Fixed ENOSPC issues with zero-granularity blocks

Result of testing on zero-granularity blocks, where the prog size and
read size equals the block size. This represents SD cards and other
traditional forms of block storage where we don't really get a benefit
from the metadata logging.

Unfortunately, since updates in both are tested by the same script,
we can't really use simple bash commands. Added a more complex
script to simulate corruption. Fortunately this should be more robust
than the previous solutions.

The main fixes were around corner cases where the commit logic fell
apart when it didn't have room to complete commits, but these were
fixable in the current design.
This commit is contained in:
Christopher Haster
2018-07-30 21:12:00 -05:00
parent 105907ba66
commit 1a58ba799c
4 changed files with 126 additions and 66 deletions

View File

@@ -59,7 +59,7 @@ tests/test.py << TEST
lfs_rename(&lfs, "b/hello", "c/hello") => 0;
lfs_unmount(&lfs) => 0;
TEST
truncate -s-7 blocks/6
tests/corrupt.py blocks/{6,7}
tests/test.py << TEST
lfs_mount(&lfs, &cfg) => 0;
lfs_dir_open(&lfs, &dir[0], "b") => 0;
@@ -86,8 +86,8 @@ tests/test.py << TEST
lfs_rename(&lfs, "c/hello", "d/hello") => 0;
lfs_unmount(&lfs) => 0;
TEST
truncate -s-7 blocks/8
truncate -s-7 blocks/a
tests/corrupt.py blocks/{8,9}
tests/corrupt.py blocks/{a,b}
tests/test.py << TEST
lfs_mount(&lfs, &cfg) => 0;
lfs_dir_open(&lfs, &dir[0], "c") => 0;
@@ -166,7 +166,7 @@ tests/test.py << TEST
lfs_rename(&lfs, "b/hi", "c/hi") => 0;
lfs_unmount(&lfs) => 0;
TEST
truncate -s-7 blocks/7
tests/corrupt.py blocks/{6,7}
tests/test.py << TEST
lfs_mount(&lfs, &cfg) => 0;
lfs_dir_open(&lfs, &dir[0], "b") => 0;
@@ -193,8 +193,8 @@ tests/test.py << TEST
lfs_rename(&lfs, "c/hi", "d/hi") => 0;
lfs_unmount(&lfs) => 0;
TEST
truncate -s-7 blocks/9
truncate -s-7 blocks/b
tests/corrupt.py blocks/{8,9}
tests/corrupt.py blocks/{a,b}
tests/test.py << TEST
lfs_mount(&lfs, &cfg) => 0;
lfs_dir_open(&lfs, &dir[0], "c") => 0;