Hi Calum,
Hi Petr,
There are two sets of changes here, for NFS client, and NFS server.
The NFS client changes have already been backported from v6.9 all the way to v5.4.
Here, Chuck is discussing the NFS server changes (and others), which were not backported from v6.9 (actually, a few were, but only to v6.8).
Thanks for info! Now I'll see the patchset "Make nfsd stats visible in network
ns" [1]. kernelnewbies [2] starts with d98416cc2154 ("nfsd: rename
NFSD_NET_* to NFSD_STATS_*"), the others are probably some preparation commits.
Anyway, I'll update the patch with NFS server patchset.
Kind regards,
Petr
[1] https://lore.kernel.org/linux-nfs/cover.1706283433.git.josef@toxicpanda.com/
[2] https://kernelnewbies.org/Linux_6.9#File_systems
Thanks,
Calum.
Sent from Outlook for Androidhttps://aka.ms/AAb9ysg
________________________________
From: Petr Vorel pvorel@suse.cz
Sent: Wednesday, August 14, 2024 8:45:59 AM
To: cel@kernel.org cel@kernel.org
Cc: stable@vger.kernel.org stable@vger.kernel.org; linux-nfs@vger.kernel.org linux-nfs@vger.kernel.org; Sherry Yang sherry.yang@oracle.com; Calum Mackay calum.mackay@oracle.com; kernel-team@fb.com kernel-team@fb.com; Chuck Lever III chuck.lever@oracle.com; Cyril Hrubis chrubis@suse.cz; ltp@lists.linux.it ltp@lists.linux.it
Subject: Re: [PATCH 6.6.y 00/12] Backport "make svc_stat per-net instead of global"
Hi Chuck,
Following up on:
https://lore.kernel.org/linux-nfs/d4b235df-4ee5-4824-9d48-e3b3c1f1f4d1@oracl...
Here is a backport series targeting origin/linux-6.6.y that closes
the information leak described in the above thread. It passes basic
NFSD regression testing.
Thank you for handling this! The link above mentions that it was already
backported to 5.4 and indeed I see at least d47151b79e322 ("nfs: expose
/proc/net/sunrpc/nfs in net namespaces") is backported in 5.4, 5.10, 5.15, 6.1.
And you're now preparing 6.6. Thus we can expect the behavior changed from
5.4 kernels.
I wonder if we consider this as a fix, thus expect any kernel newer than 5.4
should backport all these 12 patches.
Or, whether we should relax and just check if version is higher than the one
which got it in stable/LTS (e.g. >= 5.4.276 || >= 5.10.217 ...). The question is
also if enterprise distros will take this patchset.
BTW We have in LTP functionality which points as a hint to kernel fixes. But
it's usually a single commit. I might need to list all.
Kind regards,
Petr
Review comments welcome.
Chuck Lever (2):
NFSD: Rewrite synopsis of nfsd_percpu_counters_init()
NFSD: Fix frame size warning in svc_export_parse()
Josef Bacik (10):
sunrpc: don't change ->sv_stats if it doesn't exist
nfsd: stop setting ->pg_stats for unused stats
sunrpc: pass in the sv_stats struct through svc_create_pooled
sunrpc: remove ->pg_stats from svc_program
sunrpc: use the struct net as the svc proc private
nfsd: rename NFSD_NET_* to NFSD_STATS_*
nfsd: expose /proc/net/sunrpc/nfsd in net namespaces
nfsd: make all of the nfsd stats per-network namespace
nfsd: remove nfsd_stats, make th_cnt a global counter
nfsd: make svc_stat per-network namespace instead of global
fs/lockd/svc.c | 3 --
fs/nfs/callback.c | 3 --
fs/nfsd/cache.h | 2 -
fs/nfsd/export.c | 32 ++++++++++----
fs/nfsd/export.h | 4 +-
fs/nfsd/netns.h | 25 +++++++++--
fs/nfsd/nfs4proc.c | 6 +--
fs/nfsd/nfs4state.c | 3 +-
fs/nfsd/nfscache.c | 40 ++++-------------
fs/nfsd/nfsctl.c | 16 +++----
fs/nfsd/nfsd.h | 1 +
fs/nfsd/nfsfh.c | 3 +-
fs/nfsd/nfssvc.c | 14 +++---
fs/nfsd/stats.c | 54 ++++++++++-------------
fs/nfsd/stats.h | 88 ++++++++++++++------------------------
fs/nfsd/vfs.c | 6 ++-
include/linux/sunrpc/svc.h | 5 ++-
net/sunrpc/stats.c | 2 +-
net/sunrpc/svc.c | 39 +++++++++++------
19 files changed, 163 insertions(+), 183 deletions(-)