Bug 815295
| Summary: | [whql][netkvm]NDISTest6.5, "Glitch free" subtest always failed in Windows 7 64/32 bits and windows 2008 R2 | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Min Deng <mdeng> |
| Component: | virtio-win | Assignee: | Yvugenfi <yvugenfi> |
| Status: | CLOSED ERRATA | QA Contact: | Virtualization Bugs <virt-bugs> |
| Severity: | high | Docs Contact: | |
| Priority: | high | ||
| Version: | 6.3 | CC: | acathrow, bcao, bsarathy, dyasny, juzhang, lijin, michen, rhod, vrozenfe |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: |
Cause: Driver was failing WHQL certification on Windows 7 because it disregarded the request from OS to indicate only a certain number of packets during one DPC.
Consequence: WHQL test failure. Outside of test environment - OS might fail to moderate the driver workload for RX path.
Fix: Implement packet indication moderation.
Result: WHQL certification for Windows 7 pass.
|
Story Points: | --- |
| Clone Of: | Environment: | ||
| Last Closed: | 2013-02-21 10:38:36 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Min Deng
2012-04-23 10:09:16 UTC
Please attach the logs, especially of the failing test (Glitch free). Created attachment 579733 [details]
One cpk and error logs were in attachment.
Hi all, The issues was reproduced windows7 32 bits guest too.Uploaded the latest cpk and logs to the bug.Thanks. Created attachment 579771 [details]
Win7-Error-Log
windows 7 64/32 error log are in the attachment
Postponed to 6.4, and removed the blocker flag. We will use the Vista drivers for Win7 and 2K8R2, just like we did in the past. Yan Fixed the bug, still testing, but it is too delicate for 6.3. Fix: http://git.engineering.redhat.com/?p=users/yvugenfi/internal-kvm-guest-drivers-windows/.git;a=commit;h=44e34181c9774dfa1a3e41dd1a7c4a3d7b75215b Request for test: 1. Pass glitch free test 2. Test long (at least several hours) transfer scenarios (netperf and iperf). 3. Decrease "TestOnly.RXThrottle" driver parameter to 10 (from device manager -> advanced tab). Run long transfer scenarios. Additional info: 1 Test out of 30 failed. In failed cases the test was complaining that we were requested to indicate only 10 packets and indicated more. According to driver printout we were never requested to indicate less than 100 packets. If this issue will detected in WHQL environment - let's gather the logs and approach MS. During tests - test guest also as a server of iperf and netperf. Another issue with this test - from time to time test fails with "divide by zero exception". This failure was present even before my recent changes. This is a definite bug of the test, I suspect the server’s side. The root cause of this zero division is that no packets are sent and no packets are received in one of 3 test iterations. The test’s intention is to send data for 1 millisecond. It seems that all the commands are received by server but the server definitely does not send one of batches of packets. There are 3 batches to be sent – they’re to be received with limits of 10, 100 and 1000 packets per DPC. This can happen on any batch – exception terminates the test. So far,the whql passed via driver 28 (under Win7 folder) on windows 7 32/64 and windows 2008 R2 platform. The rest testing is on going QE run WHQL Test on virtoi-win-prewhql-29 and virtio-win-prewhql-31 ,did not this issue Based on above ,this issue has been fixed ald .move status to VERIFIED. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2013-0441.html |