openafs.git
21 years agoosi_linux_unmask() needs to explicitly flush_signals() before calling openafs-stable-1_2_4
Derek Atkins [Wed, 15 May 2002 21:07:29 +0000]
osi_linux_unmask() needs to explicitly flush_signals() before calling
recalc_sigpending(), otherwise recent kernels don't appear to deliver
new instances of any signals already received.

Revert back to using SIGKILL at shutdown time, but use SIGKILL instead of
explicitly using signal 9, because abstraction is your friend.

(cherry picked from commit 8251fc6871a65326150be69b929917b37d83e215)

21 years agoSeems to fix a hang at shutdown time seen on Redhat 7.3 kernels; doesn't
Derek Atkins [Wed, 15 May 2002 00:09:58 +0000]
Seems to fix a hang at shutdown time seen on Redhat 7.3 kernels; doesn't
appear to affect older kernels.

(cherry picked from commit 097df1049a6316f087e1c79b30a332358fa761c9)

21 years agoAllows us to build modules for the bigmem kernels redhat started
Derek Atkins [Tue, 14 May 2002 23:53:49 +0000]
Allows us to build modules for the bigmem kernels redhat started
shipping.

(cherry picked from commit fe6b4fd293f789c4d9f8f62401e31a78b466b09b)

21 years agoUpdate NEWS file for user visible changes in 1.2.4.
Garry Zacheiss [Mon, 13 May 2002 02:17:33 +0000]
Update NEWS file for user visible changes in 1.2.4.

(cherry picked from commit 50473f885f07857d03b963bf43c5b24771b4913b)

21 years agoAdd 1.2.4 release notes.
Garry Zacheiss [Mon, 13 May 2002 02:15:38 +0000]
Add 1.2.4 release notes.

21 years agobegin release cycle.
Garry Zacheiss [Fri, 10 May 2002 20:02:54 +0000]
begin release cycle.

21 years agoAvoid kernel panic when dynroot and memcache are used simultaneously.
Nickolai Zeldovich [Thu, 9 May 2002 23:07:12 +0000]
Avoid kernel panic when dynroot and memcache are used simultaneously.

(cherry picked from commit 81a3f0068f3488b7fb20fb9d6b7441f64f5b61e0)

21 years agoSTABLE12-dir-buffer-newslot-cant-return-zero-20020508
Derrick Brashear [Wed, 8 May 2002 19:59:53 +0000]
STABLE12-dir-buffer-newslot-cant-return-zero-20020508

since newslot will call Die() instead of returning 0 (causing assert(0))
this code can never be called. remove it.

21 years agoSTABLE12-need-prototype-for-osi-dnlc-lookup-20020503
Chas Williams [Mon, 6 May 2002 17:21:42 +0000]
STABLE12-need-prototype-for-osi-dnlc-lookup-20020503

Include afs_osidnlc.h for osi_dnlc_lookup() prototype.  Apparently
fixes a panic on ia64.

21 years agoSTABLE12-linux-include-osidnlc-header-for-ia64-20020506
Chas Williams [Mon, 6 May 2002 17:19:46 +0000]
STABLE12-linux-include-osidnlc-header-for-ia64-20020506

this is in the head and not here; ia64 needs the prototype

21 years agoInclude afs_osidnlc.h for osi_dnlc_lookup() prototype. Apparently
Chas Williams [Fri, 3 May 2002 20:06:36 +0000]
Include afs_osidnlc.h for osi_dnlc_lookup() prototype.  Apparently
fixes a panic on ia64.

21 years agokeep small globals from being gprel
Chas Williams [Thu, 2 May 2002 16:59:39 +0000]
keep small globals from being gprel

(cherry picked from commit 1e90eb264895b5bbe974c7c8c32dbf89e48d6081)

21 years agorewrite syscall stubs not in machine code
Chas Williams [Thu, 2 May 2002 16:38:45 +0000]
rewrite syscall stubs not in machine code

(cherry picked from commit 25de65bc0bba9bc2df382dc34c1c5b0c1460aca5)

21 years agoavoid ESTALE in 2.4.19-pre7
Kevin Coffman [Thu, 2 May 2002 16:36:01 +0000]
avoid ESTALE in 2.4.19-pre7

(cherry picked from commit 7147cc946bead75eee909ea1134bc542c0b09c4f)

21 years agoDon't spawn a separate kernel thread for rx kernel listener under
Nickolai Zeldovich [Mon, 29 Apr 2002 20:17:00 +0000]
Don't spawn a separate kernel thread for rx kernel listener under
Solaris, but instead just use the afsd fork for the listener. Use
of newproc() is deprecated, according to Frank.Batschulat@Sun.COM.

(cherry picked from commit be1458b07817900d15f5c069ea66b72ef50b91e9)

21 years agoIf the inode is not allocated, don't bother updating timestamps.
Nickolai Zeldovich [Mon, 29 Apr 2002 20:13:25 +0000]
If the inode is not allocated, don't bother updating timestamps.
Pointed out by Frank.Batschulat@Sun.COM.

(cherry picked from commit a31928815796bcbd1acf3005a0f628a21bbb0d8a)

21 years agoSTABLE12-darwin-use-namei-fileserver-20020429
Chaskiel M Grundman [Mon, 29 Apr 2002 16:33:27 +0000]
STABLE12-darwin-use-namei-fileserver-20020429

build namei fileserver for darwin

(cherry picked from commit c2e075fba4cec152e772ec5514d52e8ec5b29990)

21 years agoSTABLE12-cvsignore-updates-20020429
Chaskiel M Grundman [Mon, 29 Apr 2002 16:27:12 +0000]
STABLE12-cvsignore-updates-20020429

stuff missed in the stable branch

(cherry picked from commit 291e8352be1ee0c6e3da79328e752b82a7a13190)

22 years agoSTABLE12-darwin-map-more-versions-20020425
Chaskiel M Grundman [Thu, 25 Apr 2002 22:47:26 +0000]
STABLE12-darwin-map-more-versions-20020425

5.3 and 5.4 are "Same" enough

(cherry picked from commit 1ccedea1a0d978f213a9051e15b5e5efb7ccffe8)

22 years agoSTABLE12-darwin-fix-build-20020401
Derrick Brashear [Thu, 25 Apr 2002 22:22:30 +0000]
STABLE12-darwin-fix-build-20020401

add missing )

(cherry picked from commit fd318a1fc9551a156cdb43916143d3b959b6720f)

22 years agoSTABLE12-rxkad-stats-initialize-to-placate-darwin-compiler-20020425
Chaskiel M Grundman [Thu, 25 Apr 2002 22:19:52 +0000]
STABLE12-rxkad-stats-initialize-to-placate-darwin-compiler-20020425

because it deals badly with common symbols

(cherry picked from commit 0c0eb03e9a806dc84b7c3251478de4766a8fd853)

22 years agoeven though right now a vnode and an inode are basically but not exactly
Derrick Brashear [Wed, 24 Apr 2002 02:49:34 +0000]
even though right now a vnode and an inode are basically but not exactly
the same, part of the usefulness of this change is the ability to
convert to inodes linked from vcaches later. make the macros fit with
this view of the fture

(cherry picked from commit 39ee79e7c328d86738356aa5a736e24339f8b126)

22 years agoconvert vcache casting in cache manager to macros
Johan Danielsson [Wed, 24 Apr 2002 02:41:16 +0000]
convert vcache casting in cache manager to macros

should ease further changes later

(cherry picked from commit 3bf8818d956337fcf70a2734f200275f403e8814)

22 years agoprovide sysctl interface for toggling module state. provide hooks for
Kris Van Hees [Tue, 23 Apr 2002 23:22:54 +0000]
provide sysctl interface for toggling module state. provide hooks for
all currently known tweakable items

(complete earlier incomplete pullup)

22 years agoSTABLE12-linux-ia64-cleanup-casting-20020423
Chas Williams [Tue, 23 Apr 2002 14:46:55 +0000]
STABLE12-linux-ia64-cleanup-casting-20020423

so we don't need the SYSCALL2POINTER change

(cherry picked from commit 8268bf34acd715d28137a3d601252fd235d1a9fb)

22 years agoupdate setgroups32 for 2.4 and make sparc64 linux2.4 work again
Chaskiel M Grundman [Tue, 23 Apr 2002 04:32:31 +0000]
update setgroups32 for 2.4 and make sparc64 linux2.4 work again

(cherry picked from commit 5d57da192ea989d9ab4e83cb5f721a7975def9fb)

22 years agoSTABLE12-linux-ia64-updates-20020422
Chas Williams [Tue, 23 Apr 2002 04:22:29 +0000]
STABLE12-linux-ia64-updates-20020422

updates to make it work with later 2.4

SYSCALL2POINTER change may be unnecessary (sparc64 way should work)
but without any way to test...

(cherry picked from commit e12859bed90a0bd896150b7cfc137d0ef52e48c3)

22 years agoSTABLE12-darwin-vnops-make-unlink-fix-fake-references-20020422
Chaskiel M Grundman [Mon, 22 Apr 2002 17:06:36 +0000]
STABLE12-darwin-vnops-make-unlink-fix-fake-references-20020422

so unlink doesn't panic us

(cherry picked from commit 6d43ab450a907bb1eccf9561c37c108202cd8e52)

22 years agoSTABLE12-libadmin-pts-allow-listing-all-users-and-groups-20020422
Jeremy Stribling [Mon, 22 Apr 2002 16:43:27 +0000]
STABLE12-libadmin-pts-allow-listing-all-users-and-groups-20020422

so the java api can take advantage of this

(cherry picked from commit df363edad10f0cbc7f1f34fd0f79a155525415e7)

22 years agoit compiles. it's not pretty, but it does compile
Derrick Brashear [Mon, 22 Apr 2002 04:25:51 +0000]
it compiles. it's not pretty, but it does compile

(cherry picked from commit 607c17c8a97bd10367c98397089554d316c52492)

22 years agoDon't make tvc a register, since we want to dereference it later on.
Nickolai Zeldovich [Sun, 21 Apr 2002 22:16:44 +0000]
Don't make tvc a register, since we want to dereference it later on.
Pointed out by nneul@umr.edu.

(cherry picked from commit 999c7fb70dca5153c58e4f0c9bc6c1d86003865b)

22 years agoThis adds /useexisting to use current AFS creds, bypassing the initial
David Bailey [Sun, 21 Apr 2002 22:11:11 +0000]
This adds /useexisting to use current AFS creds, bypassing the initial
authentication dialog box. It's useful if you're not running kaserver
(i.e. if you've moved over to kerberos 5) and can't authenticate again
by the normal AFS means.

(cherry picked from commit b0e4e7bcbfb911563f68cd91c21d905738123d72)

22 years agolookup is sufficient for this, fix the client not to bogusly
Derrick Brashear [Sun, 21 Apr 2002 22:08:47 +0000]
lookup is sufficient for this, fix the client not to bogusly
enforce extra restrictions

(cherry picked from commit 8011b683bc72968afcf69d6c7220716281bcba55)

22 years agoavoid leaking catalog reference on failure
Nathan Neulinger [Sun, 21 Apr 2002 22:05:46 +0000]
avoid leaking catalog reference on failure

(cherry picked from commit 1ad668aee2c995e44974f89728b1d5e13470aec1)

22 years agoautoconf 2.5x turds to ignore
Nathan Neulinger [Sun, 21 Apr 2002 22:04:16 +0000]
autoconf 2.5x turds to ignore

(cherry picked from commit 13f864f21e28893cf482dcc51b498698f9ecc006)

22 years agoVerify that name lengths are at most AFSNAMEMAX and symlink
Nickolai Zeldovich [Sun, 21 Apr 2002 22:01:36 +0000]
Verify that name lengths are at most AFSNAMEMAX and symlink
contents are at most AFSPATHMAX (notably, the Linux kernel
doesn't check this for us); bug report by nneul@umr.edu.

(cherry picked from commit 41f047edf69ad7c6c83cfa94f410f0daa9bb91fd)

22 years agoInitialize tvc in afs_remove so that we don't call afs_PutVCache
Nickolai Zeldovich [Sun, 21 Apr 2002 21:58:18 +0000]
Initialize tvc in afs_remove so that we don't call afs_PutVCache
on stack garbage in case of some failure.

(cherry picked from commit 0329c5fbb42d14da20acb97ebda48b9974220b0e)

22 years agoClean up some more ViceLog calls to consistently refer to hosts in a
Nickolai Zeldovich [Sun, 21 Apr 2002 21:54:59 +0000]
Clean up some more ViceLog calls to consistently refer to hosts in a
format like 18.239.2.102:7001, rather than memory address or hex IP.

(cherry picked from commit c4ac62f829f5642b72df9e01438fc4545cda98f9)

22 years agoprototyping caught this; the usage is inconsistent with other cases and
Derrick Brashear [Sun, 21 Apr 2002 21:53:28 +0000]
prototyping caught this; the usage is inconsistent with other cases and
i believe this is the correct fix. it's not clear to me why we weren't seeing
problems from this, though

(cherry picked from commit ef1ee322deb3a2e21c9d78a1b44375b8f7768de0)

22 years agothis means to be calling afs_GetUser with a cell index, not with the cell
Nathan Neulinger [Sun, 21 Apr 2002 21:50:26 +0000]
this means to be calling afs_GetUser with a cell index, not with the cell
structure itself

(cherry picked from commit 4f1ca9bb2ebcad7e4e1c7ebea42eb361d506dcc8)

22 years agotru64 needs vrele when an error happens, rather than just returning. make
Mattias Amnefelt [Sun, 21 Apr 2002 21:45:56 +0000]
tru64 needs vrele when an error happens, rather than just returning. make
sure we do so

(cherry picked from commit 7ffbc40b212a38de92da562d285e715150e9d771)

22 years agofile server now works usefully on netbsd
Love Hörnquist-Åstrand [Sun, 21 Apr 2002 21:27:47 +0000]
file server now works usefully on netbsd

(cherry picked from commit efa250bd1c4bf2c8d6ac74d9f866a842a020d7cf)

22 years agoCall afs_PutCell after afs_GetCellByIndex, even though right
Nickolai Zeldovich [Sun, 21 Apr 2002 21:23:56 +0000]
Call afs_PutCell after afs_GetCellByIndex, even though right
now it doesn't do anything.

(cherry picked from commit 246f955432b4d00d8f1ffbe62aa9c2a048833c2f)

22 years agosuggested by lha@stacken.kth.se
Derrick Brashear [Sun, 21 Apr 2002 21:12:22 +0000]
suggested by lha@stacken.kth.se

drop locks when rxi_Alloc fails, and rework SRXAFSCB_GetCellServDB to use
afs_GetCellByIndex. He suggested same for GetLocalCell, and as long as cell 1
and not CPrimary is the local cell indicator that works, but I left that part
of the code as is for now.

(cherry picked from commit fc09b92dce32025c8e1ac762b3f8f07bc731869a)

22 years agoZero out the "allocated" flag so that we don't accidentally
Nathan Neulinger [Sun, 21 Apr 2002 20:54:17 +0000]
Zero out the "allocated" flag so that we don't accidentally
free something that wasn't allocated.

(cherry picked from commit c360bb35ce506ad9ab425fea43b96a708e0ed1d9)

22 years agouse pragmas to hit compiler into packing smb_tran2QFSInfo usefully
Yee Jiun [Sun, 21 Apr 2002 20:51:54 +0000]
use pragmas to hit compiler into packing smb_tran2QFSInfo usefully

(cherry picked from commit 908e05adbedd1130279d82cfe4ee8e9c11c9a4cc)

22 years agoFix typo: OMIT_FRAME_POINTER not FOMIT_FRAME_POINTER
Nathan Neulinger [Sun, 21 Apr 2002 20:48:34 +0000]
Fix typo: OMIT_FRAME_POINTER not FOMIT_FRAME_POINTER

(cherry picked from commit 1fe41530c1c1a6b75a7716ed849ff37be4294317)

22 years agocurrently for linux only, but allow it for others in the future
Derek Atkins [Sun, 21 Apr 2002 20:46:35 +0000]
currently for linux only, but allow it for others in the future

(cherry picked from commit edd10398afe756afb8b7d8abbb73324e3a1b2cfd)

22 years agoThe problem is that on 2.4 the 'code' variable is over-written during
Derek Atkins [Sun, 21 Apr 2002 06:13:01 +0000]
The problem is that on 2.4 the 'code' variable is over-written during
the path walk, such that by the time it's checking if the inode
exists, code is already '0'.  But if the inode does not exist, code
should be ENOENT to show that the file does not exist.

(cherry picked from commit f3bcfda51f5aa1185c3fab6b1c56d5c5165426f1)

22 years agonamei should now work on tru64 (tested on 5.0a)
Mattias Amnefelt [Sun, 21 Apr 2002 06:04:50 +0000]
namei should now work on tru64 (tested on 5.0a)

(cherry picked from commit 459815329bc52d5d264ed5f1ffbfae490d88a552)

22 years agoUse AFS_RXDEADTIME instead of hardcoding a value of 50
Nickolai Zeldovich [Sun, 21 Apr 2002 05:38:25 +0000]
Use AFS_RXDEADTIME instead of hardcoding a value of 50

(cherry picked from commit cfba7e00684a1f45f82285bcbadbc4571ab294d2)

22 years agoregression test for rxkad
Love Hörnquist-Åstrand [Sun, 21 Apr 2002 05:31:07 +0000]
regression test for rxkad

(cherry picked from commit f749ef6e024cf17911a52ade9be9b409fa25d29c)

22 years agouse rx library interface instead of tweaking in each program
Mattias Amnefelt [Sun, 21 Apr 2002 05:14:50 +0000]
use rx library interface instead of tweaking in each program

(cherry picked from commit 7138b2734cebbf93af4dd69c04cba242c20241b5)

22 years agoubik_tid and ubik_version now come from ubik-int.h
Nickolai Zeldovich [Sun, 21 Apr 2002 05:09:22 +0000]
ubik_tid and ubik_version now come from ubik-int.h

(cherry picked from commit 592d0fb0cd7a68ee5976fe3655e24b6ffb001e91)

22 years agoFix some of the fallout from having rxgen-fabricated prototypes.
Nickolai Zeldovich [Sun, 21 Apr 2002 05:01:48 +0000]
Fix some of the fallout from having rxgen-fabricated prototypes.

(cherry picked from commit ebb45c54d9a53906c2dc1245199cb95a0260252b)

22 years agoMake sure we have enough memory to allocate the return list.
Nickolai Zeldovich [Sun, 21 Apr 2002 04:56:31 +0000]
Make sure we have enough memory to allocate the return list.

(cherry picked from commit ec04dd0090b69f3cd39c93498da6175b2eb1a842)

22 years agoFix build when INSTRUMENT_LOCKS isn't defined
Nickolai Zeldovich [Sun, 21 Apr 2002 04:55:01 +0000]
Fix build when INSTRUMENT_LOCKS isn't defined

(cherry picked from commit 0dfa0954ffdf25f93e44394fce9d0f1c0872576e)

22 years agoFix minor build errors, and rename the dumpId argument of butc's
Nickolai Zeldovich [Sun, 21 Apr 2002 04:52:09 +0000]
Fix minor build errors, and rename the dumpId argument of butc's
DeleteDump to avoid collision with budb's #define dumpId.

(cherry picked from commit 1e6bc5de521a319a8a247738f41065a1e11fec2b)

22 years agominor tweaks found making solaris work
Derrick Brashear [Sat, 20 Apr 2002 19:53:56 +0000]
minor tweaks found making solaris work

(cherry picked from commit acf19ae7e92aaf4f283a6bf1fca3b4e4f24f4042)

22 years agoset $SHELL to override whatever your shell sets
Derrick Brashear [Sat, 20 Apr 2002 19:49:00 +0000]
set $SHELL to override whatever your shell sets

(cherry picked from commit 4f8f508e04ab8f4267b6950e36745c1f3f10ada9)

22 years agowas a case where we could end up testing "localcell" by accident
Derrick Brashear [Sat, 20 Apr 2002 19:45:47 +0000]
was a case where we could end up testing "localcell" by accident

(cherry picked from commit 87843056e53f7ff72d20bae0837f9be31a7a33b7)

22 years agosince configure creates the script, just chmod it
Derrick Brashear [Sat, 20 Apr 2002 19:43:41 +0000]
since configure creates the script, just chmod it

(cherry picked from commit aeeecc446b01258a443cb5705484ba90217407df)

22 years agomake it all build on solaris. provide necessary files to set up a kaserver
Derrick Brashear [Sat, 20 Apr 2002 19:38:50 +0000]
make it all build on solaris. provide necessary files to set up a kaserver
and authenticate against it (albeit with "known" keys)

22 years agocksum can be either number
Derrick Brashear [Fri, 19 Apr 2002 06:57:28 +0000]
cksum can be either number

(cherry picked from commit d92d4ac69f55e89b9b32b8927d722e1b040bc783)

22 years ago(cherry picked from commit 3a6b630ce11cf0253543fa5ba56bd1eca5b9a4f6)
Jaroslaw Polok [Fri, 19 Apr 2002 06:37:28 +0000]
(cherry picked from commit 3a6b630ce11cf0253543fa5ba56bd1eca5b9a4f6)

22 years agoUse autoconf to check for p_corefile in struct proc, rather than
Nickolai Zeldovich [Fri, 19 Apr 2002 05:47:30 +0000]
Use autoconf to check for p_corefile in struct proc, rather than
assuming it exists in Solaris 7 and above; nneul@umr.edu reports
that it doesn't for some kernel revision.

(cherry picked from commit d0ffec1b282518ab6d1395bfaad66f0ea8fe9feb)

22 years agosomehow i suspect one of our platforms doesn't do volatile and i'll
Derrick Brashear [Fri, 19 Apr 2002 05:42:11 +0000]
somehow i suspect one of our platforms doesn't do volatile and i'll
regret this, but this stops linux i386 when afs_int32 is an int from
going off into la-la land when an Rx retransmit is needed

(cherry picked from commit d7dca002beb55eefe49227db9e8e2ee4c4bb6170)

22 years agobad trace statement due to missing parameter; added
Ted Anderson [Fri, 19 Apr 2002 05:39:22 +0000]
bad trace statement due to missing parameter; added

(cherry picked from commit aef9d3502309f441640419f70f2dd8949395b32f)

22 years agothis was omitted from the original release and now can be included
Laura Stentz [Fri, 19 Apr 2002 05:34:19 +0000]
this was omitted from the original release and now can be included

(cherry picked from commit 808946cad88f45d2862345cae935eecd7b551345)

22 years agobased on feedback from ota@transarc.com
Nathan Neulinger [Fri, 19 Apr 2002 05:15:03 +0000]
based on feedback from ota@transarc.com

PrintInode returns a string, use it

(cherry picked from commit 9a6dbb4df705e063a6ab093b29e7619321aeb565)

22 years agoProperly AFS_GUNLOCK() when returning an error after having previously
Nickolai Zeldovich [Fri, 19 Apr 2002 05:10:38 +0000]
Properly AFS_GUNLOCK() when returning an error after having previously
AFS_GLOCK()'ed, in afs_mount() and afs_unmount().

(cherry picked from commit ebcef806960b3507d4de6edfd59a89c37b6ca2fd)

22 years agoAllow AFS to shutdown on Solaris (and subsequently be reloaded
Nickolai Zeldovich [Fri, 19 Apr 2002 04:55:54 +0000]
Allow AFS to shutdown on Solaris (and subsequently be reloaded
and restarted).  Still leaks memory, unlike the Linux module.

(cherry picked from commit c446c128e9be75fa8617a11403d867a8885ee80b)

22 years agoavoid case where we never come out of loop attempting to clear host callbacks
Hartmut Reuter [Fri, 19 Apr 2002 04:50:32 +0000]
avoid case where we never come out of loop attempting to clear host callbacks

(cherry picked from commit b7b178842458b274f9e5d45dbd1bb50826a3d58a)

22 years ago(cherry picked from commit 148b57dff5feed7c48b8ea83e1a570df722702b1)
Derrick Brashear [Fri, 19 Apr 2002 03:22:02 +0000]
(cherry picked from commit 148b57dff5feed7c48b8ea83e1a570df722702b1)

22 years agoSTABLE12-linux-newvcache-more-simplification-20020405
Derrick Brashear [Sat, 6 Apr 2002 05:57:23 +0000]
STABLE12-linux-newvcache-more-simplification-20020405

based on suggestion from vsrikanth@in.ibm.com

eliminate extra for loop, flatten function to drop dentries, fix
locking error i introduced

(cherry picked from commit 7747fce7e77e45ebb22728498052e96f527d2de0)

22 years agoSTABLE12-linux-dcache-create-negative-dentries-22-needs-while-still-avoiding-vlru...
Derrick Brashear [Tue, 2 Apr 2002 03:13:26 +0000]
STABLE12-linux-dcache-create-negative-dentries-22-needs-while-still-avoiding-vlru-cycle-20020328

based on patch by Srikanth Vishwanathan and modifications based on suggestions
from Ted Anderson applicable to this part of the code.

the original theory we operated under, namely that negative dentries
could be banished entirely, causes file creation to stop working
under linux 2.2 kernels. reverting that change means we have to deal with
these negative dentries, and Srikanth's original suggested patch does so.

(cherry picked from commit 4beb9adb661bade7caff8adf40ecb3882ae7deac)

22 years agoSTABLE12-revert-part-of-linux-dcache-inconsistency-fix-20020115
Derrick Brashear [Tue, 2 Apr 2002 03:12:24 +0000]
STABLE12-revert-part-of-linux-dcache-inconsistency-fix-20020115

may revisit this before release; however, open_namei on 2.2 wants a
dentry returned to it and so without this change, creating files in
afs breaks

(cherry picked from commit 372da2bebc1eb10aaacf4532d4cd5aea901a3c83)

22 years agoSTABLE12-linux-osi-alloc-up-alloc-sem-while-alloc-to-avoid-deadlock-against-glock...
Derrick Brashear [Tue, 2 Apr 2002 03:11:41 +0000]
STABLE12-linux-osi-alloc-up-alloc-sem-while-alloc-to-avoid-deadlock-against-glock-20020215

Based on report from David Howells of RedHat (via Guy Streeter)

(cherry picked from commit f6525247abd62427a05806608a6b58999cfe1a1a)

22 years agoSTABLE12-solaris-vfsck-allow-force-runs-when-old-inodes-exist-20020310
Derrick Brashear [Tue, 2 Apr 2002 03:10:46 +0000]
STABLE12-solaris-vfsck-allow-force-runs-when-old-inodes-exist-20020310

if running with -y report old inodes and continue anyway

(cherry picked from commit 2dbc58bc4a00a7ba755327b22b3fbd5fb05889a6)

22 years agoSTABLE12-linux-sysctl-fix-includes-20020206
Nickolai Zeldovich [Tue, 2 Apr 2002 03:10:03 +0000]
STABLE12-linux-sysctl-fix-includes-20020206

Include linux/sysctl.h after afsconfig.h, since we need the
modversion stuff first.

(cherry picked from commit 21455cdba29829fffbb9355495198a4c2a77bc24)

22 years agoSTABLE12-linux-provide-sysctl-interface-20020206
Kris Van Hees [Tue, 2 Apr 2002 03:07:32 +0000]
STABLE12-linux-provide-sysctl-interface-20020206

provide sysctl interface for toggling module state. provide hooks for
all currently known tweakable items

22 years agoSTABLE12-solaris-p-crlock-not-safe-to-hold-across-call-to-crset-20020206
Nickolai Zeldovich [Tue, 2 Apr 2002 03:06:32 +0000]
STABLE12-solaris-p-crlock-not-safe-to-hold-across-call-to-crset-20020206

crset() expects p_crlock to not be held when it's called; deal accordingly

(cherry picked from commit e7b3045d8ece68712d2bd07b8e92c90027dee870)

22 years agoSTABLE12-newvolumebyname-analyze-the-vrequest-actually-used-20020205
Derrick Brashear [Tue, 2 Apr 2002 03:05:15 +0000]
STABLE12-newvolumebyname-analyze-the-vrequest-actually-used-20020205

it's ok to support a null request being passed in, but if a real one is
passed in we shouldn't analyze it when it's not the one we used when
making calls.

(cherry picked from commit 918d96d66548ecff818ec438f26ae06abc3bd34e)

22 years agoAllocate the exact string length needed for the string in symlink
Nickolai Zeldovich [Mon, 18 Mar 2002 03:24:10 +0000]
Allocate the exact string length needed for the string in symlink
contents, rather than assuming that symlink data never contains
nulls (or that it's of the length we expect).

(cherry picked from commit 209f1c22577c2314a896de997bd7939fab2e3d52)

22 years agoSTABLE12-winnt-revert-64bit-change-in-stable-client-20020210
Derrick Brashear [Mon, 11 Feb 2002 22:29:48 +0000]
STABLE12-winnt-revert-64bit-change-in-stable-client-20020210

somehow this change crept in

22 years agoSTABLE12-linux-block-signals-in-cvwait-too-20020130 openafs-stable-1_2_3
Nickolai Zeldovich [Wed, 30 Jan 2002 21:15:36 +0000]
STABLE12-linux-block-signals-in-cvwait-too-20020130

Block signals when sleeping in CV_WAIT() and CV_TIMEDWAIT(), similar
to the linux-avoid-signal-wakeups-while-sleeping-20020129 change for
afs_osi_Sleep().

(cherry picked from commit 7573d5ed14df78d619a035ca50981f7ae3bfd5b6)

22 years agoSTABLE12-add-1_2_3-release-notes-20020130
Derrick Brashear [Wed, 30 Jan 2002 19:57:41 +0000]
STABLE12-add-1_2_3-release-notes-20020130

1.2.3 release notes

22 years agoSTABLE12-rx-cleanup-deadlock-and-refcnt-leak-20020121
Nickolai Zeldovich [Tue, 29 Jan 2002 19:52:14 +0000]
STABLE12-rx-cleanup-deadlock-and-refcnt-leak-20020121

The first part of the patch fixes a likely deadlock in the
case when we're destroying a client connection with active
calls.

The second part fixes a minor connection refcount leak in
my recent modification to protect against half-reachable
clients.

22 years agoSTABLE12-linux-avoid-signal-wakeups-while-sleeping-20020129
Nickolai Zeldovich [Tue, 29 Jan 2002 19:45:39 +0000]
STABLE12-linux-avoid-signal-wakeups-while-sleeping-20020129

Block all signals while sleeping; this prevents 100% cpu usage
when a process waiting for an afs operation to complete receives
a signal.

(cherry picked from commit 41fb11e6b23ba320f7fff442b2b6c3b01c339ad0)

22 years agoSTABLE12-linux-gunlock-when-mounting-twice-20020129
Nickolai Zeldovich [Tue, 29 Jan 2002 19:45:07 +0000]
STABLE12-linux-gunlock-when-mounting-twice-20020129

Properly release GLOCK when trying to mount AFS a second time.

(cherry picked from commit 905b71014b9166cc3ba7b921623a838ffbaa18b7)

22 years agoSTABLE12-create-openafs-1_2_3-20020129
Derrick Brashear [Tue, 29 Jan 2002 17:24:14 +0000]
STABLE12-create-openafs-1_2_3-20020129

begin release cycle

22 years agoSTABLE12-linux-use-fdatasync-directly-20020129
Nickolai Zeldovich [Tue, 29 Jan 2002 16:18:58 +0000]
STABLE12-linux-use-fdatasync-directly-20020129

Use filemap_fdata{sync,wait} directly (available in 2.4.5 and above)
rather than write_inode_now(), which steals afs inodes and puts them
on the global inode lists.

(cherry picked from commit 30c69dae289b0184d2d38e4d9781de5a2c81952c)

22 years agoSTABLE12-update-tests-for-stable-20020128
Derrick Brashear [Mon, 28 Jan 2002 23:56:08 +0000]
STABLE12-update-tests-for-stable-20020128

no objdir support in 1.2 yet

22 years agoSTABLE12-configure-add-darwin-52-support-as-14-since-no-change-otherwise-20020128
Derrick Brashear [Mon, 28 Jan 2002 21:31:21 +0000]
STABLE12-configure-add-darwin-52-support-as-14-since-no-change-otherwise-20020128

it was previously agreed that we wouldn't add gratuitous new sysnames unless things
(not just what the vendor calls the operating system) changed. precedent is
sun4m_412 still being "current" for sun4m_413, etc.

(cherry picked from commit b01f9b767323363c40b78f26b32f077b9c23f3da)

22 years agoSTABLE12-rx-dont-include-arraysize-in-offsetof-20020125
Nickolai Zeldovich [Sat, 26 Jan 2002 03:41:45 +0000]
STABLE12-rx-dont-include-arraysize-in-offsetof-20020125

so this compiles on irix

(cherry picked from commit 62bc57574aabfccf9e3684e7f0fe99dcbf9ad1e3)

22 years agoSTABLE12-linux-afsd-pids-block-signals-so-we-dont-need-to-flush-20020124
Derrick Brashear [Thu, 24 Jan 2002 20:16:45 +0000]
STABLE12-linux-afsd-pids-block-signals-so-we-dont-need-to-flush-20020124

no point in keeping this around

(cherry picked from commit 4ff68264b69e55cedd2f70ecc40482ab9ba0abe0)

22 years agoSTABLE12-solaris-rx-shutdown-sleep-only-if-listener-pid-not-freed-20020124
Derrick Brashear [Thu, 24 Jan 2002 10:47:24 +0000]
STABLE12-solaris-rx-shutdown-sleep-only-if-listener-pid-not-freed-20020124

replace if with while in case the listener is freed before we call sleep

(cherry picked from commit f2d38d3683345a98f118b971b5a1940e14111928)

22 years agoSTABLE12-linux-unmask-signals-in-rxk-listener-at-shutdown-20020124
Derrick Brashear [Thu, 24 Jan 2002 10:43:15 +0000]
STABLE12-linux-unmask-signals-in-rxk-listener-at-shutdown-20020124

since the rxk listener gets "kill"d at shutdown we need to unmask to wake
it up. this should probably be replaced but the logical way is evil,
involves simulating a packet to the listener so it wakes up and notices it's
time to die

(cherry picked from commit 5c2a4168a6e67733d4ec545936e13742c72ff9f8)

22 years agoSTABLE12-linux-cant-daemonize-due-to-exit-mm-call-oh-well-20020124
Derrick Brashear [Thu, 24 Jan 2002 10:09:38 +0000]
STABLE12-linux-cant-daemonize-due-to-exit-mm-call-oh-well-20020124

the call to daemonize() means we fail to shut down correctly on SMP machines;
this is due to the call to exit_mm() in daemonize()

(cherry picked from commit a89079efffb1610a55c0fdff42f2feb82b394ff1)

22 years agoSTABLE12-linux-rework-signal-blocking-for-afsdb-handler-and-clean-up-osi-invisible...
Derrick Brashear [Wed, 23 Jan 2002 19:29:49 +0000]
STABLE12-linux-rework-signal-blocking-for-afsdb-handler-and-clean-up-osi-invisible-before-im-completely-blind-20020123

based on suggestion from kolya@mit.edu

and osi_Invisible cleanup so ifdef logic would be more clear

====================
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 do this in userland cache manager

(cherry picked from commit 1f3f7bc3efea3ffca469fbfe090953dcaa668799)