We are using both your Windows and Linux test suites on the same hardware. With multiple SBC's and onboard 1.8" hard drives, on the secondary ide interface, we get a consistent failure in the hard drive test under Windows XP (latest SP's). However, we run the same hard drive test using Linux test suite (under Redhat WS 4.1, Fedora core 2.6.9) and the hard drives NEVER FAIL UNDER LINUX.
The failure under XP is "Test file could not be created". Sometimes it appears the hard drive is just dropping out (for instance if the test is run FROM the primary ide but testing against the secondary).
We are trying to determine if this could be :
A) The windows driver (atapi.sys). Have you ever seen issues in the area with current driver. This is a very mature driver I would think.
B) The hard drive itself (this particular model is a 1.8" Toshiba MK3008GAL). Would it be wise to try another 1.8" drive from different manuf?
C) Maybe it's the bios (but why would Linux behave so differently than XP if the bios hadn't initialized the drive controller properly???).
D) Difference between Linux and Windows test suites.
In another thread you guys have pointed out to me that the Linux test suite does not do the "butterfly seeking" hard drive test, but that Windows test does. Could this difference in the testing method be causing this particular error?