Bug 111871

Summary:	Computer freezes at boot every 2-3 boots on Gigabyte motherboard with MARVEL SATA ports set to AHCI
Product:	Other	Reporter:	Britt Yazel (bwyazel)
Component:	Other	Assignee:	other_other
Status:	NEW ---
Severity:	high
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	4.4.1	Subsystem:
Regression:	No	Bisected commit-id:

Description Britt Yazel 2016-02-04 08:33:43 UTC

I do not know how else to really describe this bug. My computer is running ArchLinux x64, I have an AMD fx-8120 on a 990fx chipset, and I am using an R9-290 graphics chip.

The issue is that every 2-3 boots (since around Kernel 4.0), my computer gets frozen while booting and requires a hard reset (or two) to get me to a functional desktop. It is exceedingly annoying, and I suspect it has something to do with the AMD architecture, as my laptop with an Intel i5 does not experience these issues.

My computer shows the initial lines saying systemd version 228, but then gets completely locked up.

Now, I know you need me to report some sort of logs, but I am fairly novice so please go easy on me. I just want to see this issue go away, and for my system to get back to consistently booting each and every time.

I believe my bug is related to this bug:
https://bugs.archlinux.org/task/46745

Comment 1 Britt Yazel 2016-02-05 02:58:51 UTC

I believe this issue is related to this bug:
https://bugzilla.kernel.org/show_bug.cgi?id=87581

I have traced down the root cause of this bug being when my Gigabyte 990fxa-ud3 motherboard has its off-chipset SATA ports, in this case MARVEL SATA ports, set to AHCI mode instead of IDE mode.

When the secondary SATA ports are set to AHCI, the boot process is slow and freezes every 2-3 boots. When I set these SATA ports back to IDE mode, the computer boots quickly and consitently. As far as why this issue happens, I have no idea, perhaps the other bug has a similar issue, as theirs also happens with gigabyte motherboards. Theirs was weird though in that theirs froze when their secondary SATA ports (JMicron) were set to IDA and NOT ahci. So it was the complete reverse of mine, though perhaps there is a common cause.