dragonfly.git
14 years agoMerge from vendor branch CVS:
Simon Schubert [Mon, 13 Jun 2005 22:35:54 +0000 (22:35 +0000)]
Merge from vendor branch CVS:
update cvs to 1.12.12

14 years agoupdate cvs to 1.12.12
Simon Schubert [Mon, 13 Jun 2005 22:35:54 +0000 (22:35 +0000)]
update cvs to 1.12.12

14 years agoupdate cvs to 1.12.12
Simon Schubert [Mon, 13 Jun 2005 22:35:54 +0000 (22:35 +0000)]
update cvs to 1.12.12

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 22:27:52 +0000 (22:27 +0000)]
Convert to critical sections.

14 years agoFix mismatched crit_*() pair.
Matthew Dillon [Mon, 13 Jun 2005 22:21:39 +0000 (22:21 +0000)]
Fix mismatched crit_*() pair.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 22:18:44 +0000 (22:18 +0000)]
Convert to critical sections.

14 years agoUse critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 22:14:17 +0000 (22:14 +0000)]
Use critical sections.

14 years ago- convert to critical sections
Joerg Sonnenberger [Mon, 13 Jun 2005 22:10:56 +0000 (22:10 +0000)]
- convert to critical sections
- hook up interrupt last
- use M_WAITOK for contigmalloc
- use dc_detach as common error path

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 21:53:24 +0000 (21:53 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 21:44:40 +0000 (21:44 +0000)]
Convert to critical sections.

14 years agoFix build.
Joerg Sonnenberger [Mon, 13 Jun 2005 21:39:14 +0000 (21:39 +0000)]
Fix build.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 21:38:12 +0000 (21:38 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections. Use bge_detach as common error path
Joerg Sonnenberger [Mon, 13 Jun 2005 21:24:03 +0000 (21:24 +0000)]
Convert to critical sections. Use bge_detach as common error path
in bge_attach and fix some bugs on the way. Use M_WAITOK now
that interrupts are no longer an issue there. Allow this by hooking
up the interrupt last.

14 years agoConvert to critical sections. Add a missing ether_ifdetach when
Joerg Sonnenberger [Mon, 13 Jun 2005 21:07:55 +0000 (21:07 +0000)]
Convert to critical sections. Add a missing ether_ifdetach when
bus_setup_intr failed.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 20:51:17 +0000 (20:51 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 20:36:45 +0000 (20:36 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 20:30:45 +0000 (20:30 +0000)]
Convert to critical sections.

14 years agoRemove compatiblity code.
Joerg Sonnenberger [Mon, 13 Jun 2005 20:25:56 +0000 (20:25 +0000)]
Remove compatiblity code.

14 years agoNo interrupts, no M_NOWAIT. Use M_WAITOK instead.
Joerg Sonnenberger [Mon, 13 Jun 2005 20:09:24 +0000 (20:09 +0000)]
No interrupts, no M_NOWAIT. Use M_WAITOK instead.

14 years ago- convert to critical sections
Joerg Sonnenberger [Mon, 13 Jun 2005 20:07:38 +0000 (20:07 +0000)]
- convert to critical sections
- setup interrupt last, avoid having to deal with ISRs
- use nge_detach as common error handling path

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 19:21:19 +0000 (19:21 +0000)]
Convert to critical sections.

14 years agoMissed one splnet.
Joerg Sonnenberger [Mon, 13 Jun 2005 19:19:19 +0000 (19:19 +0000)]
Missed one splnet.

14 years agoUpdate pcic(4) and add more NEWCARD manpages.
Sascha Wildner [Mon, 13 Jun 2005 19:17:29 +0000 (19:17 +0000)]
Update pcic(4) and add more NEWCARD manpages.

14 years agoConvert to critical sections. Remove compat junk. Hook up interrupt
Joerg Sonnenberger [Mon, 13 Jun 2005 19:05:19 +0000 (19:05 +0000)]
Convert to critical sections. Remove compat junk. Hook up interrupt
last. This fixes a missing interrupt teardown under error conditions.

14 years ago- convert to critical sections
Joerg Sonnenberger [Mon, 13 Jun 2005 18:43:58 +0000 (18:43 +0000)]
- convert to critical sections
- setup interrupt last in pcn_attach
- use pcn_detach for common error handling

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 18:13:20 +0000 (18:13 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections. Fix an early return without restoring
Joerg Sonnenberger [Mon, 13 Jun 2005 18:11:52 +0000 (18:11 +0000)]
Convert to critical sections. Fix an early return without restoring
interrupt protection in the progress of the conversion.

14 years agoConvert from splhigh to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 14:00:29 +0000 (14:00 +0000)]
Convert from splhigh to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 13:51:24 +0000 (13:51 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 13:47:50 +0000 (13:47 +0000)]
Convert to critical sections.

14 years agoConvert to critical sections.
Joerg Sonnenberger [Mon, 13 Jun 2005 13:31:04 +0000 (13:31 +0000)]
Convert to critical sections.

14 years agoOnly delete the miibus if it was attached first.
Joerg Sonnenberger [Mon, 13 Jun 2005 10:20:49 +0000 (10:20 +0000)]
Only delete the miibus if it was attached first.

Noticed-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoRemove PC98 support.
Sascha Wildner [Sun, 12 Jun 2005 20:55:14 +0000 (20:55 +0000)]
Remove PC98 support.

14 years agoRemove obsolete comments.
Sascha Wildner [Sun, 12 Jun 2005 19:24:10 +0000 (19:24 +0000)]
Remove obsolete comments.

14 years ago- setup interrupt last
Joerg Sonnenberger [Sun, 12 Jun 2005 17:44:29 +0000 (17:44 +0000)]
- setup interrupt last
- convert to critical sections
- use common error path in sbsh_attach by calling sbsh_detach

14 years ago- convert to critical sections
Joerg Sonnenberger [Sun, 12 Jun 2005 17:29:49 +0000 (17:29 +0000)]
- convert to critical sections
- setup interrupt last, no need to protect attach against races
- use M_WAITOK for contigmalloc
- use sf_detach for common error handling

14 years agoSwitch to critical sections, fix some possible minor nits with ISR
Joerg Sonnenberger [Sun, 12 Jun 2005 17:19:07 +0000 (17:19 +0000)]
Switch to critical sections, fix some possible minor nits with ISR
protection.

14 years agoStop abusing splbio simply because others do the same. Use critical
Joerg Sonnenberger [Sun, 12 Jun 2005 17:03:47 +0000 (17:03 +0000)]
Stop abusing splbio simply because others do the same. Use critical
sections.

14 years agoConvert splimp to critical sections.
Joerg Sonnenberger [Sun, 12 Jun 2005 16:50:20 +0000 (16:50 +0000)]
Convert splimp to critical sections.

14 years agoRemove compat junk.
Joerg Sonnenberger [Sun, 12 Jun 2005 16:32:52 +0000 (16:32 +0000)]
Remove compat junk.

14 years agoConvert from spl* to critical sections.
Joerg Sonnenberger [Sun, 12 Jun 2005 16:31:50 +0000 (16:31 +0000)]
Convert from spl* to critical sections.

14 years agoTypo.
Joerg Sonnenberger [Sun, 12 Jun 2005 15:55:59 +0000 (15:55 +0000)]
Typo.

Noticed-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoDeprecate MCLGET() in favor of m_getcl() or m_getl() in order to
Jeffrey Hsu [Sun, 12 Jun 2005 15:39:33 +0000 (15:39 +0000)]
Deprecate MCLGET() in favor of m_getcl() or m_getl() in order to
take advantage of cluster caching.

14 years agoUse m_gethdr() instead of m_get() to get a mbuf header.
Jeffrey Hsu [Sun, 12 Jun 2005 15:37:19 +0000 (15:37 +0000)]
Use m_gethdr() instead of m_get() to get a mbuf header.

14 years agoRemove NO_PKGTOOLS wrapper, it existed in top-level Makefile already.
Jeroen Ruigrok/asmodai [Sat, 11 Jun 2005 22:37:34 +0000 (22:37 +0000)]
Remove NO_PKGTOOLS wrapper, it existed in top-level Makefile already.

Reminded by: joerg

14 years agoAdd NO_PKGTOOLS to disable building of the pkg_* tools during world.
Jeroen Ruigrok/asmodai [Sat, 11 Jun 2005 21:09:51 +0000 (21:09 +0000)]
Add NO_PKGTOOLS to disable building of the pkg_* tools during world.
This is useful for people using pkgsrc.

14 years agoThe BLIST API is just as usable in userland as it is in the kernel; and
Hiten Pandya [Sat, 11 Jun 2005 16:24:36 +0000 (16:24 +0000)]
The BLIST API is just as usable in userland as it is in the kernel; and
the code already contains a userland test program.

Remove conditionals preventing BLISTs from being used in the userland.

14 years agoDispose of support for IBM's Micro Channel architecture (MCA).
Sascha Wildner [Sat, 11 Jun 2005 09:03:49 +0000 (09:03 +0000)]
Dispose of support for IBM's Micro Channel architecture (MCA).

14 years agoReally use M_WAITOK.
Joerg Sonnenberger [Sat, 11 Jun 2005 08:58:48 +0000 (08:58 +0000)]
Really use M_WAITOK.

14 years agoDon't bzero softc. Setup interrupt last to get tl_attach ISR race free.
Joerg Sonnenberger [Sat, 11 Jun 2005 08:57:26 +0000 (08:57 +0000)]
Don't bzero softc. Setup interrupt last to get tl_attach ISR race free.
Use M_WAITOK for contigmalloc. Use tl_detach for the cleanup path.

14 years agoWe know that tl_probe is run first, so turn the check for a bad device
Joerg Sonnenberger [Sat, 11 Jun 2005 08:50:21 +0000 (08:50 +0000)]
We know that tl_probe is run first, so turn the check for a bad device
into a KKASSERT.

14 years agoDeprecate MCLGET() in favor of m_getcl() or m_getl() in order to
Jeffrey Hsu [Sat, 11 Jun 2005 04:26:53 +0000 (04:26 +0000)]
Deprecate MCLGET() in favor of m_getcl() or m_getl() in order to
take advantage of cluster caching.

14 years agoMove sys/buf2.h and sys/thread2.h into the #ifdef _KERNEL section.
Matthew Dillon [Sat, 11 Jun 2005 00:35:12 +0000 (00:35 +0000)]
Move sys/buf2.h and sys/thread2.h into the #ifdef _KERNEL section.

14 years agoAdd missing #include <thread2.h> to support the critical section calls.
Matthew Dillon [Sat, 11 Jun 2005 00:29:10 +0000 (00:29 +0000)]
Add missing #include <thread2.h> to support the critical section calls.

14 years agospl->critical section conversion.
Matthew Dillon [Sat, 11 Jun 2005 00:27:09 +0000 (00:27 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversions.
Matthew Dillon [Sat, 11 Jun 2005 00:26:51 +0000 (00:26 +0000)]
spl->critical section conversions.

14 years agospl->critical section conversion.
Matthew Dillon [Sat, 11 Jun 2005 00:05:46 +0000 (00:05 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversion, plus remove some macros which are now
Matthew Dillon [Fri, 10 Jun 2005 23:59:33 +0000 (23:59 +0000)]
spl->critical section conversion, plus remove some macros which are now
unused due the removal of spls.

14 years agospl->critical section conversion.
Matthew Dillon [Fri, 10 Jun 2005 23:29:32 +0000 (23:29 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversion
Matthew Dillon [Fri, 10 Jun 2005 23:25:09 +0000 (23:25 +0000)]
spl->critical section conversion

14 years agospl->critical section conversion.
Matthew Dillon [Fri, 10 Jun 2005 23:11:55 +0000 (23:11 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversion.
Matthew Dillon [Fri, 10 Jun 2005 23:07:02 +0000 (23:07 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversion.
Matthew Dillon [Fri, 10 Jun 2005 22:44:02 +0000 (22:44 +0000)]
spl->critical section conversion.

14 years agospl->critical section conversion.
Matthew Dillon [Fri, 10 Jun 2005 22:34:51 +0000 (22:34 +0000)]
spl->critical section conversion.

14 years agoReplace SPLs with critical sections in the opencrypto code.
Matthew Dillon [Fri, 10 Jun 2005 22:16:05 +0000 (22:16 +0000)]
Replace SPLs with critical sections in the opencrypto code.

14 years agoRemove DEC Alpha support.
Sascha Wildner [Fri, 10 Jun 2005 20:16:32 +0000 (20:16 +0000)]
Remove DEC Alpha support.

14 years agoAnother terrible hack to leave interrupts disabled until the USB bus
Matthew Dillon [Fri, 10 Jun 2005 18:33:04 +0000 (18:33 +0000)]
Another terrible hack to leave interrupts disabled until the USB bus
has completed initialization.

14 years agoAttempt to avoid a livelocked USB interrupt during boot by not enabling
Matthew Dillon [Fri, 10 Jun 2005 18:21:11 +0000 (18:21 +0000)]
Attempt to avoid a livelocked USB interrupt during boot by not enabling
the EHCI interrupt until after all the companion controllers have been
attached.

14 years agoRemove spl*() in dev/raid/{mlx,mly,pst,twa,twe} and replace them
Sascha Wildner [Fri, 10 Jun 2005 17:10:26 +0000 (17:10 +0000)]
Remove spl*() in dev/raid/{mlx,mly,pst,twa,twe} and replace them
with critical sections.

14 years agoRemove a useless assignment. Move callout_reset into critical section.
Joerg Sonnenberger [Fri, 10 Jun 2005 16:19:41 +0000 (16:19 +0000)]
Remove a useless assignment. Move callout_reset into critical section.

14 years agoUse epic_detach for error cleanup during attach. Make attach interrupt
Joerg Sonnenberger [Fri, 10 Jun 2005 16:16:51 +0000 (16:16 +0000)]
Use epic_detach for error cleanup during attach. Make attach interrupt
safe by hooking the interrupt in last.

14 years agoMove the callout_reset into the critical section.
Joerg Sonnenberger [Fri, 10 Jun 2005 16:10:42 +0000 (16:10 +0000)]
Move the callout_reset into the critical section.

14 years agoUse the PCI helper functions and nuke now unused macros.
Joerg Sonnenberger [Fri, 10 Jun 2005 16:05:34 +0000 (16:05 +0000)]
Use the PCI helper functions and nuke now unused macros.

Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoif_printf / device_printf
Joerg Sonnenberger [Fri, 10 Jun 2005 16:02:47 +0000 (16:02 +0000)]
if_printf / device_printf

Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoRemove spl*() in dev/raid/{dpt,ida,iir} and replace them
Sascha Wildner [Fri, 10 Jun 2005 15:46:31 +0000 (15:46 +0000)]
Remove spl*() in dev/raid/{dpt,ida,iir} and replace them
with critical sections.

14 years agoRemove unused variables.
Sascha Wildner [Fri, 10 Jun 2005 15:29:17 +0000 (15:29 +0000)]
Remove unused variables.

14 years agoReorder critical sections to be as short as possible by moving invariants out.
Joerg Sonnenberger [Fri, 10 Jun 2005 15:13:37 +0000 (15:13 +0000)]
Reorder critical sections to be as short as possible by moving invariants out.

14 years agoAdd /usr/pkg/etc/rc.d to the rcNG search list.
Joerg Sonnenberger [Fri, 10 Jun 2005 12:53:45 +0000 (12:53 +0000)]
Add /usr/pkg/etc/rc.d to the rcNG search list.

14 years agoFix ctype::is() by using the right array
Simon Schubert [Thu, 9 Jun 2005 23:38:00 +0000 (23:38 +0000)]
Fix ctype::is() by using the right array

14 years agoRemove spl*() in dev/raid/{aac,amr,asr,ciss} and replace them
Sascha Wildner [Thu, 9 Jun 2005 20:55:05 +0000 (20:55 +0000)]
Remove spl*() in dev/raid/{aac,amr,asr,ciss} and replace them
with critical sections.

14 years agoUnbreak LINT by including <sys/thread2.h> where proper and remove some
Sascha Wildner [Thu, 9 Jun 2005 20:47:37 +0000 (20:47 +0000)]
Unbreak LINT by including <sys/thread2.h> where proper and remove some
unused s variables.

14 years agoFix typo that turns out to be harmless by accident, as MT_HEADER and
Jeffrey Hsu [Thu, 9 Jun 2005 20:17:53 +0000 (20:17 +0000)]
Fix typo that turns out to be harmless by accident, as MT_HEADER and
M_PKTHDR have the same value, 2.

14 years agoConvert to critical sections. No need to protect the interupt from racing
Joerg Sonnenberger [Thu, 9 Jun 2005 20:04:44 +0000 (20:04 +0000)]
Convert to critical sections. No need to protect the interupt from racing
against itself.

14 years agoConvert to critical sections, move timer reset into the protection.
Joerg Sonnenberger [Thu, 9 Jun 2005 19:45:12 +0000 (19:45 +0000)]
Convert to critical sections, move timer reset into the protection.

14 years agoExplicitly note that updating from pre-1.2 to PREVIEW or HEAD is not
Joerg Sonnenberger [Thu, 9 Jun 2005 19:39:18 +0000 (19:39 +0000)]
Explicitly note that updating from pre-1.2 to PREVIEW or HEAD is not
supported, the intermediate step to 1.2 is required.

14 years agoUse M_WAITOK for contigmalloc now that the attach path is interrupt-save.
Joerg Sonnenberger [Thu, 9 Jun 2005 19:29:00 +0000 (19:29 +0000)]
Use M_WAITOK for contigmalloc now that the attach path is interrupt-save.

14 years agoDon't bzero the softc, it is already zero.
Joerg Sonnenberger [Thu, 9 Jun 2005 19:26:55 +0000 (19:26 +0000)]
Don't bzero the softc, it is already zero.
Use ste_detach for cleanup if attach fails.
Setup interrupt last to prevent possible races.

Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
Inspired-by: related changes to wb(4)
14 years agoReplace the hand rolled linked list with a SLIST.
Eirik Nygaard [Thu, 9 Jun 2005 19:14:12 +0000 (19:14 +0000)]
Replace the hand rolled linked list with a SLIST.

14 years agoUse ether_crc32_be.
Joerg Sonnenberger [Thu, 9 Jun 2005 19:13:34 +0000 (19:13 +0000)]
Use ether_crc32_be.

Obtained-from: FreeBSD
Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoUse PCI helper functions instead of hand-rolling them. Remove now
Joerg Sonnenberger [Thu, 9 Jun 2005 19:10:04 +0000 (19:10 +0000)]
Use PCI helper functions instead of hand-rolling them. Remove now
unused macros.

Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agodevice_printf / if_printf and some minor nits.
Joerg Sonnenberger [Thu, 9 Jun 2005 19:05:16 +0000 (19:05 +0000)]
device_printf / if_printf and some minor nits.

Submitted-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoDeprecate MCLGET() in favor of m_getl() in order to take advantage
Jeffrey Hsu [Thu, 9 Jun 2005 18:39:05 +0000 (18:39 +0000)]
Deprecate MCLGET() in favor of m_getl() in order to take advantage
of cluster caching.

14 years agoFix a bug in the mbstats accounting. m_mbufs was being decremented for
Matthew Dillon [Thu, 9 Jun 2005 18:26:22 +0000 (18:26 +0000)]
Fix a bug in the mbstats accounting.  m_mbufs was being decremented for
normal mbufs converted to clustered mbufs, but in fact it should not be
because we decrement it properly when the normal-mbuf-with-cluster is
freed.

Change most of the m_getl()'s back to m_get()'s or m_gethdr()'s in order
to guarentee that the mbuf does NOT have a cluster associated with it
since we are going to try to share the old mbuf's cluster later on
(in m_copym()).  This bug was causing the system to run out of mbuf
clusters.

Add a number of assertions for correct operation.

14 years agoAlso preserve all the non-copied flags in the target mbuf when duplicating a
Jeffrey Hsu [Thu, 9 Jun 2005 17:14:13 +0000 (17:14 +0000)]
Also preserve all the non-copied flags in the target mbuf when duplicating a
packet header.

Requested by: Matt Dillon

14 years agoFix some bugs in the last commit. We have to call ether_ifdetach if we
Joerg Sonnenberger [Thu, 9 Jun 2005 17:09:34 +0000 (17:09 +0000)]
Fix some bugs in the last commit. We have to call ether_ifdetach if we
couldn't setup the interrupt. We should always try to detach the children.

Noticed-by: Sepherosa Ziehau <sepherosa@gmail.com>
14 years agoPreserve the target M_EXT_CLUSTER flag when duplicating a packet header.
Jeffrey Hsu [Thu, 9 Jun 2005 16:53:11 +0000 (16:53 +0000)]
Preserve the target M_EXT_CLUSTER flag when duplicating a packet header.

14 years agoAdd a missing lwkt_reltoken() in the NULL return path. Do not count NULL
Matthew Dillon [Thu, 9 Jun 2005 16:53:10 +0000 (16:53 +0000)]
Add a missing lwkt_reltoken() in the NULL return path.  Do not count NULL
returns in the gets_cumulative statistic.

14 years agoAdd directories for gcc40
Simon Schubert [Thu, 9 Jun 2005 14:35:29 +0000 (14:35 +0000)]
Add directories for gcc40

14 years agoThe header type of a mbuf doesn't change when appended onto a chain.
Jeffrey Hsu [Thu, 9 Jun 2005 02:03:39 +0000 (02:03 +0000)]
The header type of a mbuf doesn't change when appended onto a chain.

14 years agoA packet header without any packet tags is still a packet header.
Jeffrey Hsu [Thu, 9 Jun 2005 01:55:12 +0000 (01:55 +0000)]
A packet header without any packet tags is still a packet header.