* [PATCH 1/2] tools/testing/nvdimm: disable labels for nfit_test.1
@ 2017-09-19 19:51 Dan Williams
2017-09-19 19:51 ` [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash Dan Williams
0 siblings, 1 reply; 4+ messages in thread
From: Dan Williams @ 2017-09-19 19:51 UTC (permalink / raw)
To: linux-nvdimm; +Cc: linux-kernel
Improve coverage of NVDIMM-N test scenarios by providing a test bus
incapable of label operations.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
tools/testing/nvdimm/test/nfit.c | 3 ---
1 file changed, 3 deletions(-)
diff --git a/tools/testing/nvdimm/test/nfit.c b/tools/testing/nvdimm/test/nfit.c
index d20791c3f499..bef419d4266d 100644
--- a/tools/testing/nvdimm/test/nfit.c
+++ b/tools/testing/nvdimm/test/nfit.c
@@ -1527,9 +1527,6 @@ static void nfit_test1_setup(struct nfit_test *t)
set_bit(ND_CMD_ARS_START, &acpi_desc->bus_cmd_force_en);
set_bit(ND_CMD_ARS_STATUS, &acpi_desc->bus_cmd_force_en);
set_bit(ND_CMD_CLEAR_ERROR, &acpi_desc->bus_cmd_force_en);
- set_bit(ND_CMD_GET_CONFIG_SIZE, &acpi_desc->dimm_cmd_force_en);
- set_bit(ND_CMD_GET_CONFIG_DATA, &acpi_desc->dimm_cmd_force_en);
- set_bit(ND_CMD_SET_CONFIG_DATA, &acpi_desc->dimm_cmd_force_en);
}
static int nfit_test_blk_do_io(struct nd_blk_region *ndbr, resource_size_t dpa,
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash
2017-09-19 19:51 [PATCH 1/2] tools/testing/nvdimm: disable labels for nfit_test.1 Dan Williams
@ 2017-09-19 19:51 ` Dan Williams
2017-09-19 22:06 ` Saldivar, Maurice A.
0 siblings, 1 reply; 4+ messages in thread
From: Dan Williams @ 2017-09-19 19:51 UTC (permalink / raw)
To: linux-nvdimm; +Cc: Vishal Verma, linux-kernel, stable, Maurice A. Saldivar
Maurice reports:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000028
IP: holder_class_store+0x253/0x2b0 [libnvdimm]
...while trying to reconfigure an NVDIMM-N namespace into 'sector' /
'btt' mode. The crash points to this line:
(gdb) li *(holder_class_store+0x253)
0x7773 is in holder_class_store (drivers/nvdimm/namespace_devs.c:1420).
1415 for (i = 0; i < nd_region->ndr_mappings; i++) {
1416 struct nd_mapping *nd_mapping = &nd_region->mapping[i];
1417 struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
1418 struct nd_namespace_index *nsindex;
1419
1420 nsindex = to_namespace_index(ndd, ndd->ns_current);
...where we are failing because ndd is NULL due to NVDIMM-N dimms not
supporting labels.
Long story short, default to the BTTv1 format in the label-less /
NVDIMM-N case.
Fixes: 14e494542636 ("libnvdimm, btt: BTT updates for UEFI 2.7 format")
Cc: <stable@vger.kernel.org>
Cc: Vishal Verma <vishal.l.verma@intel.com>
Reported-by: Maurice A. Saldivar <maurice.a.saldivar@hpe.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
drivers/nvdimm/namespace_devs.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/drivers/nvdimm/namespace_devs.c b/drivers/nvdimm/namespace_devs.c
index 1427a386a033..3e4d1e7998da 100644
--- a/drivers/nvdimm/namespace_devs.c
+++ b/drivers/nvdimm/namespace_devs.c
@@ -1417,6 +1417,15 @@ static int btt_claim_class(struct device *dev)
struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
struct nd_namespace_index *nsindex;
+ /*
+ * If any of the DIMMs do not support labels the only
+ * possible BTT format is v1.
+ */
+ if (!ndd) {
+ loop_bitmask = 0;
+ break;
+ }
+
nsindex = to_namespace_index(ndd, ndd->ns_current);
if (nsindex == NULL)
loop_bitmask |= 1;
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash
2017-09-19 19:51 ` [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash Dan Williams
@ 2017-09-19 22:06 ` Saldivar, Maurice A.
2017-09-19 22:10 ` Dan Williams
0 siblings, 1 reply; 4+ messages in thread
From: Saldivar, Maurice A. @ 2017-09-19 22:06 UTC (permalink / raw)
To: dan.j.williams@intel.com, linux-nvdimm@lists.01.org
Cc: linux-kernel@vger.kernel.org, stable@vger.kernel.org,
vishal.l.verma@intel.com
I applied this and the kernel patch and was successfully able to
configure an NVDIMM-N into sector mode. I do have a question about
specific behavior I observed. Before I was able to configure an NVDIMM
into any mode from any mode. With these patchs if I go to sector mode
and try going to memory or dax mode it fails with:
# ndctl create-namespace -f -e namespace0.0 -m memory
Error: namespace0.0: sector_size: 4096 not supported
failed to reconfigure namespace: Invalid argument
but if I go sector->raw-><mode> I can get into other modes fine. Was
this the intended behavior?
Sincerely,
Maurice
On Tue, 2017-09-19 at 12:51 -0700, Dan Williams wrote:
> Maurice reports:
>
> BUG: unable to handle kernel NULL pointer dereference at
> 0000000000000028
> IP: holder_class_store+0x253/0x2b0 [libnvdimm]
>
> ...while trying to reconfigure an NVDIMM-N namespace into 'sector' /
> 'btt' mode. The crash points to this line:
>
> (gdb) li *(holder_class_store+0x253)
> 0x7773 is in holder_class_store
> (drivers/nvdimm/namespace_devs.c:1420).
> 1415 for (i = 0; i < nd_region->ndr_mappings; i++) {
> 1416 struct nd_mapping *nd_mapping =
> &nd_region->mapping[i];
> 1417 struct nvdimm_drvdata *ndd =
> to_ndd(nd_mapping);
> 1418 struct nd_namespace_index *nsindex;
> 1419
> 1420 nsindex = to_namespace_index(ndd, ndd-
> >ns_current);
>
> ...where we are failing because ndd is NULL due to NVDIMM-N dimms not
> supporting labels.
>
> Long story short, default to the BTTv1 format in the label-less /
> NVDIMM-N case.
>
> Fixes: 14e494542636 ("libnvdimm, btt: BTT updates for UEFI 2.7
> format")
> Cc: <stable@vger.kernel.org>
> Cc: Vishal Verma <vishal.l.verma@intel.com>
> Reported-by: Maurice A. Saldivar <maurice.a.saldivar@hpe.com>
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> ---
> drivers/nvdimm/namespace_devs.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/drivers/nvdimm/namespace_devs.c
> b/drivers/nvdimm/namespace_devs.c
> index 1427a386a033..3e4d1e7998da 100644
> --- a/drivers/nvdimm/namespace_devs.c
> +++ b/drivers/nvdimm/namespace_devs.c
> @@ -1417,6 +1417,15 @@ static int btt_claim_class(struct device *dev)
> struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
> struct nd_namespace_index *nsindex;
>
> + /*
> + * If any of the DIMMs do not support labels the
> only
> + * possible BTT format is v1.
> + */
> + if (!ndd) {
> + loop_bitmask = 0;
> + break;
> + }
> +
> nsindex = to_namespace_index(ndd, ndd->ns_current);
> if (nsindex == NULL)
> loop_bitmask |= 1;
>
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash
2017-09-19 22:06 ` Saldivar, Maurice A.
@ 2017-09-19 22:10 ` Dan Williams
0 siblings, 0 replies; 4+ messages in thread
From: Dan Williams @ 2017-09-19 22:10 UTC (permalink / raw)
To: Saldivar, Maurice A.
Cc: linux-nvdimm@lists.01.org, linux-kernel@vger.kernel.org,
stable@vger.kernel.org, Verma, Vishal L
On Tue, Sep 19, 2017 at 3:06 PM, Saldivar, Maurice A.
<maurice.a.saldivar@hpe.com> wrote:
> I applied this and the kernel patch and was successfully able to
> configure an NVDIMM-N into sector mode. I do have a question about
> specific behavior I observed. Before I was able to configure an NVDIMM
> into any mode from any mode. With these patchs if I go to sector mode
> and try going to memory or dax mode it fails with:
>
> # ndctl create-namespace -f -e namespace0.0 -m memory
> Error: namespace0.0: sector_size: 4096 not supported
> failed to reconfigure namespace: Invalid argument
>
> but if I go sector->raw-><mode> I can get into other modes fine. Was
> this the intended behavior?
>
Nope, that's another bug. I'll fix that up.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2017-09-19 22:10 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-09-19 19:51 [PATCH 1/2] tools/testing/nvdimm: disable labels for nfit_test.1 Dan Williams
2017-09-19 19:51 ` [PATCH 2/2] libnvdimm, namespace: fix btt claim class crash Dan Williams
2017-09-19 22:06 ` Saldivar, Maurice A.
2017-09-19 22:10 ` Dan Williams
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox