Skip to contentSkip to conversation

generic: test encryption key revocation during concurrent I/O

PATCH Superseded Eric Biggers 2 participants 3 messages Mar 15, 2017 fscrypt: file system level encryption support

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 generic: test encryption key revocation during concurrent I/O

From: Eric Biggers <[email protected]>

Add a test which revokes a keyring key while other processes are performing I/O on an encrypted file that was "unlocked" using that key. This crashes unpatched kernels with filesystem encryption enabled, so place it in the "dangerous" group.

[The patch to fix this is titled "fscrypt: remove broken support for detecting keyring key revocation" but hasn't been merged yet; we may want to wait a bit before merging this test.]

Cc Theodore Ts'o <[email protected]>
Cc Jaegeuk Kim <[email protected]>
Cc Richard Weinberger <[email protected]>
Cc Michael Halcrow <[email protected]>
---
 common/encrypt        |   8 ++++
 tests/generic/500     | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/500.out |   2 +
 tests/generic/group   |   1 +
 4 files changed, 120 insertions(+)
 create mode 100755 tests/generic/500
 create mode 100644 tests/generic/500.out
common/encrypt +8 -0
index f09104d1..85f71d5b 100644
@@ -144,3 +144,11 @@ _unlink_encryption_key()
 	local keyid=$($KEYCTL_PROG search @s logon $FSTYP:$keydesc)
 	$KEYCTL_PROG unlink $keyid >>$seqres.full
 }
+
+# Revoke an encryption key from the keyring, given its key descriptor.
+_revoke_encryption_key()
+{
+	local keydesc=$1
+	local keyid=$($KEYCTL_PROG search @s logon $FSTYP:$keydesc)
+	$KEYCTL_PROG revoke $keyid >>$seqres.full
+}
tests/generic/500 new file +109 -0

109 lines not on this page · raw message

tests/generic/500.out new file +2 -0
new file mode 100644
index 00000000..b46461f3
@@ -0,0 +1,2 @@
+QA output created by 500
+Didn't crash!
tests/generic/group +1 -0
index f0096bbc..f2c4755c 100644
@@ -420,3 +420,4 @@
 415 auto clone
 416 auto enospc
 417 auto quick shutdown log
+500 auto quick encrypt dangerous
Signature · 1 line
2.12.0.367.g23dc2f6d3c-goog
#1 lore raw
Quoted 10 lines · Eric Biggers

From: Eric Biggers <[email protected]>

Add a test which revokes a keyring key while other processes are performing I/O on an encrypted file that was "unlocked" using that key. This crashes unpatched kernels with filesystem encryption enabled, so place it in the "dangerous" group.

[The patch to fix this is titled "fscrypt: remove broken support for detecting keyring key revocation" but hasn't been merged yet; we may want to wait a bit before merging this test.]

Agreed.

Quoted 6 lines
Cc Theodore Ts'o <[email protected]>
Cc Jaegeuk Kim <[email protected]>
Cc Richard Weinberger <[email protected]>
Cc Michael Halcrow <[email protected]>

This looks good to me, and it crashed my 4.11-rc2 test host very quickly. Just one minor comment below.

Quoted 8 lines
---
 common/encrypt        |   8 ++++
 tests/generic/500     | 109 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/500.out |   2 +
 tests/generic/group   |   1 +
 4 files changed, 120 insertions(+)
 create mode 100755 tests/generic/500
 create mode 100644 tests/generic/500.out

...

tests/generic/500 line 51

Quoted 5 lines
tests/generic/500 new file +5 -0
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+_supported_fs ext4 f2fs

supported fs should be generic if it's in generic dir. If it only supports some specific filesystems, then it should go to shared dir.

Thanks,
Eryu

#2 lore raw
Quoted 9 lines · Eryu Guan

tests/generic/500 line 51

Quoted 5 lines
tests/generic/500 new file +5 -0
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+_supported_fs ext4 f2fs

supported fs should be generic if it's in generic dir. If it only supports some specific filesystems, then it should go to shared dir.

Indeed, it's supposed to be '_supported_fs generic'.

Thanks for reviewing!

- Eric