- 11 Apr, 2006 40 commits
-
-
Miklos Szeredi authored
A deadlock was possible, when the last reference to the superblock was held due to a background request containing a file reference. Releasing the file would release the vfsmount which in turn would release the superblock. Since sbput_sem is held during the fput() and fuse_put_super() tries to acquire this same semaphore, a deadlock results. The chosen soltuion is to get rid of sbput_sem, and instead use the spinlock to ensure the referenced inodes/file are released only once. Since the actual release may sleep, defer these outside the locked region, but using local variables instead of the structure members. This is a much more rubust solution. Signed-off-by:Miklos Szeredi <miklos@szeredi.hu>
-
git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuildLinus Torvalds authored
* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild: kconfig: fix typo in change count initialization kconfig: recenter menuconfig kconfig: revert conf behaviour change kconfig: fix default value for choice input kbuild: fix NULL dereference in scripts/mod/modpost.c kbuild: fix mode of checkstack.pl and other files. kbuild: rebuild initramfs if content of initramfs changes kbuild: properly pass options to hostcc when doing make O=.. kbuild: modules_install for external modules must not remove existing modules kbuild: fix make dir/ ver_linux: don't print reiser4progs version if none found kbuild: mips: fix sed regexp to generate asm-offset.h kbuild: fix building single targets with make O=.. single-target kbuild: use relative path to -I kbuild: fix unneeded rebuilds in drivers/net/chelsio after moving source tree kbuild: fix unneeded rebuilds in drivers/media/video after moving source tree kbuild: fix garbled text in modules.txt
-
git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6Linus Torvalds authored
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6: [IA64] Prefetch mmap_sem in ia64_do_page_fault() [IA64] Failure to resume after INIT in user space [IA64] Pass more data to the MCA/INIT notify_die hooks [IA64] always map VGA framebuffer UC, even if it supports WB [IA64] fix bug in ia64 __mutex_fastpath_trylock [IA64] for_each_possible_cpu: ia64 [IA64] update HP CSR space discovery via ACPI [IA64] Wire up new syscalls {set,get}_robust_list [IA64] 'msg' may be used uninitialized in xpc_initiate_allocate() [IA64] Wire up new syscall sync_file_range()
-
mao, bibo authored
In vsyscall function do_vgettimeofday(), some functions are declared as inlined, which is a hint for gcc to compile the function inlined but it not forced. Sometimes compiler does not compile the function as inlined, so here inline is replaced by __always_inline prefix. It does not happen in gcc compiler actually, but it possibly happens. Signed-off-by:
bibo mao <bibo.mao@intel.com> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Andi Kleen authored
Surprising that it still worked at all with this - yes it was tested. Signed-off-by:
Andi Kleen <ak@suse.de> Signed-off-by:
-
Andi Kleen authored
Nobody should pass NULL here. Could in theory make it a BUG, but the NULL pointer oops will do as well. Signed-off-by:
-
Andi Kleen authored
It's not actually needed and would break non power of two number of cores. Follows similar earlier x86-64 patch. Signed-off-by:
-
Siddha, Suresh B authored
Commit c818a181 didn't do the expected thing. This fix will remove the additional sync(cpuid) before RDTSC on Intel platforms.. Signed-off-by:
Suresh Siddha <suresh.b.siddha@intel.com> Signed-off-by:
-
Andi Kleen authored
As pointed out by Linus it is useless now because entry.S should handle it correctly in all cases. Signed-off-by:
-
Andi Kleen authored
When nolapic was passed or the local APIC was disabled for another reason ACPI would still parse the IO-APICs until these were explicitely disabled with noapic. Usually this resulted in a non booting configuration unless "nolapic noapic" was used. I also disabled the local APIC parsing in this case, although that's only cosmetic (suppresses a few printks) This hopefully makes nolapic work in all cases. Cc: Len Brown <len.brown@intel.com> Signed-off-by:
-
git://brick.kernel.dk/data/git/linux-2.6-blockLinus Torvalds authored
* 'splice' of git://brick.kernel.dk/data/git/linux-2.6-block: [PATCH] vfs: add splice_write and splice_read to documentation [PATCH] Remove sys_ prefix of new syscalls from __NR_sys_* [PATCH] splice: warning fix [PATCH] another round of fs/pipe.c cleanups [PATCH] splice: comment styles [PATCH] splice: add Ingo as addition copyright holder [PATCH] splice: unlikely() optimizations [PATCH] splice: speedups and optimizations [PATCH] pipe.c/fifo.c code cleanups [PATCH] get rid of the PIPE_*() macros [PATCH] splice: speedup __generic_file_splice_read [PATCH] splice: add direct fd <-> fd splicing support [PATCH] splice: add optional input and output offsets [PATCH] introduce a "kernel-internal pipe object" abstraction [PATCH] splice: be smarter about calling do_page_cache_readahead() [PATCH] splice: optimize the splice buffer mapping [PATCH] splice: cleanup __generic_file_splice_read() [PATCH] splice: only call wake_up_interruptible() when we really have to [PATCH] splice: potential !page dereference [PATCH] splice: mark the io page as accessed
-
Matthias Urlichs authored
Since the arrays are declared as in_urbs[N_IN_URB] and out_urbs[N_OUT_URB], both for loops go one over the end of the array. This fixes coverity id #555. Signed-off-by:
Eric Sesterhenn <snakebyte@gmx.de> Signed-Off-By:
Matthias Urlichs <smurf@smurf.noris.de> Signed-Off-By:
-
Antonino A. Daplas authored
Since the visual STATIC_PSEUDOCOLOR has a read-only colormap, use logos with 16 colors only since these logos use the console palette. This has a higher likelihood that the logo will display correctly. Signed-of-by:
Antonino Daplas <adaplas@pol.net> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
-
Antonino A. Daplas authored
Bugzilla Bug 6299: A pixel size of 8 bits produces wrong logo colors in x86_64. The driver has 2 methods for setting the color map, using the protected mode interface provided by the video BIOS and directly writing to the VGA registers. The former is not supported in x86_64 and the latter is enabled only in i386. Fix by enabling the latter method in x86_64 only if supported by the BIOS. If both methods are unsupported, change the visual of vesafb to STATIC_PSEUDOCOLOR. Signed-off-by:
-
Andrew Morton authored
sparc32 lacks vga.h, so lots of fbdev drivers won't compile. There are no sparc32 systems with PCI slots, so it's a bit moot. The patch gives sparc32 a copy of the sparc64 vga.h. It fixes sparc32 allmodconfig without mucking up fbdev Kconfig and gives us wider compile coverage. Cc: "Antonino A. Daplas" <adaplas@pol.net> Acked-by:
"David S. Miller" <davem@davemloft.net> Signed-off-by:
-
Andrew Morton authored
Heaps of build errors - disable it to keep sparc32 allmodconfig happy. Signed-off-by:
-
Adrian Bunk authored
Fix an obvious of-by-one error spotted by the Coverity checker. Signed-off-by:
Adrian Bunk <bunk@stusta.de> Cc: "Antonino A. Daplas" <adaplas@pol.net> Signed-off-by:
-
NeilBrown authored
Keep unused openowners around for at least one lease period, to avoid the need for as many open confirmations and to allow handing out more delegations. Signed-off-by:
J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by:
Neil Brown <neilb@suse.de> Signed-off-by:
-
NeilBrown authored
It's very easy for the server to DOS itself by just giving out too many delegations. For now we just solve the problem with a dumb hard limit. Eventually we'll want a smarter policy. Signed-off-by:
-
NeilBrown authored
We should be shutting down rpciod for the callback channel when we shut down the server. Also note that we do rpciod_up() and create the callback client *before* setting cb_set--the cb_set only determines whether the initial null was succesful. So cb_set is not a reliable determiner of whether we need to clean up, only cb_client is. Signed-off-by:
-
NeilBrown authored
Some obvious cleanup. Signed-off-by:
-
NeilBrown authored
We need to make sure the laundromat work doesn't reschedule itself just when we try to cancel it. Also, we shouldn't be waiting for it to finish running while holding the state lock, as that's a potential deadlock. Signed-off-by:
-
J. Bruce Fields authored
Every caller of svc_take_page ignores its return value and assumes it succeeded. So just WARN() instead of returning an ignored error. This would have saved some time debugging a recent nfsd4 problem. If there are still failure cases here, then the result is probably that we overwrite an earlier part of the reply while xdr-encoding. While the corrupted reply is a nasty bug, it would be worse to panic here and create the possibility of a remote DOS; hence WARN() instead of BUG(). Signed-off-by:
-
NeilBrown authored
We're using svc_take_page here to get another page for the tail in case one wasn't already allocated. But there isn't always guaranteed to be another page available. Also fix a typo that made us check the tail buffer for space when we meant to be checking the head buffer. Signed-off-by:
-
NeilBrown authored
Fix corruption on readdir encoding with 64k pages. Signed-off-by:
-
NeilBrown authored
In v4 we grab an extra page just for the padding of returned data. The formula that the rpc server uses to allocate pages for the response doesn't take into account this extra page. Instead of adjusting those formulae, we adopt the same solution as v2 and v3, and put the "tail" data in the same page as the "head" data. Signed-off-by:
-
NeilBrown authored
Since nfsd_setuser() is already called from any operation that uses the current filehandle (because it's called from fh_verify), there's no reason to call it from putrootfh. Signed-off-by:
Andy Adamson <andros@citi.umich.edu> Signed-off-by:
-
NeilBrown authored
In addition to setting the processes filesystem id's, nfsd_setuser also modifies the value of the rq_cred which stores the id's that originally came from the rpc call, for example to reflect root squashing. There's no real reason to do that--the only case where rqstp->rq_cred is actually used later on is in the NFSv4 SETCLIENTID/SETCLIENTID_CONFIRM operations, and there the results are the opposite of what we want--those two operations don't deal with the filesystem at all, they only record the credentials used with the rpc call for later reference (so that we may require the same credentials be used on later operations), and the credentials shouldn't vary just because there was or wasn't a previous operation in the compound that referred to some export This fixes a bug which caused mounts from Solaris clients to fail. Signed-off-by:
-
NeilBrown authored
Export a directory that does not exist: exportfs -orw,fsid=0,insecure,no_subtree_check client:/home/NFS4 Try to mount from client with nfs4. Mount hangs (I'm not sure why - that's another issue). While client is hung, back on server mkdir /home/NFS4 The server panics in dput. I traced the problem back to svc_export_parse() calling path_release() even though path_lookup() failed (it happens to fill in the nameidata structure with a negative dentry - so the test after out: succeeds). After patching, an recreating the problem, the client mount still takes some time before finally exiting with a message "couldn't read superblock". Here is a simple patch to resolve this issue: Signed-off-by:
Frank Filz <ffilzlnx@us.ibm.com> Signed-off-by:
-
NeilBrown authored
We should be using the length from the second vfs_getxattr, in case it changed. (Note: there's still a small race here; we could end up returning -ENOMEM if the length increased between the first and second call. I don't know whether it's worth spending a lot of effort to fix that.) This makes XFS ACLs usable on NFS exports, which they currently aren't, since XFS appears to be returning a too-large value for vfs_getxattr() when it's passed a NULL buffer. So there's probably an XFS bug here too, though since getxattr with a NULL buffer is usually used to decide how much memory to allocate, it may be a fairly harmless bug in most cases. Signed-off-by:
-
NeilBrown authored
We're returning -1 in a few places in the NFSv4<->POSIX acl translation code where we could return a reasonable error. Also allows some minor simplification elsewhere. Signed-off-by:
-
NeilBrown authored
this fixes coverity id #3. Coverity detected dead code, since the == -1 comparison only returns 0 or 1 to error. Therefore the if ( error < 0 ) statement was always false. Seems that this was an if( error = nfs4... ) statement some time ago, which got broken during cleanup. Signed-off-by:
-
Adrian Bunk authored
Signed-off-by:
-
NeilBrown authored
Use the fl_lmops field to identify which locks are ours, instead of trying to look them up in our private hash. This is safer and more efficient. Earlier versions of this patch used a lock flag instead, but Trond pointed out that adding a new flag for each lock manager wasn't going to scale well, and suggested this approach instead; a separate patch converts lockd to using fl_lmops in the same way. In the NFSv4 case this looks like a bit of a hack, since the NFSv4 server isn't currently actually defining a lock_manager_operations struct, so we end up defining one *just* to serve as a cookie to identify our locks. But it works, and we actually do expect to start using the lock_manager_operations at some point anyway. Signed-off-by:
Marc Eshel <eshel@almaden.ibm.com> Signed-off-by:
-
NeilBrown authored
NFSd makes sure there is enough space to hold the maximum possible reply before accepting a request. The units for this maximum is (4byte) words. However in three places, particularly for read request, the number given is a number of bytes. This means too much space is reserved which is slightly wasteful. This is the sort of patch that could uncover a deeper bug, and it is not critical, so it would be best for it to spend a while in -mm before going in to mainline. (akpm: target 2.6.17-rc2, 2.6.16.3 (approx)) Discovered-by:
"Eivind Sarto" <ivan@kasenna.com> Signed-off-by:
-
Adrian Bunk authored
CRC_CCITT is an internal helper function that should be select'ed. Signed-off-by:
Tilman Schmidt <tilman@imap.cc> Signed-off-by:
-
Adrian Bunk authored
Fix a memory leak spotted by the Coverity checker if (!try_module_get(owner)). Signed-off-by:
-
Adrian Bunk authored
- make the needlessly global gigaset_get_cs_by_tty() static - remove the unused EXPORT_SYMBOL_GPL(gigaset_debugdrivers) Signed-off-by:
-
Tilman Schmidt authored
With Hansjoerg Lipp <hjlipp@web.de> Replace some atomic_t variables in the Gigaset drivers by non-atomic ones, using spinlocks instead to assure atomicity, as proposed in discussions on the linux-kernel mailing list. Signed-off-by:
Hansjoerg Lipp <hjlipp@web.de> Signed-off-by:
-
Tilman Schmidt authored
With Hansjoerg Lipp <hjlipp@web.de> Add a README file for the Siemens Gigaset drivers to the Documentation/isdn directory. Signed-off-by:
-
