alc [Sun, 18 Feb 2007 05:54:42 +0000 (05:54 +0000)]
Enable vm_page_free() and vm_page_free_zero() to be called on some pages
without the page queues lock being held, specifically, pages that are not
contained in a vm object and not a member of a page queue.
sos [Sat, 17 Feb 2007 16:56:39 +0000 (16:56 +0000)]
Add support for chipsets that has NULL'd BAR's for legacy ports.
This allows DMA to be used on a fine little geode system I got here and
most like on lots of older systems like that.
mckusick [Sat, 17 Feb 2007 08:25:43 +0000 (08:25 +0000)]
This README file is obsolete. The cited problems were fixed long ago
and the code is installed by default so no longer requires action by
the administrator to be included.
imp [Fri, 16 Feb 2007 05:36:59 +0000 (05:36 +0000)]
Two changes:
(1) change debounce period from 1s to 250ms. This appears to be fine and
speeds things up a little.
(2) In the middle of cbb_pcic_power_disable_socket we write 0 to the EXCA_INTR
register to put the card into reset. However, this turns off CSC
interrupts for TI bridges (and maybe others). So no further card
insertion events would be noticed. To compensate, after we've gone
through the entire power down sequence, turn on EXCA_INTR_ENABLE so
that CSC events happen.
#2 should fix the 'dead slot' problem that has been reported after
card ejection (but only 16-bit cards).
mohans [Fri, 16 Feb 2007 03:46:55 +0000 (03:46 +0000)]
Backing out an earlier change. It seems harmless for NFS to miss the "force
unmount" flag, making the acquisition of the MNT_ILOCK in nfs_request() and
nfs_sigintr() unnecessary. Pointed out by tegge@.
bmah [Fri, 16 Feb 2007 02:20:11 +0000 (02:20 +0000)]
Switch to unified MI hardware notes. The hardware notes are
now contained entirely in a single DocBook article, which has
information pertinent to all architectures. This will improve
the maintainability of the hardware notes going forward by
reducing complexity of the build process, improving the visibiilty
of the architecture-specific text, and removing the need to generate
an extra version of the document for every architecture.
Remove all of the MD hardware notes stuff; all relevant content
has been folded into hardware/article.sgml.
bmah [Fri, 16 Feb 2007 02:10:12 +0000 (02:10 +0000)]
Add support for auto-generating hardware notes entries in a
manner consistent with the new MI-style hardware notes document.
man2hwnotes.pl now defaults to generating entities for the MI-style
hardware notes (i.e. hardware/article.sgml). A new -c
option causes it to generate entities for the older MD-style
hardware notes (i.e. hardware/common/dev.sgml).
The Makefile infrastructure supplies the -c option to man2hwnotes.pl
now unless the HWNOTES_MI Makefile variable is defined, so
compatiblity is preserved for hardware notes translations that
aren't converted to the new organization yet. As translations
convert, they should define HWNOTES_MI in their hardware/Makefile.
When all the relevant translations catch up, the compatibility goop
in share/mk/doc.relnotes.mk and share/sgml/Makefile can be removed.
Thanks go to simon@ for help with the backwards compatiblity
mechanism.
pjd [Thu, 15 Feb 2007 22:08:35 +0000 (22:08 +0000)]
Move vnode-to-file-handle translation from vfs_vptofh to vop_vptofh method.
This way we may support multiple structures in v_data vnode field within
one file system without using black magic.
Vnode-to-file-handle should be VOP in the first place, but was made VFS
operation to keep interface as compatible as possible with SUN's VFS.
BTW. Now Solaris also implements vnode-to-file-handle as VOP operation.
VFS_VPTOFH() was left for API backward compatibility, but is marked for
removal before 8.0-RELEASE.
Approved by: mckusick
Discussed with: many (on IRC)
Tested with: ufs, msdosfs, cd9660, nullfs and zfs
sos [Thu, 15 Feb 2007 21:51:42 +0000 (21:51 +0000)]
Support AHCI chips where the ports are not consecutively numbered as in
some incarnations of the ICH8 chip.
Also fix the panic introduced by the last commit.
bmah [Thu, 15 Feb 2007 19:33:47 +0000 (19:33 +0000)]
First step in consolidating the various MD bits of the hardware notes.
Convert MD references in the supported devices section from arch=""
attributes for conditional compilation to entities that will just
print architecture names. (The entities aren't defined yet...this
will happen in a future commit.)
luigi [Thu, 15 Feb 2007 17:21:31 +0000 (17:21 +0000)]
Cleanup and document the implementation of firmware(9) based on
a version that i posted earlier on the -current mailing list,
and subsequent feedback received.
The core of the change is just in sys/firmware.h and kern/subr_firmware.c,
while other files are just adaptation of the clients to the ABI change
(const-ification of some parameters and hiding of internal info,
so this is fully compatible at the binary level).
In detail:
- reduce the amount of information exported to clients in struct firmware,
and constify the pointer;
- internally, document and simplify the implementation of the various
functions, and make sure error conditions are dealt with properly.
The diffs are large, but the code is really straightforward now (i hope).
Note also that there is a subtle issue with the implementation of
firmware_register(): currently, as in the previous version, we just
store a reference to the 'imagename' argument, but we should rather
copy it because there is no guarantee that this is a static string.
I realised this while testing this code, but i prefer to fix it in
a later commit -- there is no regression with respect to the past.
Note, too, that the version in RELENG_6 has various bugs including
missing locks around the module release calls, mishandling of modules
loaded by /boot/loader, and so on, so an MFC is absolutely necessary
there. I was just postponing it until this cleanup to avoid doing
things twice.
rwatson [Thu, 15 Feb 2007 14:44:46 +0000 (14:44 +0000)]
Expand history and authors section of mbuf.9 man page to discuss recent
transition to mbuma (FreeBSD 5.3) and the fact that mbufs are now limited
almost entirely to packet storage, with straight UMA zones being used for
most other network data types.
rwatson [Thu, 15 Feb 2007 10:50:48 +0000 (10:50 +0000)]
Break file descriptor printing logic out of db_show_files() into
db_print_file(), and add a new "show file <ptr>" DDB command, which can
be used to print out file descriptors referenced in stack traces.
kib [Thu, 15 Feb 2007 09:53:49 +0000 (09:53 +0000)]
If both ISDOTDOT and NOCROSSMOUNT are set then lookup() might breaks out
of the special handling for ".." and perform an ISDOTDOT VOP_LOOKUP()
for a filesystem root vnode. Handle this case inside lookup().
imp [Thu, 15 Feb 2007 07:22:27 +0000 (07:22 +0000)]
It turns out that it is easier to not NULL out pccard and cardbus
device pointers. They don't change as the children device drivers
come and go. Rather, check to see if the device is attached where we
would have checked ! NULL. This solves many asymmetries in the code
that likely could lead to crashes when loading/unloading cbb without
one or more of the expected children's driver not present.
imp [Thu, 15 Feb 2007 07:13:38 +0000 (07:13 +0000)]
Fix three bugs:
o When detaching all children, try really hard to get all the children
list before giving up. This is based on an observation by hans petter
selasky in his usb p4 branch.
o When rescanning devices after a driver is added, abort if we can't get
the child list with a message.
o when rescanning devices, if the reprobe/attach is successful, save the
device for cardbus/pccard.
yongari [Thu, 15 Feb 2007 06:21:34 +0000 (06:21 +0000)]
Unbreak non-H/W VLAN extraction case.
Unlike other GigEs Yukon II always set VLAN bit when it detects VLAN
tagged packet regardless of H/W VLAN processing configuration state.
So it need to check IFCAP_VLAN_HWTAGGING bit to know whether driver
is configured to take advantage of H/W VLAN processing. If H/W VLAN
processing was disabled don't adjust received packet length such that
subsequent validation logic works for software VLAN processing.
alc [Thu, 15 Feb 2007 05:43:38 +0000 (05:43 +0000)]
Relax the page queue lock assertions in vm_page_remove() and
vm_page_free_toq() to account for recent changes that allow
vm_page_free_toq() to be called on some pages without the page queues lock
being held, specifically, pages that are not contained in a vm object and
not a member of a page queue. (Examples of such pages include page table
pages, pv entry pages, and uma small alloc pages.)
- PROT_READ, PROT_WRITE, or PROT_EXEC implies PROT_READ and PROT_EXEC.
Linux/ia64's i386 emulation layer does this and it complies with Linux
header files. This fixes mmap05 LTP test case on amd64.
- Do not adjust stack size when failure has occurred.
- Synchronize i386 mmap/mprotect with amd64.
jhb [Wed, 14 Feb 2007 22:36:27 +0000 (22:36 +0000)]
Adjust the global MSI blacklisting strategy so we don't have to explicitly
blacklist a bunch of old chipsets. If a system contains a PCI-PCI bridge
that supports PCI-X, assume the chipset supports PCI-X. If a system
contains a PCI-express root port, assume the chipset supports PCI-express.
If the chipset doesn't support either PCI-X or PCI-express, then blacklist
it by default. We should now only need to explicitly blacklist PCI-X or
PCI-express chipsets that don't properly handle MSI.
jhb [Wed, 14 Feb 2007 22:32:55 +0000 (22:32 +0000)]
- Fix an off by one error in pci_remap_msix_method() that effectively
broke the method as all the MSI-X table indices were off by one in
the backend MD code.
- Fix a cosmetic nit in the bootverbose printf in pci_alloc_msix_method().
jkim [Wed, 14 Feb 2007 19:44:15 +0000 (19:44 +0000)]
Use bge_writereg_ind() to do global reset as we did before 1.159 for certain
chipsets. It was causing 'firmware handshake timed out' errors for some
chips.
bms [Wed, 14 Feb 2007 14:17:01 +0000 (14:17 +0000)]
Retire most of the classful network behaviour of netstat -r output, for IPv4.
Without -n, we now only print a "network name" without the prefix length
under the following conditions:
1) the network address and mask matches a classful network prefix;
2) getnetbyaddr(3) returns a network name for this network address.
With -n, we unconditionally print the full unabbreviated CIDR network
prefix in the form "a.b.c.d/p". 0.0.0.0/0 is still printed as "default".
This change is in preparation for changes such as equal-cost multipath, and
to more generally assist operational deployment of FreeBSD as a modern IPv4
router. There are currently no plans to backport this change.
rwatson [Wed, 14 Feb 2007 12:22:11 +0000 (12:22 +0000)]
Change unp_mtx to supporting recursion, and do not drop the unp_mtx over
sonewconn() in unp_connect(). This avoids a race that occurs due to
v_socket being an uncounted reference, as the lock was being released in
order to call sonewconn(), which otherwise recurses into the UNIX domain
socket code via pru_attach, as well as holding the lock over a sleeping
memory allocation in uipc_attach(). Switch to a non-sleeping memory
allocation during UNIX domain socket attach.
This fix non-ideal in that it requires enabling recursion, but is a much
smaller change than moving to using true references for v_socket. The
reported panic occurs in unp_connect() following the return of
sonewconn().
luoqi [Wed, 14 Feb 2007 09:10:37 +0000 (09:10 +0000)]
The return value of aac_alloc_command() was misinterpreted in aac_ioctl_event().
Once triggered this would leak away all available commands and starve the rest
of the driver.
kientzle [Wed, 14 Feb 2007 08:16:08 +0000 (08:16 +0000)]
Correctly handle writes beyond the end of the archive entry
(as determined by the initial size given to the header).
Libarchive recently changed to correctly return the amount
of data actually consumed in this case, which revealed this
bug in bsdtar.
alc [Wed, 14 Feb 2007 07:05:55 +0000 (07:05 +0000)]
Avoid the unnecessary acquisition of the free page queues lock when a page
is actually being added to the hold queue, not the free queue. At the same
time, avoid unnecessary tests to wake up threads waiting for free memory
and the idle thread that zeroes free pages. (These tests will be performed
later when the page finally moves from the hold queue to the free queue.)
cperciva [Wed, 14 Feb 2007 05:21:22 +0000 (05:21 +0000)]
Optimize bitcount32 by replacing 6 logical operations with 2. The key
observation here is that it doesn't matter what garbage accumulates in
bits which we're going to end up masking away anyway, as long as the
garbage doesn't overflow into bits which we care about.
This improved version may not be the fastest possible on all systems,
but it's certainly going to be better than what was here before.