Get rid of static Atffiles and let bsd.test.mk generate them on the fly.

add test for PR kern/43503

Add tests for LFS resize and new fcntl commands.

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.

Add skipped test for in-kernel roll-forward agent

Centralize the base rump libraries into a variable used by all the other
Makefiles so that we can make changes to it centrally as needed and have
less mess. Fixes the sun2 build that needs rumpvfs after librump after
the latest changes.

branches: 1.1.46;
add test for PR kern/43503

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.

Use the default timeout on the autoclean tests to avoid false positives
on slower architectures.

Add in-kernel cleaner. Move cleaning-related routines into their own file.
Document the fcntl calls better in lfs.h. Bounds-check inode numbers in
segment parsing functions. Use the in-kernel cleaner in a greedy mode
for tests in tests/fs/vfs, and add a test of the kernel cleaner.

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

Use default test timeout (3min) for now for all this tests.
The timeouts cause artificial failures on many slower test setups
and the tests run sucessfull if given enough time.

Example from an sh3 machine:

t_fcntl (1/1): 4 test cases
cleanseg32: [330.482904s] Passed.
cleanseg64: [321.895482s] Passed.
coalesce32: [20.214675s] Passed.
coalesce64: [20.613933s] Passed.

We may reduce timeouts again after we have working logs from various
architectures.

Make the test files smaller so that they write more quickly. To avoid
a timeout on slower platforms reported by martin@.

Add tests for LFS resize and new fcntl commands.

Update the orphan test to use dumplfs, avoiding false negatives.
Add post-test fsck to all the LFS tests.

(This is the intended commit comment for the previous revision.)

sys/ufs/lfs/lfs_alloc.c

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.

return return atf_no_error() instead of 0 for consistency.

suggested by moritzbuhl in https://github.com/NetBSD/src/pull/11/

branches: 1.7.14;
Don't play with "../.." in includes for h_macros.h; deal with it centrally.
Minor fixes.

branches: 1.6.26;
make the timeouts a bit more generous for really slow systems

branches: 1.5.2;
double the timeout for slow hosts (from 1 to 2 seconds)

Adjusts tests after import of atf-0.12:

- The use.fs property is gone.
- Mark the tests/fs/t_create:attrs test as broken when using the default
unprivileged-user:_atf setting. This probably deserves a fix somehow
but I'm not sure at this point.

Re-add PR number and expected failure condition (as a comment, because the
problem is resolved). Per pooka's request.

Replace custom xfail settings with the new expected failure support in
atf 0.10.

add test for PR kern/43503

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

Add tests for LFS resize and new fcntl commands.

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

Convert the rfw64 test to expect success, following sys/ufs/lfs/lfs_rfw.c
rev 1.39.

Convert t_rfw:rfw64 test case to expect-fail.

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.

Add extra checks, and report intermediate status as a debugging aid.
Separate tests for 32-bit LFS (expect fail) and 64-bit (skip, unable to test).

Nix trailing whitespace.

Expand test to cover more failure cases. Change from skipped to expect fail
in anticipation of working roll-forward code.

Add skipped test for in-kernel roll-forward agent

Update the orphan test to use dumplfs, avoiding false negatives.
Add post-test fsck to all the LFS tests.

(This is the intended commit comment for the previous revision.)

sys/ufs/lfs/lfs_alloc.c

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

Use a pseudorandom number with a well-known seed to provide file data
pattern, rather than a pattern that repeats every 256 bytes. Write
pattern files to the test file system in blocks rather than one byte at a
time, significantly speeding up the test.

Explicitly specify the segment size so the future tests can use the value.
The specified size is the same that was previously auto-selected, so there
is no effective change.

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.

Add a seed parameter to differentiate between files, and correct the logic
for appending files so that the resultant file matches the expectation.
Pass the t_rfw tests again.

s/creat /create/ in few more comments and error log messages.

Explicitly specify the segment size so the future tests can use the value.
The specified size is the same that was previously auto-selected, so there
is no effective change.

Add test cases to show whether newfs_lfs creates a file system that
fsck_lfs can understand; and whether orphans (files still in use, after
removal, at the time of a system crash) are removed at fs mount time.