On Wed, 2006-12-06 at 20:34 -0500, Trond Myklebust wrote: > On Wed, 2006-12-06 at 17:16 -0800, Linus Torvalds wrote: > > > > On Wed, 6 Dec 2006, Trond Myklebust wrote: > > > > > > git pull git://git.linux-nfs.org/pub/linux/nfs-2.6.git > > > > > > This will update the following files through the appended changesets. > > > > Well, right now it conflicts with the workqueue cleanups. Can you fix up > > the conflicts and push again? Quite frankly, I could try, but since I > > don't even run NFS, I _really_ don't think you want me to do so. > > OK. I'll fix it up and resend.
I've fixed up the conflicts with the workqueue stuff. It both compiles and works for me. Please could you try pulling again. Cheers, Trond -------------------------------------------------------------------------- Hi Linus, Please pull from the repository at git pull git://git.linux-nfs.org/pub/linux/nfs-2.6.git This will update the following files through the appended changesets. Cheers, Trond ---- fs/lockd/clntproc.c | 2 fs/lockd/svc4proc.c | 2 fs/lockd/svcproc.c | 2 fs/nfs/direct.c | 6 fs/nfs/file.c | 28 + fs/nfs/inode.c | 2 fs/nfs/internal.h | 18 + fs/nfs/nfs3proc.c | 52 -- fs/nfs/nfs4proc.c | 99 ---- fs/nfs/pagelist.c | 67 ++- fs/nfs/proc.c | 31 - fs/nfs/read.c | 179 ++++---- fs/nfs/symlink.c | 2 fs/nfs/write.c | 598 +++++++++++++-------------- include/linux/nfs_fs.h | 37 -- include/linux/nfs_page.h | 7 include/linux/nfs_xdr.h | 2 include/linux/sunrpc/auth_gss.h | 2 include/linux/sunrpc/clnt.h | 1 include/linux/sunrpc/debug.h | 6 include/linux/sunrpc/gss_krb5.h | 6 include/linux/sunrpc/gss_spkm3.h | 34 +- include/linux/sunrpc/sched.h | 11 include/linux/sunrpc/xdr.h | 23 + include/linux/sunrpc/xprt.h | 37 -- net/sunrpc/auth_gss/auth_gss.c | 42 -- net/sunrpc/auth_gss/gss_krb5_crypto.c | 101 ----- net/sunrpc/auth_gss/gss_krb5_mech.c | 18 + net/sunrpc/auth_gss/gss_krb5_seal.c | 55 +- net/sunrpc/auth_gss/gss_krb5_unseal.c | 87 +--- net/sunrpc/auth_gss/gss_krb5_wrap.c | 153 ++----- net/sunrpc/auth_gss/gss_spkm3_mech.c | 131 ++---- net/sunrpc/auth_gss/gss_spkm3_seal.c | 101 ++++- net/sunrpc/auth_gss/gss_spkm3_token.c | 6 net/sunrpc/auth_gss/gss_spkm3_unseal.c | 92 ++-- net/sunrpc/clnt.c | 70 ++- net/sunrpc/pmap_clnt.c | 13 - net/sunrpc/sched.c | 137 ++++-- net/sunrpc/socklib.c | 18 - net/sunrpc/sunrpc_syms.c | 5 net/sunrpc/sysctl.c | 50 -- net/sunrpc/xdr.c | 255 +++++++---- net/sunrpc/xprt.c | 33 - net/sunrpc/xprtsock.c | 716 ++++++++++++++++++++------------ 44 files changed, 1557 insertions(+), 1780 deletions(-) commit 5847e1f4d058677c5e46dc6c3e3c70e8855ea3ba Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:36:14 2006 -0500 SUNRPC: Remove pprintk() from net/sunrpc/xprt.c These appear to be deprecated. Removing them also gets rid of some sparse noise. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit c041b5ff8d50c432698d9bfbd68cc4b76c2ea5bc Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:36:03 2006 -0500 NLM: fix print format for tk_pid The tk_pid field is an unsigned short. The proper print format specifier for that type is %5u, not %4d. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit fbf76683ff9d1462ec0b2f90ec6ea4793652318c Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:54 2006 -0500 SUNRPC: relocate the creation of socket-specific tunables Clean-up: The RPC client currently creates some sysctls that are specific to the socket transport. Move those entirely into xprtsock.c. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 282b32e17f64be2204f1ac96d7f40f92cb768cd7 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:51 2006 -0500 SUNRPC: create stubs for xprtsock init and cleanup Over time we will want to add some specific init and cleanup logic for the xprtsock implementation. Add stub routines for initialization and exit processing. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit dd4564715eae2c4136f278da9ae1c3bb5af3e509 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:44 2006 -0500 SUNRPC: Rename skb_reader_t and friends Clean-up: hch suggested that the RPC client shouldn't pollute the name space used by the generic skb manipulation routines in net/core/skbuff.c. Rename a couple of types in xdr.h to adhere to this convention. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 9d29231690925915015c21c1fff73c7118099843 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:41 2006 -0500 SUNRPC: skb_read_bits is the same as xs_tcp_copy_data Clean-up: eliminate xs_tcp_copy_data -- it's exactly the same logic as the common routine skb_read_bits. The UDP and TCP socket read code now share the same routine for copying data into an xdr_buf. Now that skb_read_bits() is exported, rename it to avoid confusing it with a generic skb_* function. As these functions are XDR-specific, they should not have names that suggest they are of generic use. Also rename skb_read_and_csum_bits() to be consistent. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 7559c7a28fbcaa0bca028eeebd5f251b09befe6b Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:37 2006 -0500 SUNRPC: Make address format buffers more generic For now we will assume that all transports will use the address format buffers in the rpc_xprt struct to store their addresses. Change rpc_peer2str() to be a generic routine to handle this, and get rid of the print_address() op in the rpc_xprt_ops vector. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 314dfd7987c71d7ba0c43ac3bf3d243c102ce025 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:34 2006 -0500 SUNRPC: move saved socket callback functions to a private data structure Move the three fields for saving socket callback functions out of the rpc_xprt structure and into a private data structure maintained in net/sunrpc/xprtsock.c. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 7c6e066ec29290bf062f5bff2984bad9be5809c7 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:30 2006 -0500 SUNRPC: Move the UDP socket bufsize parameters to a private data structure Move the socket-specific buffer size parameters for UDP sockets to a private data structure maintained in net/sunrpc/xprtsock.c. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit c8475461829fd94f30208fbfa4eab7e5584c6495 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:26 2006 -0500 SUNRPC: Move rpc_xprt socket connect fields into private data structure Move the socket-specific connection management fields out of the generic rpc_xprt structure into a private data structure maintained in net/sunrpc/xprtsock.c. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e136d0926ef6a048f6e65b35263c0a9faae3abbe Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:23 2006 -0500 SUNRPC: Move TCP state flags into xprtsock.c Move "XPRT_LAST_FRAG" and friends from xprt.h into xprtsock.c, and rename them to use the naming scheme in use in xprtsock.c. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 51971139b2342fa1005e87bbfcb52305da3fe891 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:19 2006 -0500 SUNRPC: Move TCP receive state variables into private data structure Move the TCP receive state variables from the generic rpc_xprt structure to a private structure maintained inside net/sunrpc/xprtsock.c. Also rename a function/variable pair to refer to RPC fragment headers instead of record markers, to be consistent with types defined in sunrpc/*.h. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit ee0ac0c227c2a2b6dd1b33c23831100ee895dacf Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:15 2006 -0500 SUNRPC: Remove sock and inet fields from rpc_xprt The "sock" and "inet" fields are socket-specific. Move them to a private data structure maintained entirely within net/sunrpc/xprtsock.c Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit ffc2e518c91942b7ed45fb0ab7deba1ba0c8594a Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Dec 5 16:35:11 2006 -0500 SUNRPC: Allocate a private data area for socket-specific rpc_xprt fields When setting up a new transport instance, allocate enough memory for an rpc_xprt and a private area. As part of the same memory allocation, it will be easy to find one, given a pointer to the other. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 94efa93435383b08d822a40f24ff6f6ce1a888df Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:42 2006 -0500 rpcgss: krb5: miscellaneous cleanup Miscellaneous cosmetic fixes. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 717757ad1038ab6aacb89bad579c89b006efd913 Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:41 2006 -0500 rpcgss: krb5: ignore seed We're currently not actually using seed or seed_init. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit d922a84a8bf1d627810906d033223d4fa629fdbf Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:40 2006 -0500 rpcgss: krb5: sanity check sealalg value in the downcall The sealalg is checked in several places, giving the impression it could be either SEAL_ALG_NONE or SEAL_ALG_DES. But in fact SEAL_ALG_NONE seems to be sufficient only for making mic's, and all the contexts we get must be capable of wrapping as well. So the sealalg must be SEAL_ALG_DES. As with signalg, just check for the right value on the downcall and ignore it otherwise. Similarly, tighten expectations for the sealalg on incoming tokens, in case we do support other values eventually. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 39a21dd1b0eec3f5eac84ee42bda5ab4915098ae Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:39 2006 -0500 rpcgss: krb5: clean up some goto's, etc. Remove some unnecessary goto labels; clean up some return values; etc. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit ca54f896454852f0bc8d50e6e4c55d9defedbd0a Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:38 2006 -0500 rpcgss: simplify make_checksum We're doing some pointless translation between krb5 constants and kernel crypto string names. Also clean up some related spkm3 code as necessary. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 2818bf81a8c91fb29634df68bdc3cc5e003201d0 Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:37 2006 -0500 rpcgss: krb5: kill checksum_type, miscellaneous small cleanup Previous changes reveal some obvious cruft. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 5eb064f93973def1ec2ab4a46929e94389a6283b Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:36 2006 -0500 rpcgss: krb5: expect a constant signalg value We also only ever receive one value of the signalg, so let's not pretend otherwise Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e678e06bf8fa25981a6fa1f08b979fd086d713f8 Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:35 2006 -0500 gss: krb5: remove signalg and sealalg We designed the krb5 context import without completely understanding the context. Now it's clear that there are a number of fields that we ignore, or that we depend on having one single value. In particular, we only support one value of signalg currently; so let's check the signalg field in the downcall (in case we decide there's something else we could support here eventually), but ignore it otherwise. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit adeb8133dd57f380e70a389a89a2ea3ae227f9e2 Author: Olga Kornievskaia <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:34 2006 -0500 rpc: spkm3 update This updates the spkm3 code to bring it up to date with our current understanding of the spkm3 spec. In doing so, we're changing the downcall format used by gssd in the spkm3 case, which will cause an incompatilibity with old userland spkm3 support. Since the old code a) didn't implement the protocol correctly, and b) was never distributed except in the form of some experimental patches from the citi web site, we're assuming this is OK. We do detect the old downcall format and print warning (and fail). We also include a version number in the new downcall format, to be used in the future in case any further change is required. In some more detail: - fix integrity support - removed dependency on NIDs. instead OIDs are used - known OID values for algorithms added. - fixed some context fields and types Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 37a4e6cb0391f2293ba3d59e3a63ec0e56ed720d Author: Olga Kornievskaia <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:33 2006 -0500 rpc: move process_xdr_buf Since process_xdr_buf() is useful outside of the kerberos-specific code, we move it to net/sunrpc/xdr.c, export it, and rename it in keeping with xdr_* naming convention of xdr.c. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 87d918d667e51962938392759aef6ca368d6e96d Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:32 2006 -0500 rpc: gss: fix a kmap_atomic race in krb5 code This code is never called from interrupt context; it's always run by either a user thread or rpciod. So KM_SKB_SUNRPC_DATA is inappropriate here. Thanks to Aimé Le Rouzic for capturing an oops which showed the kernel taking an interrupt while we were in this piece of code, resulting in a nested kmap_atomic(.,KM_SKB_SUNRPC_DATA) call from xdr_partial_copy_from_skb(). Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 8fc7500bb8ea3b5c909869d00628635e964ae882 Author: J. Bruce Fields <[EMAIL PROTECTED]> Date: Mon Dec 4 20:22:31 2006 -0500 rpc: gss: eliminate print_hexl()'s Dumping all this data to the logs is wasteful (even when debugging is turned off), and creates too much output to be useful when it's turned on. Fix a minor style bug or two while we're at it. Signed-off-by: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 2b577f1f14c8d83ba9400ec8accaf4a208f4f36a Author: Chuck Lever <[EMAIL PROTECTED]> Date: Thu Nov 16 15:03:38 2006 -0500 SUNRPC: another pmap wakeup fix Don't wake up bind waiters if a task finds that another task is already trying to bind. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit a18030445fd4dd20e2248007b5d1cf0b5d89c69d Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:45:19 2006 -0500 NFS: Clean up calls to mark_inode_dirty() part 2 Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 9cf85e0a243b56f4ef8439a1e6430307fd9dcda6 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:36:56 2006 -0500 NFS: Fix up writeback_control->nr_to_write accounting We're really accounting for the same page twice now: once in generic_writepages(), and once in nfs_scan_dirty(). Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 3925675cb37cc9c3fd1d3f56ce0fa729f995f863 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:36:09 2006 -0500 NFS: Fix up the dirty page accounting There is now no reason to account for the dirty pages in the NFS code, since the VM code will now do it for us via __set_page_dirty_nobuffers(), and set_page_writeback(). We still need to keep the accounting of stable writes, though. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e507d9ebbb2d5db5948a6fb3c33f015d60708d19 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:42 2006 -0500 NFS: Ensure the inode is marked as dirty if we break out of nfs_wb_all() Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit fa8d8c5b77a2dc467b5365a5651710161b84f16d Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:42 2006 -0500 NFS: Fix nfs_release_page invalidate_inode_pages2_range() will clear the PG_dirty bit before calling try_to_release_page(). Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 61822ab5e3ed09fcfc49e37227b655202adf6130 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:42 2006 -0500 NFS: Ensure we only call set_page_writeback() under the page lock Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e261f51f25b98c213e0b3d7f2109b117d714f69d Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:41 2006 -0500 NFS: Make nfs_updatepage() mark the page as dirty. This will ensure that we can call set_page_writeback() from within nfs_writepage(), which is always called with the page lock set. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 4d770ccf4257b23a7ca2a85de1b1c22657b581d8 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:41 2006 -0500 NFS: Ensure that nfs_wb_page() calls writepage when necessary. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 1a54533ec8d92a5edae97ec6ae10023ee71c4b46 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:40 2006 -0500 NFS: Add nfs_set_page_dirty() We will want to allow nfs_writepage() to distinguish between pages that have been marked as dirty by the VM, and those that have been marked as dirty by nfs_updatepage(). In the former case, the entire page will want to be written out, and so any requests that were pending need to be flushed out first. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 200baa2112012dd8a13db9da3ee6885403f9c013 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:40 2006 -0500 NFS: Remove nfs_writepage_sync() Maintaining two parallel ways of doing synchronous writes is rather pointless. This patch gets rid of the legacy nfs_writepage_sync(), and replaces it with the faster asynchronous writes. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e21195a740533348e77efa8a2e2cf03bb4092b2b Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:39 2006 -0500 NFS: More cleanups of fs/nfs/write.c Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 87a4ce16082e92b4b6d27596a517b302f3692650 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:39 2006 -0500 NFS: Remove call to igrab() from nfs_writepage() We always ensure that the nfs_open_context holds a reference to the dentry, so the test in nfs_writepage() for whether or not the inode is referenced is redundant. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 49a70f278658894d2899824cd4037095fb6711fe Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:38 2006 -0500 NFS: Cleanup: add common helper nfs_page_length() Clean up a lot of ad-hoc page length calculations in fs/nfs/write.c Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 277459d2e2cd40594967757e8fd016c4c7016614 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Dec 5 00:35:35 2006 -0500 NFS: Store pointer to the nfs_page in page->private This will allow fast lookup of the nfs_page from the struct page instead of having to search the radix tree. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 1c75950b9a2254ef08f986e00ad20266ae9ba7f1 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Mon Oct 9 16:18:38 2006 -0400 NFS: cleanup of nfs_sync_inode_wait() Allow callers to directly pass it a struct writeback_control. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 3f442547b76bf9fb70d7aecc41cf1980459253c9 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Sun Sep 17 14:46:44 2006 -0400 NFS: Clean up nfs_scan_dirty() Pass down struct writeback control. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 28c6925fce3927a9fe3c5b44af5fb266680fdcea Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Sat Sep 16 13:04:50 2006 -0400 NFS: Clean up nfs_flush_inode() Make it take a struct writepages argument, and rename to nfs_flush_mapping(). Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit c8541ecdd5692bcfbcb5305cab9a873288d29175 Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Oct 17 14:44:27 2006 -0400 SUNRPC: Make the transport-specific setup routine allocate rpc_xprt Change the location where the rpc_xprt structure is allocated so each transport implementation can allocate a private area from the same chunk of memory. Note also that xprt->ops->destroy, rather than xprt_destroy, is now responsible for freeing rpc_xprt when the transport is destroyed. Test plan: Connectathon. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e744cf2e3ab8535a8494a4cf0177de26f94586da Author: Chuck Lever <[EMAIL PROTECTED]> Date: Tue Oct 17 14:44:24 2006 -0400 SUNRPC: minor optimization of "xid" field in rpc_xprt Move the xid field in the rpc_xprt structure to be in the same cache line as the reserve_lock, since these are used at the same time. Test plan: None. Signed-off-by: Chuck Lever <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 24c5684b65ff52ebfa942e8086d91a4966121ae7 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Oct 17 15:06:22 2006 -0400 SUNRPC: Clean up xs_send_pages() Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit bee57c99c322d64407b80c8171958b4384902da4 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Mon Oct 9 22:08:22 2006 -0400 SUNRPC: Ensure xdr_buf_read_netobj() checks for memory overruns Also clean up the code... Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 4e3e43ad14c574281034a27420abf1993694ac11 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Oct 17 13:47:24 2006 -0400 SUNRPC: Add __(read|write)_bytes_from_xdr_buf Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 1e78957e0a8f882df6a3660b62f9aae441f54891 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Thu Aug 31 15:09:19 2006 -0400 SUNRPC: Clean up argument types in xdr.c Converts various integer buffer offsets and sizes to unsigned integer. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit eb5f8545ffff98a11c6656d4d2106341ab69c57d Author: Frank Filz <[EMAIL PROTECTED]> Date: Tue Oct 17 10:24:42 2006 -0700 NFS: Remove use of the Big Kernel Lock around nfs calls to readlink Remove use of the Big Kernel Lock around indirect calls to nfs3_proc_readlink and nfs4_proc_readlink, both of which basically call rpc_call_sync. Signed-off-by: Frank Filz <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit cae823c4c089d2c7c2a90f39f20376ccd85f9893 Author: Frank Filz <[EMAIL PROTECTED]> Date: Tue Oct 17 10:24:38 2006 -0700 NFS: Remove use of the Big Kernel Lock around calls to rpc_call_sync Remove use of the Big Kernel Lock around calls to rpc_call_sync. Signed-off-by: Frank Filz <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit a99b71c9c43499bf2312c64f5c1d367aaf559dc4 Author: Frank Filz <[EMAIL PROTECTED]> Date: Tue Oct 17 10:24:36 2006 -0700 NFS: Remove use of the Big Kernel Lock around calls to rpc_execute. Remove use of the Big Kernel Lock around calls to rpc_execute. Signed-off-by: Frank Filz <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 6d5fcb5a52bfd00eab3ba2c7ca890823388436ae Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Wed Oct 18 16:01:06 2006 -0400 SUNRPC: Remove BKL around the RPC socket operations etc. All internal RPC client operations should no longer depend on the BKL, however lockd and NFS callbacks may still require it. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit bbd5a1f9fc9fad0f8725812d91c51b052e847de8 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Wed Oct 18 16:01:05 2006 -0400 SUNRPC: Fix up missing BKL in asynchronous RPC callback functions Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 3e32a5d99a467b9d4d416323c8c292479b4915e5 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Thu Nov 16 11:37:27 2006 -0500 SUNRPC: Give cloned RPC clients their own rpc_pipefs directory Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 23bf85ba43650e4341672a6bc85aa3f2c02eb633 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Nov 21 10:40:23 2006 -0500 SUNRPC: Handle the cases where rpc_alloc_iostats() fails Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e8e058e830f46a76f837522e5e2df46d4303111f Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Wed Nov 15 17:31:56 2006 -0500 NFS: Fix nfs_sync_inode_wait(FLUSH_INVALIDATE) Currently nfs_sync_inode_wait() will fail to loop correctly when we call nfs_sync_inode_wait with the FLUSH_INVALIDATE argument. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit cf1308ff7829017dab0dbcc817c63dc9c212923e Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Sun Nov 19 16:44:52 2006 -0500 NFS: Fix missing page_unlock() in nfs_readpage Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 0b67130149b006628389ff3e8f46be9957af98aa Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Tue Nov 14 16:12:23 2006 -0500 NFS: Fix asynchronous read error handling We must always call ->read_done() before we truncate the page data, or decide to flag an error. The reasons are that in NFSv2, ->read_done() is where the eof flag gets set. in NFSv3/v4 ->read_done() handles EJUKEBOX-type errors, and v4 state recovery. However, we need to mark the pages as uptodate before we deal with short read errors, since we may need to modify the nfs_read_data arguments. We therefore split the current nfs_readpage_result() into two parts: nfs_readpage_result(), which calls ->read_done() etc, and nfs_readpage_retry(), which subsequently handles short reads. Note: Removing the code that retries in case of a short read also fixes a bug in nfs_direct_read_result(), which used to return a corrupted number of bytes. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 46b9f8e1484352f09f229107ba2a758fe386d7f7 Author: Andy Ryan <[EMAIL PROTECTED]> Date: Tue Nov 7 14:36:26 2006 -0600 NFS Exclusive open not supported bug When trying to open a file with the O_EXCL flag over NFS on a server that does not support exclusive mode, the file does not open. The reason, rpc_call_sync returns a errno number, and not the nfs error number. I fixed it by changing the status check in nfs3proc.c. Either this is how it should be fixed, or rpc_call_sync should be fixed to return the NFS error. Signed-off-by: Andy Ryan <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit 8aca67f0ae2d8811165c22326825a645cc8e1b48 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Mon Nov 13 16:23:44 2006 -0500 SUNRPC: Fix a potential race in rpc_wake_up_task() Use RCU to ensure that we can safely call rpc_finish_wakeup after we've called __rpc_do_wake_up_task. If not, there is a theoretical race, in which the rpc_task finishes executing, and gets freed first. Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit e6b3c4db6fbcd0d33720696f37790d6b8be12313 Author: Trond Myklebust <[EMAIL PROTECTED]> Date: Sat Nov 11 22:18:03 2006 -0500 Fix a second potential rpc_wakeup race... Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> commit cc4dc59e5580d6c0de1685a25b74d32175f43434 Author: Christophe Saout <[EMAIL PROTECTED]> Date: Sun Nov 5 18:42:48 2006 +0100 Subject: Re: [PATCH] Fix SUNRPC wakeup/execute race condition The sunrpc scheduler contains a race condition that can let an RPC task end up being neither running nor on any wait queue. The race takes place between rpc_make_runnable (called from rpc_wake_up_task) and __rpc_execute under the following condition: First __rpc_execute calls tk_action which puts the task on some wait queue. The task is dequeued by another process before __rpc_execute continues its execution. While executing rpc_make_runnable exactly after setting the task `running' bit and before clearing the `queued' bit __rpc_execute picks up execution, clears `running' and subsequently both functions fall through, both under the false assumption somebody else took the job. Swapping rpc_test_and_set_running with rpc_clear_queued in rpc_make_runnable fixes that hole. This introduces another possible race condition that can be handled by checking for `queued' after setting the `running' bit. Bug noticed on a 4-way x86_64 system under XEN with an NFSv4 server on the same physical machine, apparently one of the few ways to hit this race condition at all. Cc: Trond Myklebust <[EMAIL PROTECTED]> Cc: J. Bruce Fields <[EMAIL PROTECTED]> Signed-off-by: Christophe Saout <[EMAIL PROTECTED]> Signed-off-by: Trond Myklebust <[EMAIL PROTECTED]> - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/