gmane.linux.nfs

Re: [PATCH 6/6] NSM: Serialize SM_MON and SM_UNMON upcalls

J. Bruce Fields — 2008-12-04T03:55:24

OK, fair enough. I think it's a good catch, though, and one other motivation is that *someday* (after fixing a bunch of similar locking problems) it would be convenient if lockd could be made multithreaded. And if we haven't seen this particular problem in practice, part of the reason may be that lockd is single-threaded and entirely under the BKL. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 6/6] NSM: Serialize SM_MON and SM_UNMON upcalls

Chuck Lever — 2008-12-04T00:38:07

It's cleared in nlm_host_rebooted(), and in nsm_unmonitor(). The nlm_host_mutex already provides much the same serialization that is added in this patch, and at least on the server, nlm_lookup_host() and nlm_destroy_host() both run under the BKL. The problem (if there is one) is really in nlm_host_rebooted(), which this patch does nothing about. Let's drop this one for now. I think there is a way to get rpc.statd to report duplicates instead of ignoring them as it does now. We should get an idea of how pervasive this problem really is.

Re: [PATCH/RFC] svcgssd always sets an infinite expiry on authentication tokens etc.

Kevin Coffman — 2008-12-03T22:26:36

It took me a bit longer than I anticipated to retrofit my changes for this. I have patches that I will send out for review later tonight or tomorrow. K.C. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [NFS] export dir thru 2 diff path names

J. Bruce Fields — 2008-12-03T21:46:14

Hopefully if it turns out there's some good reason why we'd need to disable it, then we can provide some sort of backwards-compatibility flag (or at least a warning). --b. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ NFS maillist - NFS-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f< at >public.gmane.org https://lists.sourceforge.net/lists/listinfo/nfs _______________________________________________ Please note that nfs-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f< at >public.gmane.org is being discontinued. Please subscribe to linux-nfs-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org instead. http://vger.kernel.org/vger-lists.html#linux-nfs -- To unsubscribe from this list: send t

Re: [PATCH 6/6] NSM: Serialize SM_MON and SM_UNMON upcalls

J. Bruce Fields — 2008-12-03T21:41:42

... We also need to grep for places where sm_monitored is cleared, and think about what happens in a race between one of those cases and __nsm_monitor(). --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 5/6] NLM: Clean up nsm_monitor() call

J. Bruce Fields — 2008-12-03T21:28:59

Looks fine, but might prefer this slightly finer-grained: That might be a candidate to split out as a separate patch. Hm. Yeah, can't see how it could ever have been called with NULL--might be worth a changelog comment, though. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 3/6] NSM: Support IPv6 version of mon_name

Chuck Lever — 2008-12-03T21:24:38

On Dec 3, 2008, at 4:19 PM, Chuck Lever wrote: I'll clarify: a subsequent patch in my for-2.6.29 series that I haven't posted yet removes that use of nsm_use_hostnames. -- Chuck Lever chuck[dot]lever[at]oracle[dot]com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/6] NLM: Clean up nsm_monitor() call

Chuck Lever — 2008-12-03T21:20:16

On Dec 3, 2008, at 4:08 PM, J. Bruce Fields wrote: That's fine. I will probably end up reposting this series with your comments addressed, so I will split this one out. -- Chuck Lever chuck[dot]lever[at]oracle[dot]com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 3/6] NSM: Support IPv6 version of mon_name

Chuck Lever — 2008-12-03T21:19:25

Agreed. A subsequent patch in this series removes that particular use of nsm_use_hostnames. An audit of the other use(s) is a good future work item. Actually, "when lockd comes up" might be an appropriate moment to copy the value of this variable. Or a bug in rpc.statd or sm-notify... -- Chuck Lever chuck[dot]lever[at]oracle[dot]com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 4/6] NLM: Clean up nsm_monitor() call

J. Bruce Fields — 2008-12-03T21:08:43

All looks fine, thanks, however: I'd prefer to have a change in behavior split out into a separate patch from pure cleanup. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 3/6] NSM: Support IPv6 version of mon_name

J. Bruce Fields — 2008-12-03T21:01:50

At some point we need to grep for each read of nsm_use_hostnames and think about what would happen if it changed there. For example, the check of nsm_use_hostnames when searching for a match in nsm_find() could cause a spurious failure to find a host. If the nsm_find() came from nlm_host_rebooted(), we could fail to release locks from some dead host. Probably we should just forbid changing nsm_use_hostnames while the server is running or an nfs filesystem is mounted. Or, if that's not possible, allow changing the sysctl at any time, but only actually look at it (and store it) once on server startup or first mount (whichever comes first). As this requires a root user doing something wrong, fixing this bug probably isn't high priority enough to block the rest of the ipv6 patches, so we could make a note of the problem and move on for now. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majord

Re: [PATCH 2/6] NLM: Support IPv6 scope IDs innlm_display_address()

J. Bruce Fields — 2008-12-03T19:46:47

So something like just: sm_sticky : 1;/* don't unmonitor */ unsigned intsm_monitored : 1, sm_sticky : 1;/* don't unmonitor */ /* * This can hold 8 groups of colon-separated 4-hex-digit numbers, a * percent sign, and a scope id (at most 32 bits, in decimal, so): * 63 == 8*4 + 7 colons + 1 percent sign + at most 10 digits + padding: */ charsm_addrbuf[63];/* presentation address */ -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/6] NLM: Support IPv6 scope IDs innlm_display_address()

J. Bruce Fields — 2008-12-03T19:34:04

As always: what's the latest progress on getting this ipv6-address-display stuff into the core networking code? Anyway, looks fine, but looking forward to be able to delete some of this one day.... Just one more question below: Could you add a brief (1-line?) comment justifying the choice of 63? --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/6] NLM: Remove address eye-catcher buffers from nlm_host

Chuck Lever — 2008-12-03T19:10:03

It seemed like a small change (one chunk) and is tightly related to what this patch does, so I left it in. nsm_handles live longer than the nlm_hosts that reference them, so this shouldn't be a problem. -- Chuck Lever chuck[dot]lever[at]oracle[dot]com -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

nfsd fixes for 2.6.28

J. Bruce Fields — 2008-12-03T19:10:04

Please pull a few more small nfsd regression fixes for 2.6.28, from the for-2.6.28 branch at: git://linux-nfs.org/~bfields/linux.git for-2.6.28 Thanks to Jeff Moyer, Matthew Dodd, and others for testing. --b. Chuck Lever (1): NLM: client-side nlm_lookup_host() should avoid matching on srcaddr J. Bruce Fields (2): nfsd: clean up grace period on early exit nfsd: use of unitialized list head on error exit in nfs4recover.c Tom Tucker (1): Add a reference to sunrpc in svc_addsock fs/lockd/host.c | 3 ++- fs/lockd/svc.c | 1 + fs/nfsd/nfs4recover.c | 2 +- fs/nfsd/nfs4state.c | 1 + net/sunrpc/svcsock.c | 9 +++++++-- 5 files changed, 12 insertions(+), 4 deletions(-) -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/6] NLM: Remove address eye-catcher buffers fromnlm_host

J. Bruce Fields — 2008-12-03T19:03:32

As usual, whenever there's a "finally, ..." or "next, ..." I'd like a separate patch. Though this is distinct enough it's not really a problem to read. Leave it as is, OK. Hm, just checking: so the only place I can see h_nsmhandle set NULL is in nsm_unmonitor(), which is called only from nlm_destroy_host(), shortly before a kfree(host), but before a possible rpc_shutdown_client()--and doesn't look like an rpc task could be calling rpc_bind_host()? OK. And on the creation end, looks like h_nsmhandle is set on host creation. So that won't be a NULL deference. OK, looks fine. Applied. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: Is order for "exportfs -r" and rpc.mountd important?

J. Bruce Fields — 2008-12-03T17:25:18

So, I'm confused: what are the before-and-after orders of all three (modprobe, exportfs, and rpc.mountd)? What problems exactly? That does sound wrong. Perhaps rmtab should also be cleared on exportfs -r? I don't really understand how rmtab should work. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH/RFC] svcgssd always sets an infinite expiry onauthentication tokens etc.

J. Bruce Fields — 2008-12-02T23:23:44

Yes, definitely a leak. But it's funny, I have a strong memory of a conversation on one of these lists with someone who looked into this problem and found that even with some silly artificial tests that established as many contexts per second as possible, they weren't able to see significant memory consumption--but I can't find the thread now. Some limit on the size might also be nice depending on what the worst case looks like with expiries of about a day. --b. -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA< at >public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html

Re: [PATCH 0/3] NFS regression in 2.6.26?, "task blocked for more than 120 seconds"

Trond Myklebust — 2008-12-02T18:10:07

Are you seeing the same behaviour with 'netstat -t'?

[PATCH 0/3] AF_INET6 support for probe_bothports()

Chuck Lever — 2008-12-02T17:59:44

Hi Steve- Here's the next step. This small series of patches rewires probe_bothports() to support AF_INET6 addresses, now that the underlying functions can handle them. Since legacy code in other parts of the mount command still use probe_bothports() and the clnt_addr_t data type, I've added a new function call to do the IPv6 duties. The old API still exists and continues to support only AF_INET, but under the covers it calls the new code. Again, for the time being, this is used only for the legacy binary mount(2) interface. We will need this for umount later, and that is used to support both binary and text-based mounts. --- Chuck Lever (3): mount command: AF_INET6 support for probe_bothports() mount command: support AF_INET6 in probe_nfsport() and probe_mntport() mount command: full support for AF_INET6 addresses in probe_port() utils/mount/network.c | 178 +++++++++++++++++++++++++++++++++++++++---------- utils/mount/network.h | 3 + 2 files changed, 144 insertions(+), 37

[PATCH 3/3] mount command: AF_INET6 support for probe_bothports()

Chuck Lever — 2008-12-02T18:00:01

Introduce an AF_INET6 capable probe_bothports() API. This means replacing "struct sockaddr_in *" arguments with a "struct sockaddr *" and a socklen_t arguments. These functions often combine a "struct sockaddr_in" and a "struct pmap" into a single "clnt_addr_t" argument. Instead of modifying "clnt_addr_t" and all the legacy code that uses it, I'm going to create a new probe_bothports() API for the text-based mount command that takes a "struct sockaddr *" and sockaddr length, and leave the existing probe_bothports() interface, which takes "clnt_addr_t" arguments, for legacy use. Signed-off-by: Chuck Lever public.gmane.org> --- utils/mount/network.c | 86 +++++++++++++++++++++++++++++++++++++------------ utils/mount/network.h | 3 ++ 2 files changed, 68 insertions(+), 21 deletions(-) diff --git a/utils/mount/network.c b/utils/mount/network.c index 55b2cab..6a9a41a 100644 --- a/utils/mount/network.c +++ b/utils/mount/network.c < at >< at > -615,24 +615,49 < at >< at > static int nfs_