Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 638819

Summary: FEAT: port /kernel/misc/amtu to v7
Product: [Retired] Red Hat Hardware Certification Program Reporter: Qian Cai <qcai>
Component: Test Suite (tests)Assignee: Caspar Zhang <czhang>
Status: CLOSED DUPLICATE QA Contact: Red Hat Kernel QE team <kernel-qe>
Severity: medium Docs Contact:
Priority: low    
Version: 1.2CC: gnichols, qcai, rlandry, sdenham
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-10-24 06:06:44 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
amtu patch
none
New test: amtu
none
0001-add-kernel-tier-tags.patch
none
0002-re-run-required-rpm-check-after-tests-added.patch
none
0003-port-amtu-test-from-Kernel-Tier-1-tests.patch
none
updated memory patch based on the meeting result
czhang: review?
amtu supervisor -> v7/core test as qa-only czhang: review?

Description Qian Cai 2010-09-30 03:11:39 UTC
http://wiki.test.redhat.com/Kernel/TierTestsV7Overlap
# /kernel/misc/amtu

    * Memory
          o Randomly writes to areas of memory and then reads the memory back to ensure the values written remained unchanged. 
    * Memory Separation
          o Ensures that user space processes cannot read and write to areas of memory used by Video RAM, kernel code, etc. 
    * I/O Controller - Network
          o Verifies random data transmitted is also the data received for each configured network device. Only ethernet and token ring devices that are configured and up are checked. Async devices are not checked. 
    * I/O Controller - Disk
          o Verifies that information written to disks remains unchanged. Only SCSI and IDE controllers associated with mounted filesystems are checked. 
    * Supervisor Mode Instructions
          o Ensures that the enforcement of the property that privileged instructions should only be in supervisor mode is still in effect. The set privileged instructions tested to confirm this is architecture dependant.

Comment 1 Rob Landry 2010-09-30 14:06:56 UTC
This should be broken up into individual requests to avoid comments like:

Looking @ the above, the memory item above sounds redundant to the v7 memory test, the memory separation looks new, network looks redundant, disk looks redundant, supervisor mode lokks new.

...causing confusion in just one bug. :)

Comment 2 Qian Cai 2010-10-05 11:51:27 UTC
Those kernel tier tests probably better grouped together called kernel test, or merge into existing tests separately as you mentioned.

Comment 3 Rob Landry 2010-11-19 20:00:13 UTC
This will also need the infrastructure in v7 to support hwcert required tests and qa test case tests.  Marking as 1.3+.  Prsuming QA will own getting the tests ported as appropriate with discussions with gnichols.

Comment 4 Caspar Zhang 2010-12-03 04:19:54 UTC
(In reply to comment #2)
> Those kernel tier tests probably better grouped together called kernel test, or
> merge into existing tests separately as you mentioned.

I prefer merge them into existed tests.

according to amtu helping:

Usage: amtu [-dmsinph]
d      Display debug messages
m      Execute Memory Test
s      Execute Memory Separation Test
i      Execute I/O Controller - Disk Test
n      Execute I/O Controller - Network Test
p      Execute Supervisor Mode Instructions Test
h      Display help message

I port mem separation to memory test, supervisor mode instruction test to core test. The other tests are duplicated with existed v7 tests.

Comment 5 Caspar Zhang 2010-12-03 04:21:27 UTC
Created attachment 464493 [details]
amtu patch

Comment 6 Rob Landry 2010-12-03 15:38:38 UTC
Are these additions testing the hardware or verifying the software?

Comment 7 Caspar Zhang 2010-12-03 15:49:09 UTC
testing the hardware, I think. One for testing memory and the other for CPU.

Comment 8 Rob Landry 2010-12-03 16:01:44 UTC
What is the additional runtime required and is this being proposed to expand the hwcert testing requirements to make available the test?

Comment 9 Caspar Zhang 2010-12-03 16:13:06 UTC
less than 1 min needed, memory seprating and supervisor mode testing finishes very soon.

Comment 10 Rob Landry 2010-12-06 15:41:45 UTC
And the technical argument for if either test should fail on a system regarding it's ineligibility to be certified would be?

Comment 11 Caspar Zhang 2011-01-12 10:02:52 UTC
(In reply to comment #10)
> And the technical argument for if either test should fail on a system regarding
> it's ineligibility to be certified would be?

I'm not sure of "technical argument", but if the test fail, the return code is a non-zero value. Since there are some duplicated functions in amtu, I only ported mem-sep and cpu test.

Would you mind explaining "technical argument"?

Comment 12 Rob Landry 2011-01-12 15:18:29 UTC
Meaning to use it as pass/fail criteria in certification we need to be able to explain what the system is failing to do and why that is a technical problem for customers.

Comment 13 Caspar Zhang 2011-01-23 15:49:08 UTC
Created attachment 474828 [details]
New test: amtu

Re-written the new test case:

New test basic information:

Runtime: 10m

Architecture: all

Function:
* Memory: Randomly writes to areas of memory and then reads the memory
back to ensure the values written remained unchanged.
* Memory Separation: Ensures that user space processes cannot read and
write to areas of memory used by Video RAM, kernel code, etc.
* I/O Controller - Network: Verifies random data transmitted is also the
data received for each configured network device. Only ethernet and
token ring devices that are configured and up are checked. Async devices
are not checked.
* I/O Controller - Disk: Verifies that information written to disks
remains unchanged. Only SCSI and IDE controllers associated with mounted
filesystems are checked.
* Supervisor Mode Instructions: Ensures that the enforcement of the
property that privileged instructions should only be in supervisor mode
is still in effect. The set privileged instructions tested to confirm
this is architecture dependant.

PASS/FAIL Criteria: if the test returns non-zero value, one or more
sub-tests get failed. Else if all sub-tests passes, the test returns 0.

Technical Arugment: If one or more sub-tests fail in testing, the
corresponding hardware (memory, network device, disk, CPU) may be
working incorrectly. Or there may be something wrong with the kernel
functions. Since it's a existed test case in Kernel Tier 1 which is used
to test kernel usablity, it's a critical issue if the test fails.

Comment 21 Caspar Zhang 2011-02-11 12:54:07 UTC
Created attachment 478233 [details]
0001-add-kernel-tier-tags.patch

Comment 22 Caspar Zhang 2011-02-11 12:54:43 UTC
Created attachment 478234 [details]
0002-re-run-required-rpm-check-after-tests-added.patch

Comment 23 Caspar Zhang 2011-02-11 12:55:30 UTC
Created attachment 478235 [details]
0003-port-amtu-test-from-Kernel-Tier-1-tests.patch

Comment 27 Caspar Zhang 2011-03-25 09:30:38 UTC
Created attachment 487509 [details]
updated memory patch based on the meeting result

Updated a patch based our discussion result in the meeting. This merged mmap and amtu patch together to be added as amtu/memory qa-only test.

Comment 28 Caspar Zhang 2011-03-25 09:31:22 UTC
Created attachment 487510 [details]
amtu supervisor -> v7/core test as qa-only

Comment 29 Caspar Zhang 2011-10-24 06:06:44 UTC

*** This bug has been marked as a duplicate of bug 715215 ***