enable easy access to the hash chain stats. The raw prefixed versions
dump an integer array to userland with the chain lengths. This cheats
and calls it an array of 'struct int' rather than 'int' or sysctl -a
faithfully dumps out the 128K array on an average machine. The non-raw
versions return 4 integers: count, number of chains used, maximum chain
length, and percentage utilization (fixed point, multiplied by 100).
The raw forms are more useful for analyzing the hash distribution, while
the other form can be read easily by humans and stats loggers.
API for IPI's that isn't tied to the Intel APIC. MD code can still use
the apic_ipi() function or dink with the apic directly if needed to send
MD IPI's.
because:
- it used a better namespace (smp_ipi_* rather than *_ipi),
- it used better constant names for the IPI's (IPI_* rather than
X*_OFFSET), and
- this API also somewhat exists for both alpha and ia64 already.
codecs. Also, add some additional code to check for future cards without
this feature - attempting to initialise them as AC97 cards will hang the
machine.
PR: 26427
Reviewed by: cg
His description of the problem and solution follow. My own tests show
speedups on typical filesystem intensive workloads of 5% to 12% which
is very impressive considering the small amount of code change involved.
------
One day I noticed that some file operations run much faster on
small file systems then on big ones. I've looked at the ffs
algorithms, thought about them, and redesigned the dirpref algorithm.
First I want to describe the results of my tests. These results are old
and I have improved the algorithm after these tests were done. Nevertheless
they show how big the perfomance speedup may be. I have done two file/directory
intensive tests on a two OpenBSD systems with old and new dirpref algorithm.
The first test is "tar -xzf ports.tar.gz", the second is "rm -rf ports".
The ports.tar.gz file is the ports collection from the OpenBSD 2.8 release.
It contains 6596 directories and 13868 files. The test systems are:
1. Celeron-450, 128Mb, two IDE drives, the system at wd0, file system for
test is at wd1. Size of test file system is 8 Gb, number of cg=991,
size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current
from Dec 2000 with BUFCACHEPERCENT=35
2. PIII-600, 128Mb, two IBM DTLA-307045 IDE drives at i815e, the system
at wd0, file system for test is at wd1. Size of test file system is 40 Gb,
number of cg=5324, size of cg is 8m, block size = 8k, fragment size = 1k
OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=50
You can get more info about the test systems and methods at:
http://www.ptci.ru/gluk/dirpref/old/dirpref.html
Test Results
tar -xzf ports.tar.gz rm -rf ports
mode old dirpref new dirpref speedup old dirprefnew dirpref speedup
First system
normal 667 472 1.41 477 331 1.44
async 285 144 1.98 130 14 9.29
sync 768 616 1.25 477 334 1.43
softdep 413 252 1.64 241 38 6.34
Second system
normal 329 81 4.06 263.5 93.5 2.81
async 302 25.7 11.75 112 2.26 49.56
sync 281 57.0 4.93 263 90.5 2.9
softdep 341 40.6 8.4 284 4.76 59.66
"old dirpref" and "new dirpref" columns give a test time in seconds.
speedup - speed increasement in times, ie. old dirpref / new dirpref.
------
Algorithm description
The old dirpref algorithm is described in comments:
/*
* Find a cylinder to place a directory.
*
* The policy implemented by this algorithm is to select from
* among those cylinder groups with above the average number of
* free inodes, the one with the smallest number of directories.
*/
A new directory is allocated in a different cylinder groups than its
parent directory resulting in a directory tree that is spreaded across
all the cylinder groups. This spreading out results in a non-optimal
access to the directories and files. When we have a small filesystem
it is not a problem but when the filesystem is big then perfomance
degradation becomes very apparent.
What I mean by a big file system ?
1. A big filesystem is a filesystem which occupy 20-30 or more percent
of total drive space, i.e. first and last cylinder are physically
located relatively far from each other.
2. It has a relatively large number of cylinder groups, for example
more cylinder groups than 50% of the buffers in the buffer cache.
The first results in long access times, while the second results in
many buffers being used by metadata operations. Such operations use
cylinder group blocks and on-disk inode blocks. The cylinder group
block (fs->fs_cblkno) contains struct cg, inode and block bit maps.
It is 2k in size for the default filesystem parameters. If new and
parent directories are located in different cylinder groups then the
system performs more input/output operations and uses more buffers.
On filesystems with many cylinder groups, lots of cache buffers are
used for metadata operations.
My solution for this problem is very simple. I allocate many directories
in one cylinder group. I also do some things, so that the new allocation
method does not cause excessive fragmentation and all directory inodes
will not be located at a location far from its file's inodes and data.
The algorithm is:
/*
* Find a cylinder group to place a directory.
*
* The policy implemented by this algorithm is to allocate a
* directory inode in the same cylinder group as its parent
* directory, but also to reserve space for its files inodes
* and data. Restrict the number of directories which may be
* allocated one after another in the same cylinder group
* without intervening allocation of files.
*
* If we allocate a first level directory then force allocation
* in another cylinder group.
*/
My early versions of dirpref give me a good results for a wide range of
file operations and different filesystem capacities except one case:
those applications that create their entire directory structure first
and only later fill this structure with files.
My solution for such and similar cases is to limit a number of
directories which may be created one after another in the same cylinder
group without intervening file creations. For this purpose, I allocate
an array of counters at mount time. This array is linked to the superblock
fs->fs_contigdirs[cg]. Each time a directory is created the counter
increases and each time a file is created the counter decreases. A 60Gb
filesystem with 8mb/cg requires 10kb of memory for the counters array.
The maxcontigdirs is a maximum number of directories which may be created
without an intervening file creation. I found in my tests that the best
performance occurs when I restrict the number of directories in one cylinder
group such that all its files may be located in the same cylinder group.
There may be some deterioration in performance if all the file inodes
are in the same cylinder group as its containing directory, but their
data partially resides in a different cylinder group. The maxcontigdirs
value is calculated to try to prevent this condition. Since there is
no way to know how many files and directories will be allocated later
I added two optimization parameters in superblock/tunefs. They are:
int32_t fs_avgfilesize; /* expected average file size */
int32_t fs_avgfpdir; /* expected # of files per directory */
These parameters have reasonable defaults but may be tweeked for special
uses of a filesystem. They are only necessary in rare cases like better
tuning a filesystem being used to store a squid cache.
I have been using this algorithm for about 3 months. I have done
a lot of testing on filesystems with different capacities, average
filesize, average number of files per directory, and so on. I think
this algorithm has no negative impact on filesystem perfomance. It
works better than the default one in all cases. The new dirpref
will greatly improve untarring/removing/coping of big directories,
decrease load on cvs servers and much more. The new dirpref doesn't
speedup a compilation process, but also doesn't slow it down.
Obtained from: Grigoriy Orlov <gluk@ptci.ru>
s/1518/ETHER_MAX_LEN
Some style changes, add some braces, mostly residual from having
a lot of debug hooks added while working on this driver.
Bring in a plethora of changes from NetBSD:
revision 1.58
date: 2001/03/08 11:07:08; author: ichiro; state: Exp; lines: +17 -1
it wait until busy flag disappears.
it was able to prevent some cards with late initializing faling in wi_reset().
revision 1.41
date: 2000/10/13 19:15:08; author: jonathan; state: Exp; lines: +4 -2
Fix wi_intr() to avoid touching card registers during insert/remove events,
when sharing an interrupt with other devices:
check sc->sc_enabled, and drop the interrupt if its' off.
revision 1.30
date: 2000/08/18 04:11:48; author: jhawk; state: Exp; lines: +4 -4
Copy wi_{dst,src}_addr from struct wi_frame into faked-up ether_header
instead of addr1 and addr2. THis means that tcpdump -e will show the
correct MAC address for communications with access points instead of showing
the BSSID.
In the future there should be 802.11 support for bpf/libpcap/tcpdump,
but that is aways down the road.
count drops to 0 in witness_destroy, set the w_name and w_file pointers
to point to the string "(dead)" and the w_line field to 0. This way,
if a mutex of a given name is used only in a module, then as long as
all mutexes in the module are destroyed when the module is unloaded,
witness will not maintain stale references to the mutex's name in the
module's data section causing a panic later on when the w_name or w_file
field's are examined.
1. Pick up MII/PHY support for Livengood copper part (10/100/1000) from
Parag Patel. It was a fairly complete but not quite platform independent
job.
2. Finish silly offset differences that LIVENGOOD vs. WISEMAN registers
have (so the !)$*!)$*!$ fiber LIVENGOOD now works too).
3. Ansify the source.
So- we now suppor tthe PRO1000F and PRO1000T adapters.
1. The offsets for some registers change in LIVENGOOD. Gratuitously.
2. Define LIVENGOOD and LIVENGOOD_CU part numbers. Add some more
specific LIVENGOOD defaults.
3. Add definitions for PHY support for the copper LIVENGOOD part
(10/100/1000).
Since pid's are not in the kernel address space, this doesn't conflict
with the funcionality of specifying an arbitrary frame pointer to the
trace command.
- If the first function of a backtrace maps to fork_trampoline, then this
is a newly fork'd process that has not been executed yet, so just print
out the first frame and then return for that case.
- Lower the default count from 65535 to 1024. ddb doesn't trace into
userland, and if the stack gets hosed and starts looping it's less
annoying.
Parag Patel did all of the grunt work, so he gets the credit.
Register definitions and actions inferred from a Linux driver,
so Intel also gets some 'credit'.
the main benefit this gives for now is that via686 audio devices on
motherboards with ac97 codecs that do not support vra will be able to use
sample rates other than 48khz.
Add simple "xlat" converter which performs 8to8 table based conversion.
Unicode converter will be added in the near future.
Reviewed by: silence on arch@
Files placement reviewed by: bde
Obtained from: smbfs
o Change the number of init tries from 5 to a #define.
o Allow up to 5s rather than 2s for commands to complete. This
is still much less than 51 minutes, but makes my intel card init
with more reliability than before.
possible for some systems where the device is there, but the BIOS
hasn't allocated memory resources for it), we don't panic.
Submitted by: Gerard Roudier
peer out from sppp_lcp_open() to sppp_lcp_up(). For one, this makes
things look more symmetrical to sppp_lcp_close(), and somehow it also
just occurred to me that an Up event following the open caused the
value of the authentication option to be clobbered.
badaddr_read(). This fixes 'machine check in pal mode' halts on
ev5 2100As.
MFC candidate -- after spending 6 hours tracking this down, I checked and
discovered that it has been in NetBSD for over a year, so it should be safe
for MFC into 4.3-RELEASE
It's not finished yet (I still have to find a way to implement process-
dependent nodes without consuming too much memory, and the permission
system needs tightening up), but it's becoming hard to work on without
a repo (I've accidentally almost nuked it once already), and it works
(except for the lack of process-dependent nodes, that is).
I was supposed to commit this a week ago, but timed out waiting for jkh
to reply to some questions I had. Pass him a spoonful of bad karma :)
