Matthew Dillon [Wed, 10 Mar 2004 06:35:22 +0000 (06:35 +0000)]
Explain some of the instruction details in more depth.
Submitted-by: Rahul Siddharthan <rsidd@online.fr>
Additional-edits-by: Matthew Dillon
Matthew Dillon [Wed, 10 Mar 2004 02:07:52 +0000 (02:07 +0000)]
Allow the nominal NFS io block size to be set with a sysctl vfs.nfs.nfs_io_size
and default it to the largest possible block size (32K), regardless of the
network transfer size settings. nfs_iosize() is no longer based on the
network transfer size but is matched against the maximum data block size
for the protocol and transport (8K for NFSv2, 16K for NFSv3/UDP, 32K for
NFSv3/TCP). Adjust statfs() reporting to suit.
This should improve performance over high bandwidth connections, primarily
by causing the client to use larger buffer cache buffers (16K or 32K instead
of 8K prior to this commit), and also improving read-ahead (which goes by
blocks). In particular, while the largest network transfer size over UDP
is 16K, the largest transfer size of TCP is 32K, so TCP ought to reap the
largest reward w/ this commit.
Delay retrieval of mountpoint attributes until the mountpoint is actually
accessed or a df() occurs.
Submitted-by: Hiten Pandya <hmp@backplane.com>, and Matthew Dillon
Joerg Sonnenberger [Tue, 9 Mar 2004 21:39:59 +0000 (21:39 +0000)]
Remove a debug printf added with the last commit.
Joerg Sonnenberger [Tue, 9 Mar 2004 21:38:05 +0000 (21:38 +0000)]
While converting ATA to use MPIPE allocations, ata_dmaalloc was changed
to allocate the DMA buffer directly without checking if it already is.
Since atapi_queue_cmd and ad_start doesn't check this and the DMA buffer
is only freed from ad_free, ad_detach and atapi_detach, the MPIPE was
exhausted very soon. The calls the ata_dmafree has been removed in revision
1.9.
This adds an additional check wether a DMA buffer is already allocated and
reuses that instead.
Jeffrey Hsu [Tue, 9 Mar 2004 21:21:54 +0000 (21:21 +0000)]
Clarify strange ipfw byte ordering convention.
Consulted with: Luigi Rizzo <rizzo@icir.org>
Matthew Dillon [Tue, 9 Mar 2004 18:05:27 +0000 (18:05 +0000)]
ncpus2 must be initialized to 1 in the UP case. ncpus2_mask and ncpus2_shift
remain 0 in the UP case (cpu selection mask for one cpu always selects cpu #0).
This fixes a bug introduced by recent network commits, correcting a boot-time
panic on UP systems in in_pcblookup_hash().
Hiten Pandya [Tue, 9 Mar 2004 15:00:06 +0000 (15:00 +0000)]
Adjust IPFW to use M_WAITOK instead of M_NOWAIT. The M_NOWAIT flag on
DragonFly is very explicit about its behavior, which can result in a
panic.
Only spl protected calls are using M_NOWAIT.
Submitted by: Craig Dooley <craig@xlnx-x.net>
Hiten Pandya [Tue, 9 Mar 2004 14:57:18 +0000 (14:57 +0000)]
Fix violating usage of M_DONTWAIT in calls to malloc() by replacing
them with M_NOWAIT.
Submitted by: Craig Dooley <craig@xlnx-x.net>
Matthew Dillon [Tue, 9 Mar 2004 02:52:10 +0000 (02:52 +0000)]
Addendum comment to last commit. Jeffrey Hsu reminded me that kernel writes
are not subject to socket buffer limitations, so the Linux bug referenced
in the last commit should not occur even without this commit. However,
the larger socket buffers should have a positive effect on our ability to
stream received data, especially over GigE links.
Matthew Dillon [Mon, 8 Mar 2004 23:52:53 +0000 (23:52 +0000)]
Increase the default socket buffer for NFS to deal with linux bugs and to
improve performance. The default nfs socket buffer is now 65535 bytes,
settable with a sysctl (vfs.nfs.soreserve).
It is my belief that when large data block sizes (32K) are negotiated,
the larger socket buff should improve read-ahead performance and reduce
nfs socket buffer lock contention that occurs with multiple nfsd's. I
was able to do some testing over GigE and it did seem to help, but problems
with one of the machines made the tests less then reliable.
Credits: Richard Sharpe originally encountered issues with linux NFS
clients that were traced to linux doing a bad job in its delayed-ack code.
David Rhodus created an initial patch which I used as a partial basis for
this commit (circa October 2003).
Matthew Dillon [Mon, 8 Mar 2004 22:41:05 +0000 (22:41 +0000)]
Add -B to display buffer limits instead of current buffer usage.
Jeffrey Hsu [Mon, 8 Mar 2004 19:44:32 +0000 (19:44 +0000)]
Partition the TCP connection table.
Joerg Sonnenberger [Mon, 8 Mar 2004 15:13:56 +0000 (15:13 +0000)]
Handle failure in atapi_queue_cmd correctly
Joerg Sonnenberger [Mon, 8 Mar 2004 13:31:57 +0000 (13:31 +0000)]
Fix panic in acd_report_key when ai->format==DVD_INVALIDATE_AGID.
Hiten Pandya [Mon, 8 Mar 2004 12:46:12 +0000 (12:46 +0000)]
Document that kldload(2) can also return EEXIST.
Hiten Pandya [Mon, 8 Mar 2004 07:55:57 +0000 (07:55 +0000)]
Merge from FreeBSD:
Fix handling of negative hexadecimal integers.
(originally from NetBSD)
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Hiten Pandya [Mon, 8 Mar 2004 07:50:03 +0000 (07:50 +0000)]
Merge from FreeBSD:
Add more partition types.
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Hiten Pandya [Mon, 8 Mar 2004 07:47:25 +0000 (07:47 +0000)]
Merge from FreeBSD:
Clarify a sentence.
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Jeffrey Hsu [Mon, 8 Mar 2004 07:43:44 +0000 (07:43 +0000)]
To comply with the spec, do not copy the TOS from the outer IP
header to the inner IP header of the PIM Register if this is a PIM
Null-Register message.
Submitted by: Pavlin Radoslavov <pavlin@icir.org>
Jeffrey Hsu [Mon, 8 Mar 2004 07:38:20 +0000 (07:38 +0000)]
Include <sys/types.h> for autoconf/automake detection.
Submitted by: Pavlin Radoslavov <pavlin@icir.org>
Matthew Dillon [Mon, 8 Mar 2004 06:13:17 +0000 (06:13 +0000)]
The "Don't forget to do a ``make depend''" warning no longer serves any
useful purpose, remove it.
From discussions with: Chris Pressey <cpressey@catseye.mine.nu>
Matthew Dillon [Mon, 8 Mar 2004 03:28:02 +0000 (03:28 +0000)]
Config cleanup part 3/3: Remove the ns() and twisty eq() macros and replace
their use in the codebase with their contents (strdup() and !strcmp()).
Submitted-by: Chris Pressey <cpressey@catseye.mine.nu>
Matthew Dillon [Mon, 8 Mar 2004 03:24:27 +0000 (03:24 +0000)]
Config cleanup part 2/3: Remove old style C cruft.
Submitted-by: Chris Pressey <cpressey@catseye.mine.nu>
Matthew Dillon [Mon, 8 Mar 2004 03:22:46 +0000 (03:22 +0000)]
Config cleanup part 1/3: Remove old style C cruft and cleanup some variable
names.
Submitted-by: Chris Pressey <cpressey@catseye.mine.nu>
Matthew Dillon [Mon, 8 Mar 2004 03:05:27 +0000 (03:05 +0000)]
ANSIfy the tsleep() and sched_setup() procedure definitions.
Submitted-by: Peter Kadau <peter.kadau@tuebingen.mpg.de>
Matthew Dillon [Mon, 8 Mar 2004 03:03:54 +0000 (03:03 +0000)]
libcaps now compiles ipiq and token in userland, make those files compile
properly in userland.
Jeffrey Hsu [Mon, 8 Mar 2004 00:44:36 +0000 (00:44 +0000)]
Print out Early Retransmit statistics.
Jeffrey Hsu [Mon, 8 Mar 2004 00:39:00 +0000 (00:39 +0000)]
Implement Early Retransmit.
Jeffrey Hsu [Mon, 8 Mar 2004 00:36:30 +0000 (00:36 +0000)]
Cosmetic changes.
Matthew Dillon [Sun, 7 Mar 2004 23:36:45 +0000 (23:36 +0000)]
Initial CAPS IPC structural encoding and decoding support. Note that the
kernel is uninvolved (other then in supplying the IPC mechanism). After
careful consideration I decided that XML was just too much overkill but that
a human-readable format is still desireable. The encoding format supports
integers, strings, opaque data, arrays, and is extensible and structural.
An example encoding of the struct passwd record for user 'nobody':
Spasswd{F1D"nobody",F2D"%2a",F3Dfffe,F4Dfffe,F5D0,F6D"",F7D"Unprivileged%20user",F8D"/nonexistent",F9D"/sbin/nologin",FaD0}
Class elements include (S)tructure, (F)ield, (D)ata, (A)rray. Data types are
free-form numeric (stored as hex), with the source responsible for encoding
negative numbers with a '-', and quoted opaque data (e.g. strings). Any
class element may recurse using {}. Structures and Arrays always recurse.
Unknown recursions are ignored by the decoder (for future backwards
compatibility). Whitespace is allowed but most non-alpha-numeric characters
must be %-hex-escaped.
The decoder understands simple integer types, arrays, buffers, embedded
sub-structures, and string pointer structural elements. It does not (yet)
understand multi-dimensional arrays, sparse element notation, or ** pointers.
This is a nice, compact encoding format that can be used for both packetized
messages and persistently-connected streams. The initial commit only supports
full-messages, however.
Joerg Sonnenberger [Sun, 7 Mar 2004 21:35:27 +0000 (21:35 +0000)]
From FreeBSD:
When we get a packet error, move on, don't go into an infinite loop
looking at it.
fixes at least one cause of "hanging" due to this driver.
Jeffrey Hsu [Sun, 7 Mar 2004 18:38:38 +0000 (18:38 +0000)]
Send UDP packets out without a temporary connect.
Eirik Nygaard [Sun, 7 Mar 2004 12:48:34 +0000 (12:48 +0000)]
Remove perl from the i386 kernel build.
Eirik Nygaard [Sun, 7 Mar 2004 12:09:04 +0000 (12:09 +0000)]
Move the ASSERT_VOP_LOCKED and ASSERT_VOP_UNLOCKED macros into its own
functions.
Idea taken from: FreeBSD
Matthew Dillon [Sun, 7 Mar 2004 01:13:43 +0000 (01:13 +0000)]
get rid of thr{1,2,3}, which are obsolete at the moment. Keep the
caps system call tests.
Matthew Dillon [Sat, 6 Mar 2004 23:19:01 +0000 (23:19 +0000)]
The sys/xxx2.h files are supposed to be included after all the normal
sys/*.h files, not before.
Reported-by: David Rhodus <drhodus@machdep.com>
Matthew Dillon [Sat, 6 Mar 2004 22:15:00 +0000 (22:15 +0000)]
Adjust the caps client/server test code to match new CAPS features. The
code also tests fork() handling and disconnect detection.
Matthew Dillon [Sat, 6 Mar 2004 22:14:16 +0000 (22:14 +0000)]
Additional CAPS IPC work. Add additional system calls to allow a CAPS
server to set a generation number and a CAPS client to query it, which can
be used for any purpose but which is intended to allow a server to tell its
clients to invalidate their caches.
Add missing fork-handling code. CAPS links are only good on a thread-by-thread
basis. When a process forks/rforks/clones any active CAPS links will be
created as dummy entries in the forked process, causing CAPS syscalls to
return ENOTCONN. This allows code based on CAPS to detect when it has been
forked so it can re-connect to the service.
Make a slight change to the API. caps_sys_put() now returns an immediate
ENOTCONN if it forked. Note that userland CAPS code must still deal with
the case where a message has been sent and the connection is lost before the
reply is returned. The kernel automatically replies unreplied messages
with 0-length data in these cases.
Add additional flags to the API, including one that allows a client to
block when connecting to a non-existant service.
Matthew Dillon [Sat, 6 Mar 2004 19:48:22 +0000 (19:48 +0000)]
Bring libcaps in line with recent LWKT changes.
Matthew Dillon [Sat, 6 Mar 2004 19:40:32 +0000 (19:40 +0000)]
Simplify LWKT message initialization semantics to reduce API confusion.
Cleanup netisr messaging to provide more uniform error handling and to use
lwkt_replymsg() unconditionally for both async/auto-free and sync messages
as the abstraction intended. This also fixes a reply/free race.
Jeffrey Hsu [Sat, 6 Mar 2004 07:30:44 +0000 (07:30 +0000)]
Remove unused second argument to ip_stripoptions().
Jeffrey Hsu [Sat, 6 Mar 2004 05:20:31 +0000 (05:20 +0000)]
Eliminate the use of curproc in route_output() by passing down the process id
it needs.
Jeffrey Hsu [Sat, 6 Mar 2004 05:00:41 +0000 (05:00 +0000)]
Change the "struct inpcbhead *listhead" field in "struct inpcbinfo"
to "struct inpcbhead listhead" so we can have a separate list per
"struct inpcbinfo" when it becomes per-cpu.
Jeffrey Hsu [Sat, 6 Mar 2004 03:26:57 +0000 (03:26 +0000)]
Use 0 for integer value rather than NULL.
Jeffrey Hsu [Sat, 6 Mar 2004 03:20:02 +0000 (03:20 +0000)]
Correct double increment of the inp generation count.
Jeffrey Hsu [Sat, 6 Mar 2004 02:36:25 +0000 (02:36 +0000)]
Dispatch upper-half protocol request handling.
Jeffrey Hsu [Sat, 6 Mar 2004 01:59:52 +0000 (01:59 +0000)]
Dispatch upper-half protocol request handling.
Joerg Sonnenberger [Fri, 5 Mar 2004 21:38:44 +0000 (21:38 +0000)]
Propolice for GCC 3.3 based on:
http://www.trl.ibm.com/projects/security/ssp/gcc3_3/protector-3.3-7.tar.gz
protector.[ch] has been in integrated into cc3/cc_int and the protector.dif
has been splitted and integrated into cc3/cc_int/patches using the new
patch infrastructure. The old bsd-format.patch has been moved there as well.
The build order in gnu/ has been adjusted since lib/gcc3/libgcc depends on
"make depend" in usr.bin/cc3/cc_int for inclusion of the patched files.
The second stage loader has been adjusted to link when parts are compiled
using -fstack-protector.
Jeffrey Hsu [Fri, 5 Mar 2004 20:00:03 +0000 (20:00 +0000)]
Verify code assumption on number of processors with a kernel assertion.
Jeffrey Hsu [Fri, 5 Mar 2004 19:47:28 +0000 (19:47 +0000)]
Use power of 2 masking to make packet hash function fast.
Jeffrey Hsu [Fri, 5 Mar 2004 19:29:17 +0000 (19:29 +0000)]
Remember the next lowest power of 2 of "npus" in "ncpus2".
Jeffrey Hsu [Fri, 5 Mar 2004 19:17:25 +0000 (19:17 +0000)]
Propagate curproc removal changes to files compiled by LINT.
Jeffrey Hsu [Fri, 5 Mar 2004 16:57:16 +0000 (16:57 +0000)]
Once we distribute socket protocol processing requests to different
processors, we no longer have a process context to refer to, so
eliminate the use of curproc in soreserve() by passing the sockbuf
resource limit all the down from the system call code to sbreserve().
Eliminate the use of curproc in unp_attach() by passing down the
fields it needs from the proc structure. Define a pru_attach_info
structure to hold the information the attach usrreq function requires.
The thread argument to in_pcballoc() is unused, so we don't need
to pass a thread structure down to in_pcballoc().
Hiten Pandya [Fri, 5 Mar 2004 13:07:02 +0000 (13:07 +0000)]
Bring the BFE(4) manual page up-to-date with FreeBSD RELENG_4.
Submitted by: Christian Brueffer <chris@unixpages.org>
Matthew Dillon [Fri, 5 Mar 2004 08:03:31 +0000 (08:03 +0000)]
Correct bug introduced in last commit.
Submitted-by: Chris Pressey <cpressey@catseye.mine.nu>
Joerg Sonnenberger [Fri, 5 Mar 2004 01:06:50 +0000 (01:06 +0000)]
Add patch infrastructure for contrib/ and similiar directories.
This allows the inclusion of patches in SRCS in the following form:
this,is,the,source.c.patch
This will create ${.OBJDIR}/this/is/the/source.c by applying above patch
to ${CONTRIBDIR}/this/is/the/source.c. The patched files are deleted on
"make clean" and created by "make depend".
Hiten Pandya [Fri, 5 Mar 2004 00:10:31 +0000 (00:10 +0000)]
I just scored a few duh-points for myself. I committed an older version
of the patch, because the patch file name got me confused.
Correct revision 1.7 of this file by using a proper conditional block.
Eirik Nygaard [Thu, 4 Mar 2004 20:52:27 +0000 (20:52 +0000)]
this -> the fix in comment.
Remove extra space in comment while I am here.
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Eirik Nygaard [Thu, 4 Mar 2004 20:50:58 +0000 (20:50 +0000)]
Make the comment a bit clearer.
Eirik Nygaard [Thu, 4 Mar 2004 20:44:49 +0000 (20:44 +0000)]
style(9) cleanup:
o use NULL instead of 0 for pointers
o explicitly compare to 0 or NULL for non-booleans
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Eirik Nygaard [Thu, 4 Mar 2004 20:40:48 +0000 (20:40 +0000)]
style(9) cleanup:
o no space after casts
o no space, but yes parens, after sizeof and return
o spaces around binary operators
o other minor whitespace issues
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Eirik Nygaard [Thu, 4 Mar 2004 20:29:45 +0000 (20:29 +0000)]
style(9) cleanup:
o change K&R declarations to ANSI declarations
o remove ``register'' keywords
Submitted by: Chris Pressey <cpressey@catseye.mine.nu>
Jeffrey Hsu [Thu, 4 Mar 2004 10:29:24 +0000 (10:29 +0000)]
Introduce access methods for making protocol requests.
Jeffrey Hsu [Thu, 4 Mar 2004 06:13:05 +0000 (06:13 +0000)]
A UDP socket is still bound after it is disconnected, so we need to
reinsert it into the bind hash table.
Bug reported by: David Rhodus, Matt Dillon
Matthew Dillon [Thu, 4 Mar 2004 01:38:01 +0000 (01:38 +0000)]
Partitions>8: Increase the number of supported partitions from 8 to 16.
Decrease the number of supported slices from 32 to 16. Note that the
5.x boot2 code, which we adopted, was being installed just after the old
disklabel. This commit moves the boot code install to the next logical
sector (aka 4.x) in order to accomodate the larger label.
Fix newfs to not hardcode 'h' as the last partition.
Also modify 'disklabel' to not complain about preexisting garbage past
partition #8, and to detect and refuse to overwrite the old bootcode
with the new larger label until after you have installed new boot code.
Jeffrey Hsu [Thu, 4 Mar 2004 01:02:06 +0000 (01:02 +0000)]
Split out wildcarded sockets from the connection hash table.
Matthew Dillon [Thu, 4 Mar 2004 00:06:58 +0000 (00:06 +0000)]
Partitions>8: Leave a whole 512 bytes for the disklabel and squeeze code
to fit. It was a really bad idea for FBsd5 to have changed this.
Matthew Dillon [Thu, 4 Mar 2004 00:03:23 +0000 (00:03 +0000)]
Partitions>8: Do not hardwire partition limit at 'h'.
Hiten Pandya [Wed, 3 Mar 2004 22:51:28 +0000 (22:51 +0000)]
Do not print a warning about PIM sysctl node (net.inet.pim.stats)
if sysctlbyname(2) sets *errno* to ENOENT, because that means we
do not have PIM in the kernel.
Reviewed by: hsu
Hiten Pandya [Wed, 3 Mar 2004 05:16:21 +0000 (05:16 +0000)]
Include thread.h if _KERNEL_STRUCTURES is defined.
This is necessary for fixing multiple FreeBSD sysutil/ port builds.
Hiten Pandya [Wed, 3 Mar 2004 03:55:24 +0000 (03:55 +0000)]
Bring the cue(4) and miibus(4) manual page in line with
FreeBSD 4-STABLE.
Submitted by: Christian Brueffer <chris@unixpages.org>
David Rhodus [Tue, 2 Mar 2004 21:03:46 +0000 (21:03 +0000)]
* Add in support for the Silicon Image SATA controller.
David Rhodus [Tue, 2 Mar 2004 20:59:46 +0000 (20:59 +0000)]
* Correct end value calculation. This should fix the numerous problems
with accessing devices on the ATA controller channel.
David Rhodus [Tue, 2 Mar 2004 20:55:10 +0000 (20:55 +0000)]
* Remove instances of PRIBIO because DragonFly no longer
uses sleep priorities.
Robert Garrett [Tue, 2 Mar 2004 20:41:13 +0000 (20:41 +0000)]
Patch forr FreeBSD-SA-04:04.tcp
limits out of sequence reassembly queue size, to make sure we don't run
out of mbufs, resulting in a DOS attack.
This is the same as tcp47.patch
checked by Robert Garrett & Joerg Sonnenberger
Joerg Sonnenberger [Tue, 2 Mar 2004 16:04:20 +0000 (16:04 +0000)]
The slab allocator has been in for while now. Change the printf for invalid
malloc flags into a panic to allow easier tracing of any remaining incorrect
uses.
Joerg Sonnenberger [Tue, 2 Mar 2004 15:03:00 +0000 (15:03 +0000)]
We have /etc/rc.subr, don't add the dependency
Hiten Pandya [Tue, 2 Mar 2004 06:06:44 +0000 (06:06 +0000)]
Correct a filename typo, there is no such thing as machine/pcpu.h,
it is machine/globaldata.h.
Eirik Nygaard [Mon, 1 Mar 2004 20:45:57 +0000 (20:45 +0000)]
Use MALLOC_DEFINE.
Obtained from: FreeBSD
Joerg Sonnenberger [Mon, 1 Mar 2004 17:27:17 +0000 (17:27 +0000)]
Use M_WAITOK instead of M_WAIT to get memory.
Joerg Sonnenberger [Mon, 1 Mar 2004 16:57:44 +0000 (16:57 +0000)]
Make subr_bus.c more consistent with regard to style(9) and itself.
- adjust the 4 space indentation in the oldest parts
- use return(value) form
- move returns into the default case of certain switch statements
- make some pointer checks explicit against NULL
- reorder device_probe_and_attach to simplify the if's
- remove unnecessary return; at the end of functions
Joerg Sonnenberger [Mon, 1 Mar 2004 11:50:59 +0000 (11:50 +0000)]
Adjust indentation, use uint32_t and line up comments.
Matthew Dillon [Mon, 1 Mar 2004 06:33:24 +0000 (06:33 +0000)]
Newtoken commit. Change the token implementation as follows: (1) Obtaining
a token no longer enters a critical section. (2) tokens can be held through
schedular switches and blocking conditions and are effectively released and
reacquired on resume. Thus tokens serialize access only while the thread
is actually running. Serialization is not broken by preemptive interrupts.
That is, interrupt threads which preempt do no release the preempted thread's
tokens. (3) Unlike spl's, tokens will interlock w/ interrupt threads on
the same or on a different cpu.
The vnode interlock code has been rewritten and the API has changed. The
mountlist vnode scanning code has been consolidated and all known races have
been fixed. The vnode interlock is now a pool token.
The code that frees unreferenced vnodes whos last VM page has been freed has
been moved out of the low level vm_page_free() code and moved to the
periodic filesystem sycer code in vfs_msycn().
The SMP startup code and the IPI code has been cleaned up considerably.
Certain early token interactions on AP cpus have been moved to the BSP.
The LWKT rwlock API has been cleaned up and turned on.
Major testing by: David Rhodus
Joerg Sonnenberger [Sun, 29 Feb 2004 17:18:32 +0000 (17:18 +0000)]
Merge FreeBSD rev. 1.8:
Simplify the KOBJOPLOOKUP macro for the non-debug case so that gcc's
heuristics do not overestimate the code size quite so much.
Joerg Sonnenberger [Sun, 29 Feb 2004 16:55:28 +0000 (16:55 +0000)]
De-K&R-ify function prototyps and remove register keyword.
Submitted by Chris Pressey <cpressey@catseye.mine.nu>
Joerg Sonnenberger [Fri, 27 Feb 2004 11:56:12 +0000 (11:56 +0000)]
Add bfe(4) support from FreeBSD.
Initial code submitted by Peter Avalos <pavalos@theshell.com>.
Changes to the FreeBSD version:
- make the code consistent w.r.t. style(9)
- remove some unused entries from bfe_softc
- use the PCI IDs from pcidevs.h
- use BUS_DMA_WAITOK since the allocation is done in bfe_attach before
the interrupt is registered and sleeping is therefore safe
- fix some warnings in the code about signed/unsigned comparisions
Joerg Sonnenberger [Fri, 27 Feb 2004 11:27:39 +0000 (11:27 +0000)]
There are historically two families of fixed size integers, u_intX_t and
uintX_t. Since the former is BSD specific, while the later is defined for
POSIX anyway, prefer the second form.
Jeffrey Hsu [Fri, 27 Feb 2004 09:40:36 +0000 (09:40 +0000)]
Cosmetic changes.
Joerg Sonnenberger [Thu, 26 Feb 2004 14:25:29 +0000 (14:25 +0000)]
Use ifp->xname instead of if_name(ifp)
Joerg Sonnenberger [Thu, 26 Feb 2004 14:07:21 +0000 (14:07 +0000)]
Add missing return from last commit
Joerg Sonnenberger [Thu, 26 Feb 2004 14:06:13 +0000 (14:06 +0000)]
Remove unused and undocumented strhash files.
Suggested by Chris Pressey <cpressey@catseye.mine.nu>
Joerg Sonnenberger [Thu, 26 Feb 2004 13:58:26 +0000 (13:58 +0000)]
Remove parameter names, adjust white spaces in prototyps and remove
extern.
Based on patch submitted by Chris Pressey <cpressey@catseye.mine.nu>
Joerg Sonnenberger [Thu, 26 Feb 2004 13:06:19 +0000 (13:06 +0000)]
Remove the entry for pccard and allow src/include/Makefile to properly
create it as the symlink to bus/pccard.
Noticed by YONETANI Tomokazu <qhwt+dragonfly-submit@les.ath.cx>
Eirik Nygaard [Wed, 25 Feb 2004 20:44:30 +0000 (20:44 +0000)]
Update FILES section
Eirik Nygaard [Wed, 25 Feb 2004 18:58:46 +0000 (18:58 +0000)]
Remove unused settings
Joerg Sonnenberger [Wed, 25 Feb 2004 17:38:51 +0000 (17:38 +0000)]
Remove parameter names.
Submitted by Chris Pressey <cpressey@catseye.mine.nu>
Joerg Sonnenberger [Wed, 25 Feb 2004 17:35:29 +0000 (17:35 +0000)]
Don't use parameter names for kernel prototyps
Joerg Sonnenberger [Wed, 25 Feb 2004 16:30:21 +0000 (16:30 +0000)]
Cleanup emujoy_pci_probe
Joerg Sonnenberger [Wed, 25 Feb 2004 16:16:49 +0000 (16:16 +0000)]
Adjust mixer script to depend on mixer_enable="YES" and default to NO.
Joerg Sonnenberger [Wed, 25 Feb 2004 15:46:54 +0000 (15:46 +0000)]
Fix use after free / double free bugs. Return an error in error conditions.
Submitted by YONETANI Tomokazu <qhwt+dragonfly-submit@les.ath.cx>