Tom Keiser [Mon, 30 May 2005 04:39:50 +0000]
rx-fpq-bulk-free-
20050529
FIXES 19027
After profiling RX for a while, I've found a few more bottlenecks in the
packet handling code. This patch addresses a couple of these issues.
The major change in this patch is a new API to allow bulk packet
alloc/free ops on rx_queue's of packets. Benefits include reduced lock
contention on rx_freePktQ_lock, elimination of a lot of unnecessary cache
line invalidates, and reduced register window thrashing on sparc.
In addition, this patch dedicates one rx_packet per thread to rxi_SendAck,
since that function is in the critical path, and represents a large
percentage of execution time.
Tracy Di Marco White [Mon, 30 May 2005 04:35:52 +0000]
netbsd-pam-
20050529
FIXES 18924
pam support for nbsd
Roman Mitz [Wed, 25 May 2005 00:12:52 +0000]
discard-rx-streams-
20050524
FIXES 18895
they wouldn't compile anyway. remove references also.
Klas Lindfors [Wed, 25 May 2005 00:06:43 +0000]
make-clearcallback-faster-
20050524
FIXES 4413
hash callbacks twice so we can discard them more quickly.
Derrick Brashear [Tue, 24 May 2005 21:32:36 +0000]
unroll-macos-104-
20050523
we'll finish it on a branch and remerge; 1.3.83 needs to happen and 1.4 needs to go out.
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
we'll finish it on a branch and remerge; 1.3.83 needs to happen and 1.4 needs to
go out.
Asanka Herath [Tue, 24 May 2005 06:12:19 +0000]
windows-loopback-
20050524
Find the loopback adapter installer by hardware ID instead of English name.
Jeffrey Altman [Mon, 23 May 2005 14:57:55 +0000]
windows-notes-
20050523
more fixes
Jeffrey Altman [Sun, 22 May 2005 06:53:44 +0000]
windows-delonclose-
20050522
Fix support for the CreateFile(DELETE_ON_CLOSE) option.
Jim Rees [Fri, 20 May 2005 20:26:53 +0000]
bsd-config-
20050520
config stuff for OpenBSD 3.7, FreeBSD 5.4 and 6.0
Jim Rees [Fri, 20 May 2005 20:19:26 +0000]
obsd-time-
20050520
Remove extern time from OBSD/osi_machdep.h. It's in sys/kernel.h.
Suggested by irene.braun@ualberta.ca.
Jeffrey Altman [Wed, 18 May 2005 23:29:32 +0000]
windows-version-
20050518
1.3.8300
Jeffrey Altman [Wed, 18 May 2005 23:27:19 +0000]
windows-notes-
20050518
1.3.8300
Jeffrey Altman [Wed, 18 May 2005 23:22:03 +0000]
windows-afssvrmgr-
20050518
Link resulting executable to kerberos 5
Jeffrey Altman [Wed, 18 May 2005 23:01:10 +0000]
rx-makecall-race-fix-
20050518
On at least one system it was noticed that threads waiting in rx_NewCall
would starve forever (aka deadlock). This was the result of one out of
two problems related to a race condition on the RX_CONN_MAKECALL_WAITING
bit flag. This flag was set once in rx_NewCall and cleared in rx_EndCall.
However, it was possible for the flag to be cleared even though there
were additional flags waiting in rx_NewCall. This was due to a failure
to check the value of makeCallWaiters before clearing the flag and also
due to a failure to properly lock the access to the makeCallWaiters field.
The second problem was an ability to destroy a connection on which threads
are waiting within rx_NewCall.
Chas Williams [Wed, 18 May 2005 14:52:24 +0000]
linux-str-inlines-
20050518
add #include <linux/string.h> in the param file after <linux/config.h>
fixes unresolved ref to strlen from xdr.o via inline
Jeffrey Altman [Tue, 17 May 2005 19:25:21 +0000]
windows-dnlc-corruption-
20050517
It has been noticed that on startup there may be corruption of the
dnlc freelist. In particular, some items on the list are actually
in use. This patch will cause the dnlc to be purged if corruption
of the freelist is detected. It also makes changes to ensure that
items placed onto the freelist are in fact cleaned to prevent confusion.
The purge function is corrected to set that magic field to the correct
value.
The validation function is fixed to not CloseHandle the same handle
twice.
Jim Rees [Sun, 15 May 2005 21:54:50 +0000]
fix-vrefcount-
20050513
fix VREFCOUNT macro for XBSD
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
and everything else.
Jim Rees [Sun, 15 May 2005 21:40:15 +0000]
freebsd60-readme-
20050515
Add FreeBSD 6.0.
Consolidate linux kernel header notes in Linux Notes section.
Reformat system types section.
Derrick Brashear [Fri, 13 May 2005 22:43:55 +0000]
macos104-not-yet-
20050513
rewrite attrs section a bit to deal with 10.4
Derrick Brashear [Fri, 13 May 2005 04:00:24 +0000]
macos104-not-yet-
20050512
virtualize vnode refcount "greater than" testing so we can use tiger primitive
Jim Rees [Thu, 12 May 2005 21:06:59 +0000]
fix-tuiop-
20050512
tuiop, not tuio
Chaskiel M Grundman [Thu, 12 May 2005 19:30:35 +0000]
macos104-not-yet-
20050511
further work
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
ifaddr/ifnet compat macros for older versions, and clean up afs_server interface ot use it.
fix vnode_isinuse macros
====================
further work
====================
further work
Jeffrey Altman [Wed, 11 May 2005 22:33:33 +0000]
windows-scache-release-
20050511
Add additional "if !NULL" checks before calling cm_SCacheRelease
Jeffrey Altman [Tue, 10 May 2005 04:34:03 +0000]
windows-minidump-
20050509
Enable support for Minidump generation and crash reporting
Jeffrey Altman [Tue, 10 May 2005 04:33:28 +0000]
windows-server-manager-
20050509
Enable KFW support for server manager
Jeffrey Altman [Tue, 10 May 2005 04:32:30 +0000]
windows-makefile-
20050509
reorganize dependencies to allow kfw to be integrated into Server Manager
Jeffrey Altman [Tue, 10 May 2005 04:31:56 +0000]
windows-notes-
20050509
latest updates
Chaskiel M Grundman [Sun, 8 May 2005 07:49:43 +0000]
macos-104-not-yet-
20050508
shadow@dementia.org to blame for some of this.
Klas Lindfors [Sun, 8 May 2005 07:16:45 +0000]
export-calleraccess-via-pioctl-
20050508
FIXES 18442
provide pioctl to get CallerAccess bits
Chaskiel M Grundman [Sun, 8 May 2005 07:09:19 +0000]
volinfo-fixes-
20050508
FIXES 18391
large (>31bit) volume id handling and fixed handling of cloned volumes
Christopher Allen Wing [Sun, 8 May 2005 07:05:09 +0000]
linux-proc-read-seq-file-
20050508
FIXES 18612
use sequenced file interface for proc copy of CellServDB
Tom Keiser [Sun, 8 May 2005 06:59:58 +0000]
fix-64bit-assign-int64-
20050508
FIXES 18611
make this match the way the 32 bit (struct lo/hi) version works
Marc Horowitz [Sun, 8 May 2005 06:55:45 +0000]
sparc64-linux26-
20050508
FIXES 18485
add sparc64-linux24 port
Steve Roseman [Sun, 8 May 2005 06:50:23 +0000]
aix-auth-speedup-
20050508
FIXES 18467
The security module "afs_dynamic_auth" can get VERY slow if trying to
authenticate against a user in a large NIS password file. This happens
because it "replaces" getpwnam with code which loops through the
"passwd" file.
this 1) makes external programs use the "real"
getpwnam, and 2) changes the internal routine afs_getpwnam_int to use
getpwnam instead of iterating.
Chaskiel M Grundman [Sun, 8 May 2005 06:46:14 +0000]
linux-copyin-copyout-dont-check-errors-
20050508
FIXES 18293
improved macros which do error checking
Chas Williams [Sun, 8 May 2005 06:38:15 +0000]
d-revalidate-redux-
20050508
FIXES 18588
when the callback for a directory is broken the children of this vnode
are not invalidated as well. so VerifyVCache() always believes these
vnodes to be up to date since they will still have the CStatd flag.
perhaps breaking the callback on a directory should also clear CStatd
on children vnodes?
this isnt a problem for other operating system which go through the
dnlc which does purge the children entries for the vnode in question
during ClearCallBack().
since d_revalidate now avoids the dnlc (and afs_lookup) the following
patch tracks the parent vnode's DataVersion (just the lower half really
but this should be sufficient) in d_time. if the parent has changed
then the dentry is flagged as neededing updated.
Derrick Brashear [Mon, 2 May 2005 03:04:45 +0000]
macos104-not-yet-
20050501
push first group of 10.4 changes
Jim Rees [Fri, 29 Apr 2005 22:55:58 +0000]
freebsd60-mount-
20050429
Fix new FreeBSD mount vfs op.
Client works now, at least for a little while.
Jeffrey Altman [Fri, 29 Apr 2005 21:00:10 +0000]
windows-version-
20050429
1.3.8201
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
1.3.8201 (do it right this time)
Jeffrey Altman [Fri, 29 Apr 2005 20:51:18 +0000]
windows-fix-kfw-dependency-
20050429
a failure to install KFW would lead to crashes
Jeffrey Altman [Fri, 29 Apr 2005 20:49:58 +0000]
windows-notes-
20050429
1.3.8201
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
remove AFS Gateway option
Jim Rees [Fri, 29 Apr 2005 19:24:01 +0000]
freebsd-
20050429
Move module init/uninit code from module event handler to vfs ops
Use VFS_SET for FreeBSD 6.0
Jim Rees [Fri, 29 Apr 2005 18:01:08 +0000]
getpid-
20050429
rename our getpid() to osi_getpid() to avoid name collision with kernel version
Derrick Brashear [Fri, 29 Apr 2005 02:48:21 +0000]
macos-osi-flushpages-
20050428
reenable osi_FlushPages
Derrick Brashear [Thu, 28 Apr 2005 23:14:57 +0000]
split-dcache-
20050427
split for rw/ro dcaches to allow portion of cache to be dedicated
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
update
Jeffrey Altman [Thu, 28 Apr 2005 19:12:43 +0000]
windows-version-
20050428
1.3.8200
Derrick Brashear [Thu, 28 Apr 2005 06:02:19 +0000]
vnode-nusers-avoid-potential-overflow-
20050428
a byte is very small...
Chaskiel M Grundman [Thu, 28 Apr 2005 04:10:19 +0000]
osi-initcacheinfo-cleanup-
20050427
FIXES 18429
remove extraneous variable, be consistent about cacheMnt type
Peter Somogyi [Thu, 28 Apr 2005 04:04:13 +0000]
change-s390x-linker-binding-
20050427
FIXES 18447
to avoid binding against xdr_int_internal in xdr use from the pam module
Chaskiel M Grundman [Wed, 27 Apr 2005 22:43:18 +0000]
linux-always-glock-
20050427
finish this up, always glock on linux
Jeffrey Altman [Wed, 27 Apr 2005 16:32:22 +0000]
windows-callback-race-
20050427
cm_EndCallbackGrantingCall contained a race condition due to the release
of the cm_callbackLock in the middle of the for() loop. The race was
removed by optimizing out the call to cm_CallbackNotifyChange(). There
is no reason this needed to be called once per callback revoke in the
list.
Jeffrey Altman [Wed, 27 Apr 2005 16:28:00 +0000]
winnotes-
20050427
more updates
Derrick Brashear [Wed, 27 Apr 2005 02:49:47 +0000]
add-ptprocs-prototype-use-
20050426
thanks to Chris Wing for pointing this out
Derrick Brashear [Wed, 27 Apr 2005 02:36:40 +0000]
daemonize-log-somewhere-so-fds-can-close-without-harm-
20050426
FIXES 18405
switch to VLog from printf
Tom Keiser [Mon, 25 Apr 2005 21:52:59 +0000]
rx-fpq-optimize-
20050425
FIXES 17805
here's a patch that reduces the overhead of transfers
between the local and global free packet queues. The old algorithm was
O(n) in the number of store instructions -- 7 per rx_packet. I've added
some bulk transfer macros to the rx_queue package. Now, the number of
store instructions is O(1) -- 6 total. This should help reduce bus
contention and cache line invalidates on SMPs.
Derrick Brashear [Mon, 25 Apr 2005 18:20:16 +0000]
linux-global-sunlock-always-
20050424
ok, well, in the land of pre-emption i guess we need this always
Chas Williams [Mon, 25 Apr 2005 15:54:54 +0000]
linux-dentry-revalidate-one-more-
20050424
further updates
Chaskiel M Grundman [Sun, 24 Apr 2005 15:54:03 +0000]
add-missing-inotify-test-
20050424
so we actually detect the inotify sem
Chas Williams [Sun, 24 Apr 2005 15:26:53 +0000]
dentry-revalidate-cleanup-
20050424
clean up dentry_revalidate one more time so performance improves
Niklas Edmundsson [Sun, 24 Apr 2005 15:24:57 +0000]
aix53-
20050424
add param file for aix 53
check for cache partition type
Rainer Toebbicke [Sun, 24 Apr 2005 15:19:55 +0000]
maintain-volume-filecounts-
20050424
maintain file counts in volume header
Christopher Allen Wing [Sun, 24 Apr 2005 15:16:21 +0000]
use-daemon-in-bosserver-and-afsd-
20050424
provide daemon() if none exists. call it in bosserver and in afsd for afsdb handler
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
add the file, too
Chas Williams [Sun, 24 Apr 2005 15:01:29 +0000]
avoid-leaking-stat-info-
20050424
otherwise we are effectively giving stat powers to people who shouldn't have it
Chaskiel M Grundman [Sun, 24 Apr 2005 02:02:52 +0000]
osi-file-use-dentry-open-
20050423
FIXES 18335
use dentry_open et al instead of doing the work ourselves
Derrick Brashear [Sun, 24 Apr 2005 01:54:42 +0000]
linux-osi-module-proc-csdb-sprintf-bug-
20050423
FIXES 18334
remove extraneous %s
Jim Rees [Fri, 22 Apr 2005 15:57:09 +0000]
freebsd60-
20050422
Preliminary FreeBSD 6.0 support.
Builds, but unlikely to work.
Jim Rees [Thu, 21 Apr 2005 23:08:12 +0000]
ignore-ytab-
20050421
add y.tab.h to .cvsignore
Michael Niksch [Thu, 21 Apr 2005 13:03:43 +0000]
restore-tab-
20050421
Restore lost tab.
Jeffrey Altman [Thu, 21 Apr 2005 03:16:57 +0000]
windows-shell-ext-remove-mtpt-
20050420
Conditionalize the AFS->Mount Point->Remove menu item based upon whether
or not the selected item(s) contain mount points
Tom Keiser [Wed, 20 Apr 2005 22:21:16 +0000]
rx_fpq_take_three-
20050420
FIXES 17805
I've been stress testing a patch all weekend that changes the way
thread-local packet quotas are computed. I was able to replicate the
unbounded packet alloc problem on unix server components by eliminating my
code from rxi_StartServerProcs that sets the maximum number of expected
threads. This patch makes the upper thread limit get computed on the fly,
adds some thread-local free packet queue statistics, and a few other minor
tweaks. I still don't have a working windows development environment, so
I can't say whether this will fix all the windows client problems. But,
the unbounded packet allocation problem should go away with this patch.
stress testing on windows succeeds as well. <jaltman@secure-endpoints.com>
Jim Rees [Wed, 20 Apr 2005 17:35:44 +0000]
fix-vc-v-macro-
20050420
Fix IsAfsVnode and friends for OpenBSD and NetBSD.
Use "v" instead of "vc" for vnode args to macros (cosmetic only).
Redefine OpenBSD macros instead of using #ifdefs in afs_osi.h.
Jeffrey Altman [Tue, 19 Apr 2005 22:06:22 +0000]
windows-hookdll-
20050419
Some more tweaks for hook dll processing.
Chas Williams [Tue, 19 Apr 2005 17:27:25 +0000]
linux-conflate-rdwrs-
20050419
FIXES 17561
conflate the 2 rdwr routines we have
Niklas Edmundsson [Tue, 19 Apr 2005 16:50:25 +0000]
aix53-
20050419
add aix 5.3 support
Jeffrey Altman [Tue, 19 Apr 2005 07:26:27 +0000]
windows-hooks-
20050418
Add new hooks from "AfsdHook.dll" to the afsd_service.exe
AfsdInitHook - where you currently have one
AfsdRxStartedHook - after RxInit but before SmbInit
AfsdSmbStartedHook - after SmbInit
AfsdBkDaemonHook - called by the BkDaemon thread and is executed once per
cycle
AfsdStoppingHook - called just after a shutdown event has been
received but before any shutdown has been performed
AfsdStoppedHook - called just after all shutdown operations after
completed but before the service terminates
The type of the function is
BOOL ( APIENTRY * AfsdHook )(void)
in all cases. Returning FALSE will cause the service to shutdown.
Jeffrey Altman [Tue, 19 Apr 2005 06:04:35 +0000]
windows-O_NOBLOCK-
20050418
O_NOBLOCK is not defined on Windows
Jeffrey Altman [Tue, 19 Apr 2005 06:04:03 +0000]
windows-rx-ts-fpq-
20050418
FIXES 17805
Add a new function called rx_StartClientThread() which parallels the
behavior of rx_StartServer but is used for initializing client threads.
====================
This delta was composed from multiple commits as part of the CVS->Git migration.
The checkin message with each commit was inconsistent.
The following are the additional commit messages.
====================
FIXES 17805
Add a new function called rx_StartClientThread() which parallels the
behavior of rx_StartServer but is used for initializing client threads.
Also, disable RX_TSFPQ_ENABLE for Windows as it slows the client down.
Jeffrey Altman [Tue, 19 Apr 2005 06:03:10 +0000]
windows-formatting-
20050418
keep formatting consistent
Jeffrey Altman [Tue, 19 Apr 2005 06:02:48 +0000]
windows-free-lib-
20050418
call FreeLibrary not CloseHandle
Derrick Brashear [Mon, 18 Apr 2005 21:09:38 +0000]
macos104-not-yet-
20050418
push first group of 10.4 changes
Jeffrey Altman [Mon, 18 Apr 2005 15:43:50 +0000]
windows-syscfg-
20050418
GetAdaptersAddressess -> GetAdaptersAddresses
Chaskiel M Grundman [Mon, 18 Apr 2005 15:30:53 +0000]
linux-fmode-set-
20050418
to avoid problems where we don't get truncated
Rainer Toebbicke [Mon, 18 Apr 2005 15:28:05 +0000]
logging-in-linux-vnodeops-sucks-
20050418
given that we keep ending up with these weird errors, it's just not worth it.
Robert Banz [Mon, 18 Apr 2005 15:25:34 +0000]
solaris-largefile-not-on-ia32-
20050417
disable largefile on solaris x86 32 bit
Jeffrey Altman [Mon, 18 Apr 2005 03:28:31 +0000]
missing-eof-
20050417
add missing eof
Derrick Brashear [Fri, 15 Apr 2005 20:39:33 +0000]
ptserver-restricted-mode-
20050415
add restricted mode to pts
only admins can make changes
Derrick Brashear [Fri, 15 Apr 2005 20:32:31 +0000]
make-linux-proc-cellservdb-info-
20050415
use info hook instead of read so we can have larger than a slab
Derrick Brashear [Fri, 15 Apr 2005 20:04:55 +0000]
license-compromise-
20050415
just copy the full list of licenses here. bleh.
Derrick Brashear [Fri, 15 Apr 2005 20:02:08 +0000]
license-update-
20050415
FIXES 18131
collect all licenses here
Kris Van Hees [Fri, 15 Apr 2005 19:43:57 +0000]
vos-release-compare-update-timestamps-
20050415
FIXES 17758
use last updated timestamps to determine if a release needs to be done
Peter Somogyi [Fri, 15 Apr 2005 19:34:30 +0000]
fix-rxkad-prototype-
20050415
FIXES 18079
fix cbc_encrypt prototype
Derrick Brashear [Fri, 15 Apr 2005 19:33:22 +0000]
kill-struct-osisocket-
20050415
FIXES 18079
more complete version of what's in the patch, kill all refs to struct osi_socket in favor of just osi_socket
from psomogyi@gamax.hu
Chas Williams [Fri, 15 Apr 2005 19:23:06 +0000]
viced-already-regd-warning-
20050415
FIXES 18049
the ethernet address is irrelevant
Jeffrey Altman [Fri, 15 Apr 2005 17:01:40 +0000]
windows-locking-freeacls-
20050415
Ensure that the appropriate mutex locks are held on the stat cache
entry referred to by the acl back pointer when performing a cleanup
Jeffrey Altman [Fri, 15 Apr 2005 17:00:15 +0000]
windows-memmap-adjust-
20050415
Prepare to allow the maxVolumes and maxCells parameters to become
independent of the MaxStats value
Chas Williams [Fri, 15 Apr 2005 14:40:31 +0000]
sema-init-takes-2-args-
20050415
FIXES 18271
init sema correctly
Derrick Brashear [Fri, 15 Apr 2005 14:31:42 +0000]
fifo-logs-
20050415
use O_NONBLOCK during open if it's a fifo
suppress SIGPIPE in case we care (EPIPE handling still not needed: return value
of write is discarded)
Jeffrey Altman [Thu, 14 Apr 2005 06:46:34 +0000]
windows-winlogon-logon-event-
20050414
Apparently the problem with multi-domain forests with cross-
realm trusts to non-Windows realms was not entirely solved.
The authentication to the AFS SMB service failed because
the wrong name was being used. Using ASU as an example,
the authentication was being performed with the name
"QAAD\user" (an account in the forest root) and not
"user@ASU.EDU (the MIT Kerberos principal used to login with)
The solution was to add an additional dependency on KFW
in order or to be able to easily obtain the client principal
name stored in the MSLSA ccache TGT. This information is
used in two locations:
- the pioctl() function
- a new WinLogon Event Handler for the "logon" event.
The pioctl function will now be able to use the correct
name when calling WNetAddConnection2() and the "logon"
event handler will now be able to call WNetAddConnection2().
The hope is that the "logon" event handler will be called
before the profile is loaded but I have not guarrantee
that will happen.
Derrick Brashear [Thu, 14 Apr 2005 05:41:59 +0000]
support-named-pipe-logs-
20050414
so log files can be named pipes
Derrick Brashear [Thu, 14 Apr 2005 05:29:34 +0000]
renamefile-for-mrafslogs-
20050414
call the renamefile wrapper so mrafs style logs on windows work like other logs