Reflect a recent change in /etc/rc.subr that made ``-M''
appear in tmpmfs and varmfs default flags explicitly.
Explain why -M is good for these file systems (it maximizes
performance and makes the system more stable at low memory
conditions by reducing the chance of thrashing.)
Add a new struct buf flag bit, B_PERSISTENT, and use it to tag
struct bufs that are persistently held by ext2fs. Ignore any buffers
with this flag in the code in boot() that counts "busy" and dirty
buffers and attempts to sync the dirty buffers, which is done before
attempting to unmount all the file systems during shutdown.
This fixes the problem caused by any ext2fs file systems that are
mounted at system shutdown time, which caused boot() to give up on
a non-zero number of buffers and skip the call to vfs_unmountall().
This left all the mounted file systems in a dirty state and caused
them to all require cleanup by fsck on reboot.
Move the two separate copies of the "busy" buffer test in boot()
to a separate function.
Nuke the useless spl() stuff in the ext2fs ULCK_BUF() macro.
Bring the PRINT_BUF_FLAGS definition in sys/buf.h up to date with
this and previous flag changes.
PR: kern/56675, kern/85163
Tested by: "Matthias Andree" matthias.andree at gmx.de
Reviewed by: bde
MFC after: 3 days
Import new version of the HPTMV driver from Highpoint. The major change
here is the support for amd64, as well as possible support for PAE. Many
thanks to Highpoint for continuing to support FreeBSD.
Obtained from: Steve Chang @ Highpoint
MFC After: 3 days.
o Note that the first 255 locations are reserved for JEDEC Ids from
publication 106 (current revision Q, each one verified with
JEDEC and the PMCICA).
o Move ADAPTEC2 to the right section.
o Sort TOSHIBA2 numerically.
Free the bpf descriptor array after we are done with it. This probably isnt
that critical as the program exits after this point anyway, but this may
not always be the case.
Several changes: Added a BUGS section with several bugs. And
-- Made the synopses more precise.
-- Added argument to flag in option description.
-- Moved -b default and limits to option description (to un-hide).
-- Noted several behaviors that were not mentioned.
-- A few more trivial changes.
PR: docs/46787
Approved by: keramida
MFC after: 3 days
Merge bpfstat's functionality into the netstat(1) utility. This adds
a -B option which causes bpf peers to be printed. This option can be
used in conjunction with -I if information about specific interfaces
is desired. This is similar to what NetBSD added to their version of
netstat.
-Add bpf.c which stores all the code for retrieving and parsing bpf
related statistics.
-Modify main.c to add support for the -B option and hook it into the
program logic.
-Add bpf.c to the build.
-Document this new functionality in the man page and bump the revision
date.
-Add prototype for bpf_stats function.
IPv6 was improperly defining its malloc type the same as IPv4 (M_IPMADDR,
M_IPMOPTS, M_MRTABLE). Thus we had conflicting instantiations.
Create an IPv6-specific type to overcome this.
Unbreak release builds. Now that NO_INFO also affects the installation
of texinfo and we need makeinfo to build world, we cannot have NO_INFO
anymore when we populate the chroot.
Minor reorg of pccard attach code to fix ax88x90 case broken after fixing
dl100xx case.
o We no longer acquire and release resources during attach many times. We now
do it once at the beginning.
o Move setting the resource offsets to just after acquiring the ports in
attach.
o Move ax88x90 code to the end of the file, just after the dl100xx specific
code.
o Rename ed_pccard_Linksys to ed_pccard_dl100xx to reflect the underlying
chipset.
o Pass the ed_product structure into ed_pccard_{dl100xx,ax88x90} and have
those routines test the flags to see if this card should be probed in that
way.
o transition from ed_probe_Novell to ed_probe_Novell_generic since we already
have the resources setup.
o Move use of ed_probe_Novell_generic into ed_pccard_dl100xx to be more
consistant with ax88x90 case.
o simplify the code where we probe for the chipsets
o Expose ed_probe_Novel_generic, it is useful for pccard case.
o Fix a now stale comment. There are likely many others like them, but this
one definitely hasn't been true since DELAY was introduced into the tree.
Fix DL100xx based cards. Don't drop resources. It was approrpiate in
the probe code that this used to be part of, but as part of the
attach, we shouldn't be dropping the resources here.
Also, allocate the proper rid in the ax88x90 setup.
Work around a bug I've seen on Linksys EC2T cards. For some reason,
as yet unknown, those cards report their MAC address a byte at a time.
However, other AX88x90 cards report the MAC address a word at a time.
Add a heuristic which looks at the high order bytes of the first 6
words. If they are all '0', assume the card is behaving like the
Linksys EC2T card. Since the default prefix for these cards appears
to be 00:e0:98, this appears to be a safe heuristic. While some cards
have been observed with different prefixes, they all work with this
heuristic.
I'm unsure if this is a bug in the EC2T card, or if it is a bug in the
initialization of the card. No other OS has this heuristic (although
w/o it, the MAC address that is used works).
It turns out that many of the ambicom cards have their CIS resources
listed in different orders. Since it is easy to identify the Modem
resources vs the Ethernet resources by looking at the size, use that
rather than hard coded rids. For such parts, go ahead and guess which
rid we should use based on the size. This guess appears reliable for
the two example cards that I have with different CIS info.
Add support for multicast to the bridge and allow inet6 addresses to be
assigned to the interface.
IPv6 auto-configuration is disabled. An IPv6 link-local address has a
link-local scope within one link, the spec is unclear for the bridge case and
it may cause scope violation.
An address can be assigned in the usual way;
ifconfig bridge0 inet6 xxxx:...
The sub-expression "(nulonly || 1)" always evaluates to true and
according to CVS logs seems to be just a left-over from some
debugging and introduced by accident. Removing the sub-expression
doesn't change semantics and a code inspection showed that the
variable "nulonly" is also not necessary here in any way (and the
expression would require fixing instead of removing).
2. Remove dead code:
- if (backslash && c == '\\') {
- if (read(STDIN_FILENO, &c, 1) != 1) {
- status = 1;
- break;
- }
- STPUTC(c, p);
- } else if (ap[1] != NULL && strchr(ifs, c) != NULL) {
+ if (ap[1] != NULL && strchr(ifs, c) != NULL) {
Inspection of the control and data flow showed that variable
"backslash" is always false (0) when the "if"-expression is
evaluated, hence the whole block is effectively dead code.
Additionally, the skipping of characters after a backslash is already
performed correctly a few lines above, so this code is also not
needed at all. According to the CVS logs and the ASH 0.2 sources,
this code existed in this way already since its early days.
The expression wants to ensure the trap is not assigned the empty
string. But the "!" operator has higher precedence than "==", so the
comparison should be put into parenthesis to form the intended way of
expression. Nevertheless the code was effectively not really broken
as both particular NUL comparisons are semantically equal, of course.
But the parenthesized version is a lot more intuitive.
4. Remove shadowing variable declaration:
- char *q;
The declaration of symbol "q" hides another identical declaration of
"q" in the same context. As the other "q" is already reused multiple
times and also can be reused again without negative side-effects,
just remove the shadowing declaration.
5. Just small cosmetics:
- if (ifsset() != 0)
+ if (ifsset())
The ifsset() macro is already coded by returning the boolean result
of a comparison operator, so no need to compare this boolean result
again against a numerical value. This also aligns the macros usage to
the remaining existing code.
Teach portsnap how to ignore unwanted parts of the ports tree. A line
of the form "REFUSE foo" in portsnap.conf will result in parts of the
tree matching "^foo" being (a) not extracted by "portsnap extract", (b)
not updated by "portsnap update", and (c) not having any patches or new
ports downloaded by "portsnap fetch" or "portsnap cron". The example
shown in portsnap.conf demonstrates ignoring all the language categories.
As mentioned in portsnap.conf.5, the use of an imcomplete ports tree is
not officially supported; but this is something which many users have
requested, so I'm adding it anyway.
PR: bin/85619 (but not the patch provided therein)
MFC after: 1 month
When we read data from socket buffer using soreceive() the socket layer
does not clear m_nextpkt for us. The mbufs are sent into netgraph and
then, if they contain a TCP packet delivered locally, they will enter
socket code again. They can pass the first assert in sbappendstream()
because m_nextpkt may be set not in the first mbuf, but deeper in the
chain. So the problem will trigger much later, when local program
reads the data from socket, and an mbuf with m_nextpkt becomes a
first one.
This bug was demasked by revision 1.54, when I made upcall queueable.
Before revision 1.54 there was a very small probability to have 2
mbufs in GRE socket buffer, because ng_ksocket_incoming2() dequeued
the first one immediately.
- in ng_ksocket_incoming2() clear m_nextpkt on all mbufs
read from socket.
- restore rev. 1.54 change in ng_ksocket_incoming().
PR: kern/84952
PR: kern/82413
In collaboration with: rwatson
Raise one more bit in READER_MASK. I believe that before this change
it was possible to have 1 reader and 1 writer thread working on
a node simultaneously.
sam [Tue, 6 Sep 2005 03:59:22 +0000 (03:59 +0000)]
updates:
o note all pci/cardbus parts are supported (modulo hal updates)
o use ath_rate_sample instead of ath_rate_onoe
o note SuperG support is missing
o note WPA not supported on 5210
o remove stuff about needing a better tx rate control algorithm
Convert the primary ACL allocator from malloc(9) to using a UMA zone instead.
Also introduce an aclinit function which will be used to create the UMA zone
for use by file systems at system start up.
Now that the PID is constantly refreshed, introduce a pidname function which
when given a PID consults the kern.proc.pid sysctl variable to pull out the
pcomm data.
Instead of caching the PID which opened the bpf descriptor, continuously
refresh the PID which has the descriptor open. The PID is refreshed in various
operations like ioctl(2), kevent(2) or poll(2). This produces more accurate
information about current bpf consumers. While we are here remove the bd_pcomm
member of the bpf stats structure because now that we have an accurate PID we
can lookup the via the kern.proc.pid sysctl variable. This is the trick that
NetBSD decided to use to deal with this issue.
Special care needs to be taken when MFC'ing this change, as we have made a
change to the bpf stats structure. What will end up happening is we will leave
the pcomm structure but just mark it as being un-used. This way we keep the ABI
in tact.
MFC after: 1 month
Discussed with: Rui Paulo < rpaulo at NetBSD dot org >
o In pmap_remove_pte: always invalidate the page. Previously the page
was not invalidated if the PTE was not actually being removed. In
an UP kernel this didn't cause problems, because the new mapping
would preempt the old one. In an SMP kernel this could lead to the
use of stale translations when processes move between CPUs at the
"right" moment. This fixes the last of the obvious SMP problems
and it should be safe to enable SMP by default now.
o In pmap_remove_pte: minor code refactoring to avoid duplication.
o Test all PTE pointers against NULL. Don't use implicit boolean
tests.
Fix panic when loading libiconv.ko on sparc64.
It seems that this issue only become obvious when compiled with -O2
on sparc64.
Since each struct iconv_converter_class has been initialized by
DEFINE_CLASS macro, not all members of struct iconv_converter_class
has been allocated on memory and cc_link member has not been
initialized, while iconv_register_converter() wanted to access it
with TAILQ.
Now we modify KICONV_CONVERTER macro and fix this bug.
Problem reported on: freebsd-sparc64
Pointed out by: yongari
Discussed with: yongari
Tested by: yongari
MFC after: 3 days
Remove Giant mutex from polling(4) and use a separate poll_mtx(4)
instead. Detailed changelist:
o Add flags field to struct pollrec, to indicate that
are particular entry is being worked on.
o Define a macro PR_VALID() to check that a pollrec
is valid and pollable.
o Mark ISRs as mpsafe.
o ether_poll()
- Acquire poll_mtx while traversing pollrec array.
- Skip pollrecs, that are being worked on.
- Conditionally acquire Giant when entering handler.
o netisr_pollmore()
- Conditionally assert Giant.
- Acquire poll_mtx while working with statistics.
o netisr_poll()
- Conditionally assert Giant.
- Acquire poll_mtx while working with statistics
and traversing pollrec array.
o ether_poll_register(), ether_poll_deregister()
- Conditionally assert Giant.
- Acquire poll_mtx while working with pollrec array.
o poll_idle()
- Remove all strange manipulations with Giant.
In collaboration with: ru, pjd
In collaboration with: Oleg Bulyzhin <oleg rinet.ru>
In collaboration with: dima <_pppp mail.ru>
In accordance with my intentions announced (and not objected to)
on -arch, and RFC 4159 (http://www.rfc-editor.org/rfc/rfc4159.txt)
which officially deprecates all usage of IP6.INT, remove the
reference to that zone from the example named.conf file.
Add a regression test to test two variations on the same bug: joining a
multicast group using a raw socket, then removing the interface on which
the group is found, and joining a multicast group using a udp socket,
then removing the interface on which the group is found. An if_disc
interface is used as the interface on which to attach.
NB: A panic currently results from running this regression test, so do
so with caution.
PR: 77665
Reported by: Gavin Atkinson <gavin dot atkinson at ury dot york dot ac dot uk>
Reported by: Brooks Davis <brooks at FreeBSD dot org>
Attempt to complete the userspace integration of POSIX.1e extended ACLs.
This includes adding support for ACLs into cp(1) and mv(1) userspace
utilities.
For mv(1), if _PC_ACL_EXTENDED is in effect for the source AND destination
operands, the destination file's ACLs shall reflect the source.
For cp(1), if _PC_ACL_EXTENDED is in effect for both source and destination
operands, and -p has been specified, the ACLs from the source shall be
preserved on the destination.
Remove a race condition that could result in processes being stuck
waiting for geom events to happen:
Instead of maintaining a count of outstanding events, simply look if
the queue is empty. Make sure to not remove events from the queue
until they are executed in order to not open a new race.
Much work by: pjd
Tested by: kris
MT6: yes, should be.
Eliminate unnecessary TLB invalidations by pmap_enter(). Specifically,
eliminate TLB invalidations when permissions are relaxed, such as when a
read-only mapping is changed to a read/write mapping. Additionally,
eliminate TLB invalidations when bits that are ignored by the hardware,
such as PG_W ("wired mapping"), are changed.
When bge_stop() is called from bge_detach(), mii is already NULL, so check
before trying to use it.
This fix allows to 'kldunload if_bge' without panicing.
o s/vhpt_size/pmap_vhpt_log2size/g
o s/vhpt_base/pmap_vhpt_base/g
o s/vhpt_bucket/pmap_vhpt_bucket/g
o Declare the above in <machine/pmap.h>
o Move the vm.stats.vhpt.* sysctls to machdep.vhpt.*
o Create a tunable machdep.vhpt.log2size, with corresponding sysctl.
The tunable allows the user to specify the VHPT size from the loader.
o Don't keep track of the number of PTEs in the VHPT. Calculate the
population when necessary by iterating the buckets and summing up
the length of the buckets.
o Don't perform the tpa instruction with a bucket lock held. The
instruction can (theoretically) fault and locking is not needed.
More consistantly return the correct BAR size. Before, we'd only
return the correct bar size if we encountered a 64-bit BAR that had
its resources already assigned. If the resources weren't yet
assigned, we'd bogusly assume it was a 32-bit bar and return 1.
Fix collision chain termination checks. The result of IA64_PHYS_TO_RR7
is never 0, so one cannot test for a NULL pointer after a physical
address is translated into a virtual pointer with said macro. Instead,
keep the physical address around and test it against 0. Note that
this obviously implies that a PTE can never be allocated at physical
address 0. This isn't exactly guaranteed, but hasn't been a problem
so far. We test the physical address against 0 for as long as the ia64
port exists...
If we ignore an unknown % sequence, we must stop interpreting the
remaining % arguments because the varargs are now out of sync and
there is a risk that we might for instance dereference an integer
in a %s argument.