[PATCH] Fix tests/test_device.py

[PATCH] Fix tests/test_device.py

[Bob Pearson]

Removed a test case which requires vendor_part_id to be non zero.
Per IBTA A3.3.1 VENDOR INFORMATION it is not required that the
vendor part ID field be set to a non-zero value:

	The following components are vendor specific: VendorID, DeviceID, De-
	vice Version, Subsystem VendorID, SubsystemID, ID String.

	The vendor places its IEEE assigned Organization Unique Identifier
	(OUI) in the VendorId field and *MAY PLACE ANY VALUE IN THE DEVICEID* and
	Device Version fields. The vendor may also provide an ASCII string of its
	choice in the ID String field.

	The Subsystem VendorID and SubsystemID provide additional informa-
	tion when a subsystem vendor uses components provided by other ven-
	dors. In this case the subsystem vendor provides its OUI in the Subsystem
	VendorID field and may specify any value in the SubsystemD field.
	A vendor that produces a generic controller (i.e., one that supports a stan-
	dard I/O protocol such as SRP), which does not have vendor specific de-
	vice drivers, may use the value of 0xFFFFFF in the VendorID field.
	However, such a value prevents the vendor from ever providing vendor
	specific drivers for the product.

Signed-off-by: Bob Pearson <rpearson@hpe.com>
---
 tests/test_device.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tests/test_device.py b/tests/test_device.py
index c23caae1..d552f402 100644
--- a/tests/test_device.py
+++ b/tests/test_device.py
@@ -86,7 +86,6 @@ class DeviceTest(PyverbsAPITestCase):
         assert attr.max_mr_size > PAGE_SIZE
         assert attr.page_size_cap >= PAGE_SIZE
         assert attr.vendor_id != 0
-        assert attr.vendor_part_id != 0
         assert attr.max_qp > 0
         assert attr.max_qp_wr > 0
         assert attr.max_sge > 0
-- 
2.25.1

Re: [PATCH] Fix tests/test_device.py

[Leon Romanovsky]

On Mon, Aug 24, 2020 at 08:23:44PM -0500, Bob Pearson wrote:
> Removed a test case which requires vendor_part_id to be non zero.
> Per IBTA A3.3.1 VENDOR INFORMATION it is not required that the
> vendor part ID field be set to a non-zero value:
>
> 	The following components are vendor specific: VendorID, DeviceID, De-
> 	vice Version, Subsystem VendorID, SubsystemID, ID String.
>
> 	The vendor places its IEEE assigned Organization Unique Identifier
> 	(OUI) in the VendorId field and *MAY PLACE ANY VALUE IN THE DEVICEID* and
> 	Device Version fields. The vendor may also provide an ASCII string of its
> 	choice in the ID String field.
>
> 	The Subsystem VendorID and SubsystemID provide additional informa-
> 	tion when a subsystem vendor uses components provided by other ven-
> 	dors. In this case the subsystem vendor provides its OUI in the Subsystem
> 	VendorID field and may specify any value in the SubsystemD field.
> 	A vendor that produces a generic controller (i.e., one that supports a stan-
> 	dard I/O protocol such as SRP), which does not have vendor specific de-
> 	vice drivers, may use the value of 0xFFFFFF in the VendorID field.
> 	However, such a value prevents the vendor from ever providing vendor
> 	specific drivers for the product.
>
> Signed-off-by: Bob Pearson <rpearson@hpe.com>
> ---
>  tests/test_device.py | 1 -
>  1 file changed, 1 deletion(-)


Thanks, I'll pick to rdma-core later this week if no objections come.