openafs.git
23 years agoinitial-freebsd-port-work-20010414
Tom Maher [Sat, 14 Apr 2001 18:27:12 +0000]
initial-freebsd-port-work-20010414

start of work on freebsd port

23 years agolinux-inode-and-dentry-validation-rework-20010413
Chas Williams [Fri, 13 Apr 2001 22:56:00 +0000]
linux-inode-and-dentry-validation-rework-20010413

gets us closer to the ideal in inode and dentry revalidation

23 years agomake-use-of-int-preblock-specific-to-s390-linux-20010412
Derrick Brashear [Thu, 12 Apr 2001 21:39:33 +0000]
make-use-of-int-preblock-specific-to-s390-linux-20010412

based on report Pavel Semerad <semerad@ss1000.ms.mff.cuni.cz>, but fixing the
root cause.

23 years agobetter-type-guessing-for-readdir-hint-20010411
Chaskiel M Grundman [Wed, 11 Apr 2001 20:17:51 +0000]
better-type-guessing-for-readdir-hint-20010411

instead of just saying unknown, if we know for sure, provide the info

23 years agowindows-compile-fixes-20010411
Volker Holfeld [Wed, 11 Apr 2001 19:57:23 +0000]
windows-compile-fixes-20010411

ignore serverLog on nt

====================
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.
====================

correct typo for nt

====================

update volinfo -filenames feature for nt

23 years agorx-keep-track-of-resent-packets-20010406
Nickolai Zeldovich [Fri, 6 Apr 2001 19:22:01 +0000]
rx-keep-track-of-resent-packets-20010406

previously the resent packet counter was never incremented

23 years agodont-fflush-fds-20010406
Nickolai Zeldovich [Fri, 6 Apr 2001 19:13:19 +0000]
dont-fflush-fds-20010406

bogus call to fflush in server log code, trying to fflush a file descriptor

23 years agopartition-include-headers-to-get-opendir-etc-20010406
Nickolai Zeldovich [Fri, 6 Apr 2001 19:11:44 +0000]
partition-include-headers-to-get-opendir-etc-20010406

so namei code comes closer to being useful in more cases

23 years agolinux24-new-afs-xsetgroups32-implementation-20010406
Chaskiel M Grundman [Fri, 6 Apr 2001 19:08:09 +0000]
linux24-new-afs-xsetgroups32-implementation-20010406

afs_xsetgroups32 is now a copy of afs_xsetgroups, with the call to
sys_setgroupsp changed to sys_setgroups32p

no change to sparc64 port stuff

23 years agodarwin-pagincred-will-notice-pags-20010406
Chaskiel M Grundman [Fri, 6 Apr 2001 19:03:46 +0000]
darwin-pagincred-will-notice-pags-20010406

make PagInCred notice pags on darwin

23 years agoremove-endif-warning-20010405
Nathan Neulinger [Thu, 5 Apr 2001 20:32:41 +0000]
remove-endif-warning-20010405

endif comment should be commentized

23 years agouse-consistent-data-typing-for-hosts-20010404
Nathan Neulinger [Wed, 4 Apr 2001 21:36:19 +0000]
use-consistent-data-typing-for-hosts-20010404

use afs_uint32 everywhere

23 years agoplacate-gmake-on-solaris-when-building-libafs-20010404
Nathan Neulinger [Wed, 4 Apr 2001 21:29:31 +0000]
placate-gmake-on-solaris-when-building-libafs-20010404

gmake tries to build something named libafs, make it not

23 years agovfsck-doesnt-build-on-hpux11-20010404
Nathan Neulinger [Wed, 4 Apr 2001 21:27:43 +0000]
vfsck-doesnt-build-on-hpux11-20010404

previous revisions caused it to be attempted to be built

23 years agomake-getfreetag-find-correct-free-tags-20010404
Hartmut Reuter [Wed, 4 Apr 2001 18:28:05 +0000]
make-getfreetag-find-correct-free-tags-20010404

"I have found a "nice" bug in namei_ops.c:

In GetFreeTag() the following code was used to find a free place in the
link table entry:

    /* Now find a free column in this row and claim it. */
    coldata = 0x7;
    for (col = 0; col<NAMEI_MAXVOLS; col++) {
        coldata <<= col * 3;
        if ((row & coldata) == 0)
            break;
    }

What happens?
With col == 0 we look for the bits 0x7. This is ok.
With col == 1 we look for the bits 0x38. This is also ok.
With col == 2 we look for the bits 0xe00 instead of 0x1c0 !!!
But later the allocation is done correctly with 0x40.
Thus the next time he will find again a free place for col == 2 !
For higher values of col the test is also incorrect, of course."

23 years agoremove-redundant-makefile-rules-20010404
Nathan Neulinger [Wed, 4 Apr 2001 18:24:02 +0000]
remove-redundant-makefile-rules-20010404

another set of redundant makefile rules removed

23 years agochanges-to-allow-identifying-threads-in-log-file-output-20010404
Hartmut Reuter [Wed, 4 Apr 2001 18:21:53 +0000]
changes-to-allow-identifying-threads-in-log-file-output-20010404

"This is an addition for MR-AFS log files where the thread
name is shown. rx switches the general server threads to listeners and
vice versa. To have reasonable names the listener must get back his old
name when converting himself to a server thread. To not bother other
programs with unresolved references I do this with two function pointers
which are initialized in the main() programs for MR-AFS."

====================
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.
====================

use function pointers here instead of threadname

23 years agomake-vol-package-deal-with-namei-correctly-on-systems-which-can-be-compiled-with...
Hartmut Reuter [Wed, 4 Apr 2001 18:14:23 +0000]
make-vol-package-deal-with-namei-correctly-on-systems-which-can-be-compiled-with-or-without-namei-fileserver-20010404

If compiled without AFS_NAMEI_ENV existence
of AFSIDat must stop us. If compiled with AFS_NAMEI_ENV volume headers
in a partition, but no AFSIDat must stop us.

23 years agosalvager-syslog-support-20010404
Nathan Neulinger [Wed, 4 Apr 2001 18:12:18 +0000]
salvager-syslog-support-20010404

support for salvager to syslog instead of logging to file

23 years agopartial-hpux-10-support-20010404
Nathan Neulinger [Wed, 4 Apr 2001 18:08:51 +0000]
partial-hpux-10-support-20010404

similar to HPUX 11, just make what can be compiled now compile

23 years agosimplify-salvager-target-20010404
Nathan Neulinger [Wed, 4 Apr 2001 18:00:44 +0000]
simplify-salvager-target-20010404

it's the same for linux as for everyone else now

23 years agoenable-getcrypt-command-20010404
Derrick Brashear [Wed, 4 Apr 2001 13:48:23 +0000]
enable-getcrypt-command-20010404

provide cmd hook for getcrypt command (support was already present)

23 years agowinnt-enable-cryptall-support-20010404
Volker Holfeld [Wed, 4 Apr 2001 13:46:19 +0000]
winnt-enable-cryptall-support-20010404

introduce fs setcrypt/fs getcrypt features to nt client
add registry parameter SecurityLevel to allow persistent setting

23 years agowinnt-create-globalautomapper-registry-key-20010404
Volker Holfeld [Wed, 4 Apr 2001 13:43:45 +0000]
winnt-create-globalautomapper-registry-key-20010404

if GlobalAutoMapper key is missing when a global drive is defined with the
gui it will be created

23 years agocorrect-udebug-flag-name-20010403
Nathan Neulinger [Tue, 3 Apr 2001 21:26:31 +0000]
correct-udebug-flag-name-20010403

-server takes one server arg, so fix its name to be -server and not -servers

23 years agofix-ubik-lowest-host-comparison-algorithm-20010403
Nathan Neulinger [Tue, 3 Apr 2001 21:22:05 +0000]
fix-ubik-lowest-host-comparison-algorithm-20010403

be more particular about how we do comparison, and fix printing of the host
address for byte order problem

23 years agoprefer-ubik-print-to-printf-20010403
Nathan Neulinger [Tue, 3 Apr 2001 21:19:34 +0000]
prefer-ubik-print-to-printf-20010403

changes to prefer ubik_print &c to printf to make syslogging easier

23 years agolink-libs-into-windows-butc-in-useful-order-20010403
Volker Holfeld [Tue, 3 Apr 2001 20:06:24 +0000]
link-libs-into-windows-butc-in-useful-order-20010403

reorder libs so threaded afsrpc.lib is before afsrx.lib

23 years agoi386-linux24-includes-long-long-types-20010403
Derrick Brashear [Tue, 3 Apr 2001 07:55:34 +0000]
i386-linux24-includes-long-long-types-20010403

define AFS_64BIT_ENV for linux24 i386
basically denotes existance of long long

23 years agoattempt-more-portable-64bit-type-support-for-namei-fileserver-20010403
Derrick Brashear [Tue, 3 Apr 2001 07:52:56 +0000]
attempt-more-portable-64bit-type-support-for-namei-fileserver-20010403

be at least as portable as existing code

23 years agoadd-volinfo-filenames-arg-in-namei-environment-20010402
Thomas Mueller [Tue, 3 Apr 2001 00:13:04 +0000]
add-volinfo-filenames-arg-in-namei-environment-20010402

show filenames used by volumes in namei fileserver

23 years agolinux24-use-kmap-instead-of-afs-linux-page-address-20010402
Chas Williams [Tue, 3 Apr 2001 00:09:51 +0000]
linux24-use-kmap-instead-of-afs-linux-page-address-20010402

kmap is the preferred interface for 2.4 now

23 years agolinux-dentry-validation-changes-20010402
Chas Williams [Mon, 2 Apr 2001 23:48:22 +0000]
linux-dentry-validation-changes-20010402

change afs_linux_dentry_revalidate to call afs_linux_revalidate

23 years agos390-updates-20010330
Adam Thornton [Sat, 31 Mar 2001 00:20:23 +0000]
s390-updates-20010330

fix PRE_Block and lwp problems

23 years agoadd-missing-exports-for-64bit-types-for-windows-20010329
Volker Holfeld [Fri, 30 Mar 2001 05:10:59 +0000]
add-missing-exports-for-64bit-types-for-windows-20010329

23 years agodrivemap-box-error-for-3-character-cellnames-20010329
Silvia Gross [Fri, 30 Mar 2001 00:18:55 +0000]
drivemap-box-error-for-3-character-cellnames-20010329

"In the original version the path name in the box showing the network
drives were wrong if a part of the name after afs had only 3 characters:
afs\ipp\u became afs\u. Also the small hooks in front of the names
disappeared immediately."

23 years agofix-readdir-hint-20010329
Sam Hartman [Fri, 30 Mar 2001 00:12:38 +0000]
fix-readdir-hint-20010329

"Under 2.4.x kernels, there is a file type in the result of readdir to
help applications optimize directory traversal.  With usfficiently new
libc compiled against sufficiently new kernel headers, this is exported to
applications.

Several applications, including fileutils use this information.  AFS
asserts that all files are directories.  This sucks, because among
other things it causes rm -rf not to work."

23 years agos390-libafs-makefile-changes-20010329
Michael Durket [Thu, 29 Mar 2001 23:56:18 +0000]
s390-libafs-makefile-changes-20010329

use -O and not -O2 to deal with some bad versions of gcc for now

link in appropriate asm directory

23 years agolinux24-setgroups32-support-20010329
Thomas Mueller [Thu, 29 Mar 2001 18:07:41 +0000]
linux24-setgroups32-support-20010329

setgroups32 support for linux2.4 machines (required for modern glibc)

23 years agowinnt-bufio-mods-20010328
David Bailey [Wed, 28 Mar 2001 17:54:21 +0000]
winnt-bufio-mods-20010328

fix warnings in fileutil.c

23 years agolinux-s390-page-offset-fix-20010328
Carsten Jacobi [Wed, 28 Mar 2001 17:50:19 +0000]
linux-s390-page-offset-fix-20010328

use PAGE_OFFSET for s390

23 years agoubik-clone-support-20010212
Hartmut Reuter [Wed, 28 Mar 2001 07:51:25 +0000]
ubik-clone-support-20010212

ubik library support for readonly clones

23 years agofirst-cut-autoconf-support-20010327
Johan Danielsson [Wed, 28 Mar 2001 00:01:45 +0000]
first-cut-autoconf-support-20010327

Makefile.am and configure.in first cuts. not yet useful

====================
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.
====================

oops; this belongs on the branch created for autoconf development

23 years agodarwin-cacheops-initialization-check-20010327 BP-openafs-devel-autoconf
Chaskiel M Grundman [Tue, 27 Mar 2001 23:22:05 +0000]
darwin-cacheops-initialization-check-20010327

make sure nothing happens before cacheops init

23 years agodarwin port was unhappy because of this
Chaskiel M Grundman [Tue, 27 Mar 2001 23:18:09 +0000]
darwin port was unhappy because of this

23 years agoremove-libadmin-real-dependancies-20010327
Derrick Brashear [Tue, 27 Mar 2001 23:10:13 +0000]
remove-libadmin-real-dependancies-20010327

23 years agodeath-to-permit-xprt-h-20010327
Derrick Brashear [Tue, 27 Mar 2001 23:02:13 +0000]
death-to-permit-xprt-h-20010327

permit_xprt.h, we hardly knew ye

23 years agobegone-with-xprt-CoerceLevel-20010327
Derrick Brashear [Tue, 27 Mar 2001 22:31:02 +0000]
begone-with-xprt-CoerceLevel-20010327

coersion is a crime. last step in the death of permit_xprt.h

23 years agobegone-with-xprt-cryptok-20010327
Derrick Brashear [Tue, 27 Mar 2001 22:28:17 +0000]
begone-with-xprt-cryptok-20010327

de-cripple rxkad-crypt for all cases

23 years agobegone-with-AFS-HIDE-20010327
Derrick Brashear [Tue, 27 Mar 2001 22:18:26 +0000]
begone-with-AFS-HIDE-20010327

start of the death of permit_xprt.h

23 years agoallow-building-subset-of-afs-20010327
Chaskiel M Grundman [Tue, 27 Mar 2001 21:33:17 +0000]
allow-building-subset-of-afs-20010327

allow overriding finale with e.g. blibafs to build a subset of afs

23 years agowinnt-lacks-snprintf-so-build-ours-20010327
Derrick Brashear [Tue, 27 Mar 2001 21:05:20 +0000]
winnt-lacks-snprintf-so-build-ours-20010327

apparently windows lacks snprintf

23 years agodont-install-uklibs-unless-linktest-works-20010327
Chaskiel M Grundman [Tue, 27 Mar 2001 20:37:56 +0000]
dont-install-uklibs-unless-linktest-works-20010327

follow the intent of this target

23 years agodont-build-clients-for-hpux11-20010327
Nathan Neulinger [Tue, 27 Mar 2001 20:33:52 +0000]
dont-build-clients-for-hpux11-20010327

build tools and servers but no kernel support for hpux since it doesn't work
currently

23 years agoinstall-util-archive-also-as-libafsutil-20010327
Derrick Brashear [Tue, 27 Mar 2001 20:29:04 +0000]
install-util-archive-also-as-libafsutil-20010327

It's about time you can link to this with -lsomething

23 years agoupdate-libuafs-install-procedure-20010327
Derrick Brashear [Tue, 27 Mar 2001 09:57:15 +0000]
update-libuafs-install-procedure-20010327

to deal with tweaks make so pinstall can be left uninstalled on darwin port

23 years agoadd-pts-global-reader-support-20010327
Derrick Brashear [Tue, 27 Mar 2001 09:49:41 +0000]
add-pts-global-reader-support-20010327

in the same way members of system:administrators have all pts powers,
members of system:ptsviewers have all pts read powers by virtue of
said membership

23 years agorudimentary-support-in-kpasswd-and-kas-for-mit-v4-string-to-keyed-passwords-20010327
Derrick Brashear [Tue, 27 Mar 2001 09:36:15 +0000]
rudimentary-support-in-kpasswd-and-kas-for-mit-v4-string-to-keyed-passwords-20010327

klog supports such passwords but without this such users can't run kas or kpasswd

23 years agouse-afs-int64-types-in-util-package-20010327
Derrick Brashear [Tue, 27 Mar 2001 09:16:43 +0000]
use-afs-int64-types-in-util-package-20010327

we already defined "portable" types for this; use them

23 years agomake-namei-fileserver-compile-on-solaris8-20010327
Nickolai Zeldovich [Tue, 27 Mar 2001 08:58:49 +0000]
make-namei-fileserver-compile-on-solaris8-20010327

use uint64_t in preference to u_int64_t as solaris has only the latter and
linux appears to have both. also, use lockf where flock was used in the
solaris case

23 years agoinitial-darwin-support-20010327
Chaskiel M Grundman [Tue, 27 Mar 2001 08:43:13 +0000]
initial-darwin-support-20010327

setup for darwin port

====================
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.
====================
changes for darwin port

====================

added files for the darwin port

====================
changes for darwin port

23 years agorxdebug-changes-for-peers-and-to-show-rtt-values-20010326
Nickolai Zeldovich [Tue, 27 Mar 2001 01:53:27 +0000]
rxdebug-changes-for-peers-and-to-show-rtt-values-20010326

"   rxdebug: allow the use of -onlyhost and -onlyport options with
    -peers, and also list the rtt, rttdev and retransmit timeout
    values."

23 years agofix-rx-rtt-computation-when-delay-is-involved-20010326
Nickolai Zeldovich [Tue, 27 Mar 2001 01:51:35 +0000]
fix-rx-rtt-computation-when-delay-is-involved-20010326

"   rx: attempt to fix incorrect rtt computation, possibly winning
    on performance for links with some packet loss.

    Currently, the rtt is computed as the time from the transmission
    of a packet until the receipt of the corresponding ack packet.
    But this includes a potential delay (rx_{soft,hard,last}AckDelay)
    on the client which results in an abnormally high rtt, and even
    more erroneous rtt_dev and retransmit timeout values.

    The proposed solution is to ignore RX_ACK_DELAY type ack packets
    for the purposes of rtt computations, and introduce another ack
    packet type for RX_ACK_DELAY-type acks which are sent immediately
    by the client.

    This should of course be backwards-compatible with existing clients.
"

23 years agofirst-pass-at-alpha-linux-support-20010312
Michael Stroucken [Tue, 27 Mar 2001 00:53:26 +0000]
first-pass-at-alpha-linux-support-20010312

Changes for Alpha to previous 64 bit changes for sparc (sparc64 isn't
necessarily analogous to alpha or any other 64 bit linux platform)

====================
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.
====================
use mkstemp

====================
Include necessary file for alpha config

====================
Deal with 64 bit time_t

====================
Don't do increments in a way that can be miscompiled

====================
configure for alpha linux

====================
deal with time and initialize a variable

====================
don't use ViceIoctl32 for alpha linux

====================
deal with compiling kdump for alpha linux

====================
mods in callback.c analogous to dux port; mods in viced to deal with
thread creation function return convention

====================
config for linux alpha

====================
ids for alpha linux consistent with new-style sysname grouping

====================
makefile and config files for new-style sysnames

====================
MakefileProto changes for alpha linux with new-style sysnames. A change
in config/mc.c will be needed for this to configure correctly, and it will
be added soon

====================
note alpha port and the state it's in to hopefully avoid questions

23 years agoadd-server-syslog-logging-capabilities-20010326
Nathan Neulinger [Mon, 26 Mar 2001 20:26:17 +0000]
add-server-syslog-logging-capabilities-20010326

Add support for syslogging server log messages

====================
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 syslogging capability to ptserver

====================
add syslog capability to fileserver

====================
Add syslog capability to the volserver

====================
Add syslog capability to the vlserver

====================
fix my own error when hand-applying nathan's patch

====================
remove unused function

23 years agomake-fileserver-deal-correctly-with-client-which-changed-addresses-but-was-offline...
Derek Atkins [Mon, 26 Mar 2001 17:51:34 +0000]
make-fileserver-deal-correctly-with-client-which-changed-addresses-but-was-offline-when-a-callback-to-it-was-revoked-20010326

"First, some background: The AFS Fileserver tries really hard to keep
track of all the "interfaces" of a client.  Generally this is for a
multi-homed client, so that the server realizes that you are the same
client when you come from multiple addresses.  However, this also
winds up applying to a mobile host whose IP address changes over time.

When the Fileserver sees a "new" address, it asks the client for its
Uuid and, if that Uuid already exists, it adds this new address to the
list of interfaces for the existing host.  However, it keeps a
callback connection open to the original address.

Here's the problem: Assume the client has callbacks registered with
the server and then disappears from the network.  While the client is
off the net, someone else makes a change that causes that callback to
be broken.  The fileserver can't reach the client, so the break gets
added to the delayed callback list.  The logic is such that no client
requests will be processed by a host while there are outstanding
delayed callbacks to that client.

Now, if the client comes back on the same IP Address, everything works
fine.  The fileserver uses the cached callback connection and the
callbacks are cleared successfully.  However, if the client returns to
the network under a different address, this new address is added to
the existing host structure and then the delayed callbacks are
attempted.  Unfortunately it is using the (invalid) cached connection
to the old IP Address, so the delayed break fails.  Therefore, this
client is locked out of the fileserver until:

        1) the fileserver reboots,
        2) the client returns to the original IP Address, or
        3) All the callbacks timeout on their own.

This patch will fix this problem.  When the client makes a request and
the fileserver tries to break the delayed callbacks, if the breaking
fails then the fileserver will attempt to find a 'working' interface
by probing all the host interfaces for one that responds with the
correct Uuid.  If that succeeds then it resets the cached callback
connection and then breaks the delayed callbacks, thereby regaining
the connection to the client and proceeding with the proper cleanup
before the original request is completed."

23 years agobuild-vfsck-for-solaris8-20010326
Derrick Brashear [Mon, 26 Mar 2001 17:26:29 +0000]
build-vfsck-for-solaris8-20010326

Before it was being explicitly skipped; Build it

23 years agonull-terminate-cell-when-copying-from-lrealm-20010326
Harald Barth [Mon, 26 Mar 2001 17:18:42 +0000]
null-terminate-cell-when-copying-from-lrealm-20010326

Avoid condition where we could read past the end of lrealm (unlikely
to occur in practice and not remotely exploitable; cell and lrealm are
both MAXKTCREALMLEN and lrealm is configured locally and not read from the
network)

23 years agodo-bounds-checking-in-psetvolumestatus-20010315
Nickolai Zeldovich [Fri, 16 Mar 2001 03:35:07 +0000]
do-bounds-checking-in-psetvolumestatus-20010315

Previously it was possible to overflow the variables used to construct a
SetVolumeStatus call

23 years agofix-for-cache-consistency-on-linux-24-smp-systems-20010315
Erik J. Burckart [Fri, 16 Mar 2001 03:27:05 +0000]
fix-for-cache-consistency-on-linux-24-smp-systems-20010315

"Here are the symptoms.  I have 2 OpenAFS  Linux 2.4 machines, one SMP and
one not.
--With both of them I access a series of directories and files within
(/afs/w/x/y/z/).
--I then move to the root directory.
--From a third machine, I remove the directories x, y, and z. Then re-add
them (same names) with some other files.
--From the SMP box, I do an ls in /afs/w and see x.  But x is inaccessible.
--From the UP box I can see and access x.
"

====================
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.
====================
Missed earlier: not part of the IBM delta.

====================
Part of the alpha patch slipped into the delta; Revert to previous
revision

23 years agodux-advfs-client-support-20010315
Johan Danielsson [Thu, 15 Mar 2001 18:55:20 +0000]
dux-advfs-client-support-20010315

necessary changes to allow a cache in an advfs

23 years agobosserver-syslog-option-including-selectable-facility-20010315
Nathan Neulinger [Thu, 15 Mar 2001 18:47:50 +0000]
bosserver-syslog-option-including-selectable-facility-20010315

"Simple patch to add "-syslog" option to bosserver. Causes it to direct all
logging activity to syslogd instead of to logs/BosLog"

currently unimplemented for NT

====================
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.
====================
Don't use a static outside its scope

23 years agoinitial-alpha-linux-support-20010315
Michael Stroucken [Thu, 15 Mar 2001 18:15:46 +0000]
initial-alpha-linux-support-20010315

top-level makefile mods for 64 bit system

23 years agopts-args-checking-when-calling-com-err-20010314
Nickolai Zeldovich [Thu, 15 Mar 2001 17:50:34 +0000]
pts-args-checking-when-calling-com-err-20010314

"This prevents pts from crashing like this:

  pepsi-one~> pts createg a:b -id -1234
  /bin/athena/pts: Permission denied ; unable to create group a:b with id -1234
owned by 'Segmentation fault
"

23 years agofix-locking-on-linux-sparc64-20010312
Pavel Semerad [Thu, 15 Mar 2001 17:47:45 +0000]
fix-locking-on-linux-sparc64-20010312

" I have found problem with file locking and 64bit linux (tested on
ultrasparc). When issued fcntl(fd, f_SETLK, ..) with l_start = 0 and l_len
= 0, it sets l_len to LONG_MAX, so it is needed to test this value"

23 years agosimplification-of-code-in-cache-manager-to-make-dux-advfs-support-simpler-20010315
Johan Danielsson [Thu, 15 Mar 2001 17:43:11 +0000]
simplification-of-code-in-cache-manager-to-make-dux-advfs-support-simpler-20010315

with advfs you can't VTOI inodes so provide a wrapper function for dux to ease
this problem

23 years agomake-comerr-tolerate-different-eol-convention-20010306
Eric Lorenzo [Tue, 6 Mar 2001 17:57:00 +0000]
make-comerr-tolerate-different-eol-convention-20010306

23 years agomake-afsd-be-less-annoying-20010306
Derrick Brashear [Tue, 6 Mar 2001 16:02:00 +0000]
make-afsd-be-less-annoying-20010306

23 years agoxlc-r-on-path-not-usr-bin-20010306
Derrick Brashear [Tue, 6 Mar 2001 06:16:17 +0000]
xlc-r-on-path-not-usr-bin-20010306

By default it ends up in /usr/vac/bin

23 years agouse-dollar-cc-not-cc-for-aix-exporter-20010305
Derrick Brashear [Mon, 5 Mar 2001 22:57:57 +0000]
use-dollar-cc-not-cc-for-aix-exporter-20010305

necessary to build aix afs with gcc

23 years agomake-snprintf-compile-on-aix-433-20010305
Derrick Brashear [Mon, 5 Mar 2001 22:47:54 +0000]
make-snprintf-compile-on-aix-433-20010305

AF_INET is in sys/socket.h on AIX 4.3.3

23 years agomake-bosserver-be-less-annoying-20010305
Derrick Brashear [Mon, 5 Mar 2001 19:13:24 +0000]
make-bosserver-be-less-annoying-20010305

23 years agorestore-rxdebug-to-the-living-20010305
Derrick Brashear [Mon, 5 Mar 2001 18:55:13 +0000]
restore-rxdebug-to-the-living-20010305

Give us back rxdebug

23 years agosupport-namei-fileserver-on-solaris8-20010305
Hartmut Reuter [Mon, 5 Mar 2001 16:39:05 +0000]
support-namei-fileserver-on-solaris8-20010305

So the namei fileserver can be used on Solaris 8

23 years agoafs-linux-s390-updates-20010305
Adam Thornton [Mon, 5 Mar 2001 16:34:32 +0000]
afs-linux-s390-updates-20010305

Updates for Linux s/390 AFS port

23 years agoreturn-truncated-reason-from-krb-udp-interface-20010305
Derrick Brashear [Mon, 5 Mar 2001 16:26:04 +0000]
return-truncated-reason-from-krb-udp-interface-20010305

The remainder of the patch the snprintf support was intended to be used for

23 years agopam-afs-password-changing-support-20010305
Ulrich Hahn [Mon, 5 Mar 2001 16:22:53 +0000]
pam-afs-password-changing-support-20010305

support for changing passwords when user has kaserver password only

23 years agoallow-shadowed-src-trees-20010305
Hartmut Reuter [Mon, 5 Mar 2001 15:37:46 +0000]
allow-shadowed-src-trees-20010305

So "make links" will work if src was a symlink.

23 years agontp-for-solaris-58-20010305
Hartmut Reuter [Mon, 5 Mar 2001 15:34:14 +0000]
ntp-for-solaris-58-20010305

Let ntp build on Solaris 8

23 years agosupport-dux-51-20010305
Johan Danielsson [Mon, 5 Mar 2001 15:12:42 +0000]
support-dux-51-20010305

Changes needed for OpenAFS Tru64 5.1 support

23 years agolinux-libafs-with-modversions-20010228
Michael Welsh Duggan [Wed, 28 Feb 2001 19:22:44 +0000]
linux-libafs-with-modversions-20010228

Use symbol versioning when compiling against a kernel with symbol versioning
enabled.

23 years agoafs-lookup-check-evalmountpoint-return-20010221
Nickolai Zeldovich [Sat, 24 Feb 2001 15:35:05 +0000]
afs-lookup-check-evalmountpoint-return-20010221

Make check the return code from EvalMountPoint()

====================
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.
====================
maintain old behavior consistent with checking eval return

23 years agobackup-addvolentry-fix-20010221
Mattias Amnefelt [Thu, 22 Feb 2001 03:06:04 +0000]
backup-addvolentry-fix-20010221

23 years agobos-mrafs-support-20010221
Derrick Brashear [Thu, 22 Feb 2001 00:34:45 +0000]
bos-mrafs-support-20010221

Make it work as intended

23 years agoupdate-README-20010213
Nathan Neulinger [Tue, 13 Feb 2001 16:53:48 +0000]
update-README-20010213

Irix compile completes

23 years agosnprintf-for-aix-not-irix-20010212
Derrick Brashear [Tue, 13 Feb 2001 03:54:04 +0000]
snprintf-for-aix-not-irix-20010212

Irix 6.5 has snprintf

23 years agoprovide-snprintf-for-irix-dux-20010212
Derrick Brashear [Tue, 13 Feb 2001 03:46:15 +0000]
provide-snprintf-for-irix-dux-20010212

So we can use snprintf, provide it for platforms that don't have it

23 years agobos-mrafs-support-20010212
Hartmut Reuter [Tue, 13 Feb 2001 00:28:22 +0000]
bos-mrafs-support-20010212

Necessary additions to bos for MR-AFS support

23 years agorevert-xdr-int64-in-favor-of-common-makefile-20010212
Hartmut Reuter [Tue, 13 Feb 2001 00:15:25 +0000]
revert-xdr-int64-in-favor-of-common-makefile-20010212

the new version of the patch put these in Makefile.common

23 years agoinclude-xdr-int64-in-libafs-20010212
Hartmut Reuter [Mon, 12 Feb 2001 23:56:16 +0000]
include-xdr-int64-in-libafs-20010212

for new pioctl calls for mrafs

====================
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.
====================
for new pioctl calls for mrafs; convert spaces to tab