summaryrefslogtreecommitdiff
path: root/tools/perf/scripts/python/export-to-postgresql.py
diff options
context:
space:
mode:
authorNeilBrown <neil@brown.name>2025-03-19 14:01:33 +1100
committerChristian Brauner <brauner@kernel.org>2025-04-07 09:25:32 +0200
commit8ad9248471138e08a91d5e54b74d95d02f9518b2 (patch)
tree32e0ef72ff7d3e239aefe8b45f5784cb8e0b7844 /tools/perf/scripts/python/export-to-postgresql.py
parent5741909697a31cfb08e45d56b4211959fb791487 (diff)
nfsd: Use lookup_one() rather than lookup_one_len()
nfsd uses some VFS interfaces (such as vfs_mkdir) which take an explicit mnt_idmap, and it passes &nop_mnt_idmap as nfsd doesn't yet support idmapped mounts. It also uses the lookup_one_len() family of functions which implicitly use &nop_mnt_idmap. This mixture of implicit and explicit could be confusing. When we eventually update nfsd to support idmap mounts it would be best if all places which need an idmap determined from the mount point were similar and easily found. So this patch changes nfsd to use lookup_one(), lookup_one_unlocked(), and lookup_one_positive_unlocked(), passing &nop_mnt_idmap. This has the benefit of removing some uses of the lookup_one_len functions where permission checking is actually needed. Many callers don't care about permission checking and using these function only where permission checking is needed is a valuable simplification. This change requires passing the name in a qstr. Currently this is a little clumsy, but if nfsd is changed to use qstr more broadly it will result in a net improvement. Signed-off-by: NeilBrown <neil@brown.name> Link: https://lore.kernel.org/r/20250319031545.2999807-3-neil@brown.name Signed-off-by: Christian Brauner <brauner@kernel.org>
Diffstat (limited to 'tools/perf/scripts/python/export-to-postgresql.py')
0 files changed, 0 insertions, 0 deletions