs - this is particularly useful
peter [Tue, 17 Oct 2000 00:41:36 +0000 (00:41 +0000)]
Implement simple write combining for newfs - this is particularly useful
for large scsi disks with WCE = 0.  This yields around a 7 times speedup
on elapsed newfs time on test disks here.  64k clusters seems to be the
sweet spot for scsi disks using our present drivers.

23 years agoSome linux apps, such as IBM's JDK 1.3, will attempt to mmap thread
gallatin [Tue, 17 Oct 2000 00:25:43 +0000 (00:25 +0000)]
Some linux apps, such as IBM's JDK 1.3, will attempt to mmap thread
stacks near the top of their address space.  If their TOS is greater
than vm_maxsaddr, vm_map_growstack() will confuse the thread stack
with the process stack and deliver a SEGV if they attempt to grow the
thread stack past their current stacksize rlimit.  To avoid this,
adjust vm_maxsaddr upwards to reflect the current stacksize rlimit
rather than the maximum possible stacksize.  It would be better to
adjust the mmap'ed region, but some apps (again, IBM's JDK 1.3) do not
check mmap's return value..

This commit (in conjunction with setting MINSIGSTKSZ to 2048 &
rebuilding your kernel and modules) will get IBM's JDK 1.3 working
with FreeBSD at least well enough to run many of the example applets.

Reviewed by: marcel
Tested by:   sto@stat.duke.edu, many others on freebsd-java@

23 years agoMention the new Adaptec RAID controllers that are supported by the asr driver.
scottl [Tue, 17 Oct 2000 00:15:00 +0000 (00:15 +0000)]
Mention the new Adaptec RAID controllers that are supported by the asr driver.

Submitted by:  mark_salyzyn@adaptec.com

23 years agoregen after addition of linux_rt_sigreturn
gallatin [Tue, 17 Oct 2000 00:03:02 +0000 (00:03 +0000)]
regen after addition of linux_rt_sigreturn

23 years agoA start at an implemention of linux_rt_sendsig & linux_rt_sigreturn
gallatin [Tue, 17 Oct 2000 00:00:20 +0000 (00:00 +0000)]
A start at an implemention of linux_rt_sendsig & linux_rt_sigreturn
and associated user-level signal trampoline glue.

Without this patch, an SA_SIGINFO style handler can be installed by a linux
app, but if the handler accesses its sip argument, it will get a garbage
pointer and likely segfault.

We currently supply a valid pointer, but its contents are mainly
garbage.  Filling this in properly is future work.

This is the second of 3 commits that will get IBM's JDK 1.3 working with
FreeBSD ...

23 years agoAdd support for cardbus card's chips. This will make the 3c575 cards
imp [Mon, 16 Oct 2000 23:16:02 +0000 (23:16 +0000)]
Add support for cardbus card's chips.  This will make the 3c575 cards
work once the rest of the cardbus infrastructure has been committed.

Submitted by: Jonathan Chen <jon@spook.org>

23 years agoVery early and very *very* lightly tested support for LIVENGOOD chipset
mjacob [Mon, 16 Oct 2000 23:08:45 +0000 (23:08 +0000)]
Very early and very *very* lightly tested support for LIVENGOOD chipset
(followon to WISEMAN). Presumably some flavors are also no multimode copper
as well.

23 years agoWhen wierdreset flag is set, turn on the DISADVFD flag when we reset
imp [Mon, 16 Oct 2000 23:02:21 +0000 (23:02 +0000)]
When wierdreset flag is set, turn on the DISADVFD flag when we reset
rather than all the flags.  This prevents setting being read from ROM,
which is a problem.  If this breaks anything, it will only break the
3C556B cards minipci cards, which mainly exist at rpi as far as rpi
has been able to tell.

Submitted by: Louis Gerbarg <gerbal@rpi.edu>

23 years agoFix problems booting large kernels on alphas. The symptom is that the kernel
gallatin [Mon, 16 Oct 2000 20:15:43 +0000 (20:15 +0000)]
Fix problems booting large kernels on alphas.   The symptom is that the kernel
loads, prints the copyright, and either hangs or locks solid.  The
PC tends to be in the data segment and the RA is in XentMM

Doug really came up with the fix, I'm just the monkey typing.  Doug says:
The alpha can only support 64k of globals with $gp pointing at
base+32k so that the code can use 16bit signed offsets from $gp to
access it. ....  it is possible to have multiple .got subsections
and the linker handles this with the relocations for 'ldgp' pseudo
instructions.  [Without this patch] the code in exception.s has been
linked  to use a different gp from locore.s (where pal_kgp is set).

Reviewed by: dfr

23 years agoAdd section for building ia64 kernels.
dfr [Mon, 16 Oct 2000 20:04:52 +0000 (20:04 +0000)]
Add section for building ia64 kernels.

23 years agoThe PowerPC conversion was
obrien [Mon, 16 Oct 2000 20:03:53 +0000 (20:03 +0000)]
The PowerPC conversion was

Submitted by: Benno Rice <benno@jeamland.net>

23 years agoConvert from the Alpha compontents to PowerPC ones.
obrien [Mon, 16 Oct 2000 19:58:32 +0000 (19:58 +0000)]
Convert from the Alpha compontents to PowerPC ones.

23 years agoRepo copied from src/sys/boot/alpha/libalpha/alpha_module.c,v
obrien [Mon, 16 Oct 2000 19:55:06 +0000 (19:55 +0000)]
Repo copied from src/sys/boot/alpha/libalpha/alpha_module.c,v

23 years agoRepo copied from src/sys/boot/alpha/libalpha/prom.c,v
obrien [Mon, 16 Oct 2000 19:51:48 +0000 (19:51 +0000)]
Repo copied from src/sys/boot/alpha/libalpha/prom.c,v

23 years agoRepo copied from src/sys/boot/alpha/libalpha/devicename.c,v
obrien [Mon, 16 Oct 2000 19:51:29 +0000 (19:51 +0000)]
Repo copied from src/sys/boot/alpha/libalpha/devicename.c,v

23 years agoAdd types and prototypes.
imp [Mon, 16 Oct 2000 19:49:30 +0000 (19:49 +0000)]
Add types and prototypes.

Submitted by: msmith

23 years agoForgot to commit this last night.
imp [Mon, 16 Oct 2000 19:43:44 +0000 (19:43 +0000)]
Forgot to commit this last night.

Define interrupt routing method.

Submitted by: msmith

23 years agoAdd an optional start/stop argument to pccard_ether, and migrate
joe [Mon, 16 Oct 2000 19:11:11 +0000 (19:11 +0000)]
Add an optional start/stop argument to pccard_ether, and migrate
the 'ifconfig ifN delete' into pccard_ether under the 'stop'
function.

In addition start dhclient with a pid file based on the interface
name, e.g. /var/run/dhclient.ep0.pid, and kill the correct dhclient
(or dhcpc) process when the card is removed.

23 years agoAvoid passing the optional ifconfig arguments to dhcpc as well as ifconfig.
joe [Mon, 16 Oct 2000 19:03:48 +0000 (19:03 +0000)]
Avoid passing the optional ifconfig arguments to dhcpc as well as ifconfig.
This bug has been here since revision 1.8.

23 years agoRemove an errant splimp() that I missed when I went through this driver
wpaul [Mon, 16 Oct 2000 18:51:32 +0000 (18:51 +0000)]
Remove an errant splimp() that I missed when I went through this driver
the first time.

23 years agoUndo previous commit. <machine/clock.h> is needed for the prototype
n_hibma [Mon, 16 Oct 2000 18:50:00 +0000 (18:50 +0000)]
Undo previous commit. <machine/clock.h> is needed for the prototype
for DELAY().

23 years agoSync with NetBSD:
n_hibma [Mon, 16 Oct 2000 18:13:59 +0000 (18:13 +0000)]
Sync with NetBSD:

K&R -> ANSI

Bugfix: 'Keep the bit position even when the report descriptor says POP.'

Add hid_use_report_desc, hid_parse_usage_page, hid_parse_usage_in_page.

Changed iface for hid_report_size.

23 years agoSync with NetBSD:
n_hibma [Mon, 16 Oct 2000 17:45:27 +0000 (17:45 +0000)]
Sync with NetBSD:

Add more Usages.

23 years agoSync with NetBSD:
n_hibma [Mon, 16 Oct 2000 17:40:59 +0000 (17:40 +0000)]
Sync with NetBSD:

Add seatbelts.
Be more lenient towards restarting the ports: If a port manages to connect
successfully reset the restartcount.

23 years agoMove __uint16_swap_uint32, __uint8_swap_uint32 and
brian [Mon, 16 Oct 2000 17:06:48 +0000 (17:06 +0000)]
Move __uint16_swap_uint32, __uint8_swap_uint32 and
__uint8_swap_uint16 inside __GNUC__.

Reviewed by: bde

23 years agoIn pmap_remove_pv(), only manipulate the page's list if the pv is
dfr [Mon, 16 Oct 2000 17:06:32 +0000 (17:06 +0000)]
In pmap_remove_pv(), only manipulate the page's list if the pv is
managed.

23 years agoDo a full exception_restore after an execve syscall to ensure that the
dfr [Mon, 16 Oct 2000 17:03:51 +0000 (17:03 +0000)]
Do a full exception_restore after an execve syscall to ensure that the
new program gets the right values for its arguments etc.

23 years agoClear the register stack frame before using loadrs to invalidate the
dfr [Mon, 16 Oct 2000 16:59:32 +0000 (16:59 +0000)]
Clear the register stack frame before using loadrs to invalidate the
stacked registers.

23 years agoClear ar.pfs for the child process in cpu_fork - switch_trampoline
dfr [Mon, 16 Oct 2000 16:55:59 +0000 (16:55 +0000)]
Clear ar.pfs for the child process in cpu_fork - switch_trampoline
doesn't want a stack frame.

23 years agoReturn zero from setjmp() and _setjmp() for now.
dfr [Mon, 16 Oct 2000 16:51:50 +0000 (16:51 +0000)]
Return zero from setjmp() and _setjmp() for now.

23 years agoFix a couple of dumb mistakes.
dfr [Mon, 16 Oct 2000 16:48:48 +0000 (16:48 +0000)]
Fix a couple of dumb mistakes.

23 years agoMerge from stable: Update of comments
n_hibma [Mon, 16 Oct 2000 16:38:04 +0000 (16:38 +0000)]
Merge from stable: Update of comments

23 years agoThe previous commit broke kernel builds on alpha (and probably ia64).
gallatin [Mon, 16 Oct 2000 15:38:11 +0000 (15:38 +0000)]
The previous commit broke kernel builds on alpha (and probably ia64).
#ifdef away the offending code until somebody with more newbus fu than
me can figure out where to put a default function that returns 255
without touching each alpha chipset driver..

23 years ago"Ok, my loader's now up to putting up a prompt. It probes disks partially
obrien [Mon, 16 Oct 2000 10:46:22 +0000 (10:46 +0000)]
"Ok, my loader's now up to putting up a prompt.  It probes disks partially
but can't boot from them yet."

Thanks to Stephane Potvin for the some of the code in this set.

Submitted by: Benno Rice <benno@jeamland.net>

23 years agoBacked out part of the revision 1.10.
ru [Mon, 16 Oct 2000 10:00:50 +0000 (10:00 +0000)]
Backed out part of the revision 1.10.
BLOCKSIZE could specify arbitrary block sizes, not exactly 1K blocks.