The mbuf and mcluster free lists now each "own" a condition variable,
m_starved.
- Clean up minor indentention issues in sys/mbuf.h caused by previous
commit.
to not using IO_SYNC. Expose a sysctl (debug.ufs_extattr_sync) for
enabling the use of IO_SYNC.
- Use of IO_SYNC substantially degrades ACL performance when a
default ACL is set on a directory, as there are four synchronous
writes initiated to define both supporting EAs for new
sub-directories, and to set the data; two for new files. Later, this
may be optimized to two writes for sub-directories, one for new
files.
- IO_SYNC does not substantially improve consistency properties due
to the poor consistency properties of existing permissions (which
ACLs are a superset of), due to interaction with soft updates,
and due to differences in handling consistency for data and file
system meta-data.
- In macro-benchmarks, this reduces the overhead of setting default
ACLs down to the same overhead as enabling ACLs on a file system
and not using them. Enabling ACLs still introduces a small
overhead (I measure 7% on a -j 2 buildworld with pre-allocated
EA backing store, but this is not rigorous testing, nor in any way
optimized).
- The sysctl will probably change to another administration method
(or at least, a better name) in the near future, but consistency
properties of EAs are still being worked out. The toggle is defined
right now to allow easier performance analysis and exploration
of possible guarantees.
Obtained from: TrustedBSD Project
Don't use atomic operations for the stats updating, instead protect
the counts with the mbuf mutex. Most twiddling of the stats was
done right before or after releasing a mutex. By doing this we
reduce the number of locked ops needed as well as allow a sysctl
to gain a consitant view of the entire stats structure.
In the future...
This will allow us to chain common mbuf operations that would
normally need to aquire/release 2 or 3 of the locks to build an
mbuf with a cluster or external data attached into a single op
requiring only one lock.
Simplify the per-cpu locks that are planned.
There's also some if (1) code that should check if the "how"
operation specifies blocking/non-blocking behavior, we _could_ make
it so that we hold onto the mutex through calls into kmem_alloc
when non-blocking requests are made, but for safety reasons we
currently drop and reaquire the mutex around the calls.
Also, note that calling kmem_alloc is rare and only happens during
a shortage so drop/re-getting the mutex will not be a common
occurance.
Remove some #define's that seemed to obfuscate the code to me.
Remove an extranious comment.
Remove an XXX, including mutex.h isn't a crime.
Reviewed by: bmilekic
avoid silly lock contention on sched_lock since in 2 out of the 3 places
that we call stop(), we get sched_lock right after calling it and we were
locking sched_lock inside of stop() anyways.
failures in MOD_LOAD.
Dodge duplicate make_dev() calls by (ab)using dev->si_drv2 to
remember if we created the device node via a dev_clone callback
before the d_open call.
Without this, ifpromisc() always fails (after setting the IFF_PROMISC
bit in ifp->if_flags) and bpf never bothers to turn promiscuous mode off.
PR: 20188
SIGCHLD to our parent process. Otherwise, we could block while obtaining
the process lock for our parent process and switch out while we were
in SSTOP. Even worse, when we try to resume from the mutex being blocked
on our p_stat will be SRUN, not SSTOP.
- Fix a comment above stop() to indicate that it requires that the proc lock
be held, not a proctree lock.
Reported by: markm
Sleuthing by: jake
under heavy use when default ACLs were bgin inherited by new files
or directories. This is done by removing a bug in default ACL
reading, and improving error handling for this failure case:
- Move the setting of the buffer length (len) variable to above the
ACL type (ap->a_type) switch rather than having it only for
ACL_TYPE_ACCESS. Otherwise, the len variable is unitialized in
the ACL_TYPE_DEFAULT case, which generally worked right, but could
result in failure.
- Add a check for a short/long read of the ACL_TYPE_DEFAULT type from
the underlying EA, resulting in EPERM rather than passing a
potentially corrupted ACL back to the caller (resulting "cleaner"
failures if the EA is damaged: right now, the caller will almost
always panic in the presence of a corrupted EA). This code is similar
to code in the ACL_TYPE_ACCESS handling in the previous switch case.
- While I'm fixing this code, remove a redundant bzero() of the ACL
reader buffer; it need only be initialized above the acl_type
switch.
Obtained from: TrustedBSD Project
operations on file descriptors, which complement the existing set of
calls, extattr_{delete,get,set}_file() which act on paths. In doing
so, restructure the system call implementation such that the two sets
of functions share most of the relevant code, rather than duplicating
it. This pushes the vnode locking into the shared code, but keeps
the copying in of some arguments in the system call code. Allowing
access via file descriptors reduces the opportunity for race
conditions when managing extended attributes.
Obtained from: TrustedBSD Project
ps_showallprocs such that if superuser is present to override process
hiding, the search falls through [to success]. When additional
restrictions are placed on process visibility, such as MAC, new clauses
will be placed above the return(0).
Obtained from: TrustedBSD Project
than a NOP. bounds_check_with_label() would return -1 yet NOT set any
of the bio flags to show an error. This meant the caller would not
properly see that bounds_check_with_label() did not do any work. This
prevented newfs(8) from being able to write a file system on any partition
other than `c' on a `ccd'.
The logs of this file do not tell _why_ bounds_check_with_label() was
emasculated. Nor are there any `XXX' comments. So we'll unemasculated
it, and see what breaks.
Submitted by: gallatin
a #defined constant, wrap a few long lines, etc... Also remove stupid
'all your base are belong to us' joke from comment that I don't really
care to see immortalized in the source tree.
- Added 4 speaker enable to initialization sequence.
- Removed delays between register pokes which appear to aggravate a
problem this card has sampling at 44.1kHz. With any form of delay,
skew relative to system clock at 44.1kHz is usually in range 0-25%
(now 0-3%). No other rates exhibit this problem.
- Changed structs cmi_* to sc_*.
Approved by: Cameron Grant <gandalf@vilnya.demon.co.uk>
aic7xxx_pci.c:
Enable board generation of interrupts only once our handler is
in place and all other setup has occurred.
aic7xxx.c:
More conversion of data types to ahc_* names. tmode_tstate and
tmode_lstate are the latest victims.
Clean up the check condition path by branching early rather
than indenting a giant block of code.
Add support for target mode initiated sync negotiation.
The code has been tested by forcing the feature on for
all devices, but for the moment is left inaccesible until
a decent mechanism for controlling the behavior is complete.
Implementing this feature required the removal of the
old "target message request" mechanism. The old method
required setting one of the 16 bit fields to initiate
negotiation with a particular target. This had the nice
effect of being easy to change the request and have it
effect the next command. We now set the MK_MESSAGE bit
on any new command when negotiation is required. When
the negotiation is successful, we walk through and clean
up the bit on any pending commands. Since we have to walk
the commands to reset the SCSI syncrate values so no additional
work is required. The only drawback of this approach is that
the negotiation is deferred until the next command is queued to
the controller. On the plus side, we regain two bytes of
sequencer scratch ram and 6 sequencer instructions.
When cleaning up a target mode instance, never remove the
"master" target mode state object. The master contains
all of the saved SEEPROM settings that control things like
transfer negotiations. This data will be cloned as the
defaults if a target mode instance is re-instantiated.
Correct a bug in ahc_set_width(). We neglected to update
the pending scbs to reflect the new parameters. Since
wide negotiation is almost always followed by sync
negotiation it is doubtful that this had any real
effect.
When in the target role, don't complain about
"Target Initiated" negotiation requests when an initiator
negotiates with us.
Defer enabling board interrupts until after ahc_intr_enable()
is called.
Pull all info that used to be in ahc_timeout for the FreeBSD
OSM into ahc_dump_card_state(). This info should be printed
out on all platforms.
aic7xxx.h:
Add the SCB_AUTO_NEGOITATE scb flag. This allows us to
discern the reason the MK_MESSAGE flag is set in the hscb
control byte. We only want to clear MK_MESSAGE in
ahc_update_pending_scbs() if the MK_MESSAGE was set due
to an auto transfer negotiation.
Add the auto_negotiate bitfield for each tstate so that
behavior can be controlled for each of our enabled SCSI
IDs.
Use a bus interrupt handler vector in our softc rather
than hard coding the PCI interrupt handler. This makes
it easier to build the different bus attachments to
the aic7xxx driver as modules.
aic7xxx.reg:
Remove the TARGET_MSG_REQUEST definition for sequencer ram.
aic7xxx.seq:
Fix a few target mode bugs:
o If MK_MESSAGE is set in an SCB, transition to
message in phase and notify the kernel so that
message delivery can occur. This is currently
only used for target mode initiated transfer
negotiation.
o Allow a continue target I/O to compile without
executing a status phase or disconnecting. If
we have not been granted the disconnect privledge
but this transfer is larger than MAXPHYS, it may
take several CTIOs to get the job done.
Remove the tests of the TARGET_MSG_REQUEST field in scratch ram.
aic7xxx_freebsd.c:
Add support for CTIOs that don't disconnect. We now defer
the clearing of our pending target state until we see a
CTIO for that device that has completed sucessfully.
Be sure to return early if we are in a target only role
and see an initiator only CCB type in our action routine.
If a CTIO has the CAM_DIS_DISCONNECT flag set, propogate
this flag to the SCB. This flag has no effect if we've
been asked to deliver status as well. We will complete
the command and release the bus in that case.
Handle the new auto_negotiate field in the tstate correctly.
Make sure that SCBs for "immediate" (i.e. to continue a non
disconnected transaction) CTIO requests get a proper mapping
in the SCB lookup table. Without this, we'll complain when
the transaction completes.
Update ahc_timeout() to reflect the changes to ahc_dump_card_state().
aic7xxx_inline.h:
Use ahc->bus_intr rather than ahc_pci_intr.
two subject ucreds. Unlike p_cansee(), u_cansee() doesn't have
process lock requirements, only valid ucred reference requirements,
so is prefered as process locking improves. For now, back p_cansee()
into u_cansee(), but eventually p_cansee() will go away.
Reviewed by: jhb, tmm
Obtained from: TrustedBSD Project
