Currently AFS doesn't handle Explorer operations 'rename' and 'move'
correctly. Normal operation of 'rename' and 'move' is to warn the operator
when the destination file name already exist.
makefile changes.
initialize variables to NULL before use
clean up CellServDB parsing
make sure files do not get timestamp of -1 (1969)
update product version
Garry Zacheiss [Wed, 20 Nov 2002 03:18:47 +0000 (03:18 +0000)]
Make this work with Linux 2.2 kernels again. In particular,
- Use the DCOUNT() macro from afs.h instead of atomic_read, because
the d_count member of a struct dentry is an int in the 2.2 series,
not an atomic_t.
- Use list_del() + INIT_LIST_HEAD() on 2.2, which doesn't have list_del_init().
- BUG() only exists on 2.4 kernels. Just don't use it on 2.2.
Derrick Brashear [Fri, 15 Nov 2002 13:45:31 +0000 (13:45 +0000)]
readd missing statistics
make cm and ukernel compile with new stuff
====================
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 missing include path
====================
un-nocopy bg-fcrypt
====================
and readd rx/rx.h so private_data's struct clock ref is resolved
hold the dcache lock only briefly; instead put dcache entries to be junked
onto a free list, by unhashing them, then grab the dcache lock, junk them,
and release it.
Garry Zacheiss [Fri, 15 Nov 2002 04:08:57 +0000 (04:08 +0000)]
fix typo introduced during merge.
====================
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.
====================
rxkad fcrypt from bg@pdc.kth.se. we should also add autoconf test for what
to use for EFF_NTOHL
====================
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.
====================
rxkad fcrypt from bg@pdc.kth.se. we should also add autoconf test for what
to use for EFF_NTOHL
Added support to rxkad for servers accepting Kerberos V5 tickets and
truncated "proposal 2b" tickets. When used with an appropriate aklog
or krb524d (such as shipped with Heimdal and MIT Kerberos), this allows
the use of Kerberos V5 tickets (with DES session keys) to authenticate
connections to OpenAFS servers.
Under Solaris, only clean up the open count in VOP_INACTIVE when
the vcache is mvstat 0 (necessary because executables don't get
VOP_CLOSE'd). Volume roots (mvstat 2) need to keep their open
counts, because under fakestat, the mountpoint above it is still
considered open by the kernel.
The attached patch attempts to deal with the issue that the volserver
blocks if the fileserver's fssync interface isn't responsive by doing a
select (IOMGR_Select) before trying to read the server's response.
If the user calls rx_EndCall and we still have an outstanding
delayack, force-send it now, to be nice to the server (otherwise
the server will keep retransmitting to us for a while).
Change the meaning of the -fakestat switch to only enable fakestat
for cross-cell mountpoints. The -fakestat-all switch can now be
used to enable fakestat for all mountpoints (old behavior).
Nathan Neulinger [Thu, 14 Nov 2002 20:01:59 +0000 (20:01 +0000)]
If we're doing our own xdr, we should be doing ALL of of own xdr, not
piecemeal. This impacted on HP w/ duplicate symbols if -lnsl, and
missing symbols if no -lnsl.
David Howells [Thu, 14 Nov 2002 19:53:00 +0000 (19:53 +0000)]
init linux rwlocks we missed before
====================
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.
====================
Nick Ingolia [Thu, 14 Nov 2002 19:47:12 +0000 (19:47 +0000)]
If ngroups in the parent is 0, but the first two elements of the groups
array represent a valid PAG (perhaps because they weren't zeroed), then
set_pag_in_parent won't add 2 to ngroups and the parent process won't
join the PAG.
====================
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.
====================
Tom Maher [Tue, 10 Sep 2002 23:33:24 +0000 (23:33 +0000)]
a) a freebsd client that compiles, and to a certain extent, works. Should
not be used except by people doing freebsd client development
b) configure glue for freebsd 4.3, 4.4, and 4.5
c) freebsd tfileserver support. Note that, by default, this isn't very
interesting, as the supported freebsd pthreads are just as cooperative as
LWP. if you install the linuxthreads port and swap the MT_* variables in
osconf.m4, you will get a more traditional tfileserver though. configure
glue for that will be forthcoming at some point.
d) a maybe-fix that sets sin_len in sockaddrs on platforms that have them.
(and also a fix for the AC_PROG_CC issue which affects autoconf 2.13 and 2.5x
on same conf files issue)
====================
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.
====================
Derrick Brashear [Wed, 28 Aug 2002 08:20:25 +0000 (08:20 +0000)]
Irix needs this. Put it back.
Note: This was part of the delta prototypes-fixes-20020821 on the mainline,
but this chunk is needed on the 1.2.x branch, and it would be appropriate
to pull in the entire delta in which it appeared.
Make "fs newcell" work even when there are no pre-existing cells in
the kernel yet. Previously, newcell tried to issue a GetCell pioctl
to figure out what version of the pioctl interface the kernel is
using. Lacking such information (if the pioctl returns an error),
we now assume the newer interface.
Garry Zacheiss [Wed, 28 Aug 2002 04:37:27 +0000 (04:37 +0000)]
Don't return ELOOP when crossing a mountpoint boundary from a BK volume to
a BK volume. find treats it as a fatal error, and it's annoying not to be
able to naively use find to recurse through a volume with its backup volume
mounted. Never noticed this before because we used to ignore this return
value. Return ENODEV instead.
Always verify that the client is fully reachable before talking
to it, in the fileserver. This prevents some additional lossage
cases with poorly-connected clients.
Disable same feature for other Rx services, since they don't
need to incur extra RTT delay.