Attachment 1473944 Details for Bug 1612914 – Patch - mark removed cache entry negative

[patch] Patch - mark removed cache entry negative

autofs-5.1.4-mark-removed-cache-entry-negative.patch (text/plain), 2.17 KB, created by Ian Kent on 2018-08-07 10:14:35 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Ian Kent

Created: 2018-08-07 10:14:35 UTC

Size: 2.17 KB

patch

obsolete

>autofs-5.1.4 - mark removed cache entry negative
>
>From: Ian Kent <raven@themaw.net>
>
>When re-reading a map entries that have been removed are detected
>and deleted from the map entry cache by lookup_prune_cache().
>
>If a removed map entry is mounted at the time lookup_prune_cache()
>is called the map entry is skippedi. This is done becuase the next
>lookup (following the mount expire, which needs the cache entry to
>remain) will detect the stale cache entry and a map update done
>resulting in the istale entry being removed.
>
>But if a map re-read is performed while the cache entry is mounted
>the cache will appear to up to date so the removed entry will remain
>valid even after it has expired.
>
>To cover this case it's sufficient to mark the mounted cache entry
>negative during the cache prune which prevent further lookups from
>using the stale entry.
>
>Signed-off-by: Ian Kent <raven@themaw.net>
>---
> CHANGELOG       |    1 +
> daemon/lookup.c |   12 +++++++++---
> 2 files changed, 10 insertions(+), 3 deletions(-)
>
>--- autofs-5.0.7.orig/CHANGELOG
>+++ autofs-5.0.7/CHANGELOG
>@@ -302,6 +302,7 @@
> - add units After line to include statd service.
> - use systemd sd_notify() at startup.
> - fix update_negative_cache() map source usage.
>+- mark removed cache entry negative.
> 
> 25/07/2012 autofs-5.0.7
> =======================
>--- autofs-5.0.7.orig/daemon/lookup.c
>+++ autofs-5.0.7/daemon/lookup.c
>@@ -1411,15 +1411,21 @@ void lookup_prune_one_cache(struct autof
> 		if (valid && valid->mc == mc) {
> 			 /*
> 			  * We've found a map entry that has been removed from
>-			  * the current cache so it isn't really valid.
>+			  * the current cache so it isn't really valid. Set the
>+			  * mapent negative to prevent further mount requests
>+			  * using the cache entry.
> 			  */
>+			debug(ap->logopt, "removed map entry detected, mark negative");
>+			if (valid->mapent) {
>+				free(valid->mapent);
>+				valid->mapent = NULL;
>+			}
> 			cache_unlock(valid->mc);
> 			valid = NULL;
> 		}
> 		if (!valid &&
> 		    is_mounted(_PATH_MOUNTED, path, MNTS_REAL)) {
>-			debug(ap->logopt,
>-			      "prune check posponed, %s mounted", path);
>+			debug(ap->logopt, "prune posponed, %s mounted", path);
> 			free(key);
> 			free(path);
> 			continue;

Actions: View | Diff

Attachments on bug 1612914: 1473944