SATA fix for 3snic raid controller

[steven.song]

I fixed some issues when using SATA disks.

Please see attached log files.

[steven.song]

smartctl -x

[Christian Franke]

Thanks for the patch. Looks like the CK_COND issue is fixed now.

Remaining issues, some found by cppcheck (​https://cppcheck.sourceforge.io/):

• Line 1511: Non-boolean value returned from function returning bool: return (r) always returns nonzero, which is interpreted as true. IMPORTANT: Before submitting future patches, make sure to check also for sane behavior if ioctl() fails.
• Line 1520: The cast of the first parameter is useless because it is already a pointer.
• Line 1520: Always use C++-style casts instead of C-style casts for T=>pointer conversions, for example: memcpy(iop->sensep, reinterpret_cast<void *>(io_hdr_v4.response), len);
• Hint: Never use legacy NULL (which is the same as 0 in C++) in new C++ code, use C++11 nullptr instead. Using the pointer as a boolean in conditions is allowed, for example if (iop->sensep && len > 0) ....
• Hint: Lines 1449-1455: All three initial memset(...) are not needed, C++11 zero initialization is more safe. Simply append {}, for example: struct sg_io_v4 io_hdr_v4{};
• Line 1515: Remove trailing whitespaces.
• As usual provide a ChangeLog entry - in same diff/patch file please.

[steven.song]

updated smartctl-x.log

[Christian Franke]

Applied in r5457, thanks!