-
Bug
-
Resolution: Fixed
-
Medium
-
None
-
None
-
3
-
9223372036854775807
in LU-13805 and LU-18006 the following code was introduced:
if (rw == WRITE) {
spin_lock(&sdio->csd_lock);
...
/* fault in the entire userspace iovec */
if (rw == WRITE) {
if (unlikely(ll_iov_iter_fault_in_readable(iter, count)))
I think it's a bad idea to repopulate userspace pages holding a spinlock - this can involve page allocation and I/O.
- is duplicated by
-
LU-19932 client write hang in osc_extent_wait()
-
- Resolved
-
- is related to
-
LU-21159 hybrid IO: buffered read()/write() fails with ENOMEM or EINVAL when the DIO bounce page pool is short
-
- Open
-
-
LU-20317 sanity/398u fails on aarch64 - incorrect PAGE_SIZE
-
- Resolved
-
-
DDN-6788 Loading...
- is related to
-
LU-19932 client write hang in osc_extent_wait()
-
- Resolved
-
- relates to failure
-
DDN-6837 Loading...
(1 relates to failure)