Messages in this thread | | | From | Vitaly Kuznetsov <> | Subject | [PATCH v3 0/4] lib/string_helpers: fix precision issues and introduce tests | Date | Thu, 29 Oct 2015 17:30:46 +0100 |
| |
Linux always lies about your storage size when it has 4k sectors and its size is big enough. E.g. a device with 8192 4k sectors will be reported as "32.7 MB/32 MiB" while "33.5 MB/32 MiB" is expected. This series is supposed to fix the issue by fixing calculation precision in string_get_size() for all possible inputs.
PATCH 1/4 is a preparatory change, PATCH 2/4 adds additional protection against blk_size=0 (nobody is supposed to call string_get_size() with with blk_size=0, but better safe than sorry), PATCH 3/4 re-factors string_get_size() fixing the issue, PATCH 4/4 introduces tests for string_get_size().
PATCH 4/4 was previously sent as part of "lib/string_helpers.c: fix infinite loop in string_get_size()" series but it is still not merged upstream. In this submission I improve it and add additional tests to it.
Changes since v2: - Separate blk_size check from Patch 3/4 to new Patch 2/4 [Andy Shevchenko] - Slightly change the algorithm in Patch 3/4 [Rasmus Villemoes]
Vitaly Kuznetsov (4): lib/string_helpers: change blk_size to u32 for string_get_size() interface lib/string_helpers.c: protect string_get_size() against blk_size=0 lib/string_helpers.c: don't lose precision in string_get_size() lib/test-string_helpers.c: add string_get_size() tests
include/linux/string_helpers.h | 2 +- lib/string_helpers.c | 38 ++++++++++++++------------- lib/test-string_helpers.c | 58 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 79 insertions(+), 19 deletions(-)
-- 2.4.3
| |