On Sun, Sep 22, 2013 at 5:59 PM, Tejun Heo <tj@xxxxxxxxxx> wrote: > On Sun, Sep 22, 2013 at 03:56:35AM +0200, Patrik Jakobsson wrote: >> Where REGS_PER_GTF is defined as 7 which is correct according to ACPI specs. >> Since I'm getting a length of 8 I started digging in the ACPI code and found >> that the length is rounded up to acpi_size (u32 or u64 depending on arch). I >> cannot find any commits that recently touched this though I didn't really dig >> through it all. > > It means your bios is buggy. I must be confusing the actual buffer length with the allocated buffer length. It is not entirely clear what is what in the ACPI-CA code. What I don't understand is how a simple "Return (GTF0)" where GTF0 is 56 bits gets padded to 64 bits by the firmware. Maybe the iasl disassembly is lying to me. >> I've disassembled the SSDT for the "SataAhci" and everything looks ok. That >> code returns a 56 bit buffer. I also tried calling the _GTF method manually >> which returned the following (padded to 8 bytes by the ACPI code ofc). >> >> {0x10, 0x03, 0x00, 0x00, 0x00, 0xa0, 0xef, 0x00} >> >> Doing a quick google search gives me a few of these _GTF length errors dating >> back to at least 2011. What's going on here? Is this a known error? > > Failing to execute _GTF is usally fine to the point where I sometimes > wonder why we bother with it at all. The above command is SET > FEATURES / Enable SATA Feature / DIPM, which enables device side link > powersaving and is a weird thing to do from _GTF anyway as host side > needs configuration too. You can properly configure it using libata > link PM. Thanks for the info. I tested "fixing up" the code to let the _GTF slide through and I couldn't see it make any difference whatsoever. As you say, we can safely ignore this. Cheers Patrik > > Thanks. > > -- > tejun -- To unsubscribe from this list: send the line "unsubscribe linux-ide" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html