Hello,
According to DM8168 errata, silicon revision 2.0 has problem with initializing link to SATA GEN3 drives.
In our case, there are two drives, Kingston V300 and SanDisk U110, both are SSD.
With Kingston, it takes up to 10 seconds to initialize the drive (but it succeedes), and with SanDisk, when doing power cycling test, once in a while SATA controller gets stuck during initializing with the drive, as described in errata.
Since we have pretty big batch of boards with 2.0 samples, we can't simply switch to 2.1 which supposedly solved this problem. Also, these SSD drives don't have jumpers to cap them to GEN2, so we have to assert this in other way.
1. Are there some register settings for SATA controller which may help preventing the stuck on SanDisk?
2. Errata mentiones issuing SATA reset command after unsuccessfull initialization. Did anyone tried that in Linux and where is the right place to do this?
Thanks!
