lvm2/0095-filesystem-fix-nofs-early-return-checking-wrong-stru.patch
Marian Csontos 0d41e7e8af Additional patches for 9.9.0 lvm2
Patches from upstream up to 2.03.41.

Resolves: RHEL-174324
2026-06-04 21:29:42 +02:00

40 lines
1.2 KiB
Diff

From 80326bcb7fa8638fd14ce8900a30e59b9c9dc2f4 Mon Sep 17 00:00:00 2001
From: Zdenek Kabelac <zkabelac@redhat.com>
Date: Fri, 3 Apr 2026 13:32:46 +0200
Subject: [PATCH 095/211] filesystem: fix nofs early-return checking wrong
struct
fs_get_info() checked fsi->nofs (caller's zeroed output parameter)
instead of info.nofs (the local struct populated by fs_get_blkid).
The early-return for no-filesystem case never triggered, causing
unnecessary crypto_LUKS, mount, and resize checks.
Result was still correct because *fsi = info copy at line 342
propagated nofs to the caller, but the early exit was dead code.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
(cherry picked from commit 3606b9749042fb02117bebb0ba84f4e95ebeef57)
---
lib/device/filesystem.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/lib/device/filesystem.c b/lib/device/filesystem.c
index adb8f0174..87f97130d 100644
--- a/lib/device/filesystem.c
+++ b/lib/device/filesystem.c
@@ -284,8 +284,10 @@ int fs_get_info(struct cmd_context *cmd, struct logical_volume *lv, struct fs_in
return 0;
}
- if (fsi->nofs)
+ if (info.nofs) {
+ fsi->nofs = 1;
return 1;
+ }
/*
* If there's a LUKS dm-crypt layer over the LV, then
--
2.54.0