Kernel commit 52f019d43c22 ("block: add a hard-readonly flag to struct
gendisk") broke the read-only management test, by fixing the broken
behavior that BLKROSET could make a block device read-write even when the
disk is read-only. The fix (see Link:) propagates changes of the region
read-only state to the underlying disk. Add ndctl_region_set_ro() ahead of
BLKROSET so that BLKROSET does not conflict the block_device state with the
disk state.Link: http://lore.kernel.org/r/161534060720.528671.2341213328968989192.st...@dwillia2-desk3.amr.corp.intel.com Reported-by: kernel test robot <l...@intel.com> Reported-by: Vishal Verma <vishal.l.ve...@intel.com> Signed-off-by: Dan Williams <dan.j.willi...@intel.com> --- test/libndctl.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/test/libndctl.c b/test/libndctl.c index faa308ade8fc..391b94086dae 100644 --- a/test/libndctl.c +++ b/test/libndctl.c @@ -1541,6 +1541,7 @@ static int validate_bdev(const char *devname, struct ndctl_btt *btt, struct ndctl_pfn *pfn, struct ndctl_namespace *ndns, struct namespace *namespace, void *buf) { + struct ndctl_region *region = ndctl_namespace_get_region(ndns); char bdevpath[50]; int fd, rc, ro; @@ -1578,6 +1579,13 @@ static int validate_bdev(const char *devname, struct ndctl_btt *btt, } ro = 0; + rc = ndctl_region_set_ro(region, ro); + if (rc < 0) { + fprintf(stderr, "%s: ndctl_region_set_ro failed\n", devname); + rc = -errno; + goto out; + } + rc = ioctl(fd, BLKROSET, &ro); if (rc < 0) { fprintf(stderr, "%s: BLKROSET failed\n", @@ -1605,8 +1613,16 @@ static int validate_bdev(const char *devname, struct ndctl_btt *btt, rc = -ENXIO; goto out; } + + rc = ndctl_region_set_ro(region, namespace->ro); + if (rc < 0) { + fprintf(stderr, "%s: ndctl_region_set_ro reset failed\n", devname); + rc = -errno; + goto out; + } + rc = 0; - out: +out: close(fd); return rc; } _______________________________________________ Linux-nvdimm mailing list -- linux-nvdimm@lists.01.org To unsubscribe send an email to linux-nvdimm-le...@lists.01.org
