progs: tests: Avoid multidevice test on 32bit platforms
PATCH 1/1 progs: tests: Avoid multidevice test on 32bit platforms
From: Marcos Paulo de Souza <[email protected]>
This test uses truncate utility to create a 6E file but this fails currently fails for PPC32[1], but it will also fail to other 32bit platforms, so skip this test in these platforms.
[1]: https://github.com/kdave/btrfs-progs/issues/192
---
I couldn't find a way to make truncate to use O_LARGEFILE option. I though even to write a helper to open the file using this flag and later on call ftruncate, but isn't it one more helper to maintain?
If there are better ideas to avoid skipping this test, please let me know!
tests/common | 15 +++++++++++++++ tests/mkfs-tests/018-multidevice-overflow/test.sh | 1 + 2 files changed, 16 insertions(+)
tests/common +15 -0
@@ -580,6 +580,21 @@ check_min_kernel_version() return 0 } +check_32bit_machine() +{ + local msg + + msg="$1" + if [ -z "$msg" ]; then + msg="Skipping test on 32bit machines" + fi + + long_bit=$(getconf LONG_BIT) + if [ $long_bit -eq 32 ]; then + _not_run "$msg" + fi +} + # how many files to create. DATASET_SIZE=50
tests/mkfs-tests/018-multidevice-overflow/test.sh +1 -0
@@ -5,6 +5,7 @@ source "$TEST_TOP/common" check_prereq mkfs.btrfs check_prereq btrfs +check_32bit_machine "32bit machines can't handle 6E file sizes" setup_root_helper prepare_test_dev
Signature · 1 line
2.24.0
Quoted 8 lines · Marcos Paulo de Souza
From: Marcos Paulo de Souza <[email protected]>
This test uses truncate utility to create a 6E file but this fails currently fails for PPC32[1], but it will also fail to other 32bit platforms, so skip this test in these platforms.
Quoted 8 lines
Signed-off-by Marcos Paulo de Souza <[email protected]>---
I couldn't find a way to make truncate to use
O_LARGEFILEoption. I though even to write a helper to open the file using this flag and later on call ftruncate, but isn't it one more helper to maintain?If there are better ideas to avoid skipping this test, please let me know!
What about testing the first 6E file allocation?
If the first failed, then it means the system can't handle, and we skip the test.
Skipping the whole 32bit systems looks overkilled to me.
Thanks,
Qu
Quoted 44 lines
tests/common | 15 +++++++++++++++ tests/mkfs-tests/018-multidevice-overflow/test.sh | 1 + 2 files changed, 16 insertions(+)tests/common +15
-0@@ -580,6 +580,21 @@ check_min_kernel_version() return 0 } +check_32bit_machine() +{ + local msg + + msg="$1" + if [ -z "$msg" ]; then + msg="Skipping test on 32bit machines" + fi + + long_bit=$(getconf LONG_BIT) + if [ $long_bit -eq 32 ]; then + _not_run "$msg" + fi +} + # how many files to create. DATASET_SIZE=50tests/mkfs-tests/018-multidevice-overflow/test.sh +1
-0@@ -5,6 +5,7 @@ source "$TEST_TOP/common" check_prereq mkfs.btrfs check_prereq btrfs +check_32bit_machine "32bit machines can't handle 6E file sizes" setup_root_helper prepare_test_dev