Linux ATA/IDE development
 help / color / mirror / Atom feed
From: sonofagun@openmailbox.org
To: linux-ide@vger.kernel.org
Subject: Re: AM4 B350 chipset Sata/IDE problem
Date: Wed, 22 Nov 2017 11:02:54 -0000	[thread overview]
Message-ID: <20171122110254.ADB344E0090@mta-1.openmailbox.og> (raw)

 Hello, you have a brand new AMD box and an old seagate HDD. That does NOT sound good... I am an AMD fan as well. Many years ago my brother and I bought our first AMD board with SB700. A good chip that had many annoying errata... A SATA II chip paired with the best iGPU of its era. Only a good HDD was suitable for such a board! That was an HDD from the seagate 7200.10 family. A family full of innovations. The first SATA II family, the first PMR HDD family and the first family offering 5 year warranty. Those disks are very reliable compared to 7200.7 and to 7200.11!!! Actually their access time is like 7200RPM disks while the 7200.11 had access times of 5400RPM class disks... Unfortunately those disks had a compatibility issue with the AMD AHCI implementation. All revisions of the SB7x0 southbridge family were affected and transfer speed was hurt badly when NCQ was enabled! I wrote about it(https://ata.wiki.kernel.org/index.php/Known_issues#SATA_hard_drives_which_show_poor_performance_with_sequential_reads_.28e.g._hdparm_-t.29) so that everyone will know it.
At that time there was no official firmware for any SATA disk. Firmware was given only when something was really wrong such as the deathstars. That attitude applied to all HDD manufacturers. My brother had to phone seagate at their headquarters in order to ask for a fix. Our next seagate disks(7200.11) were all having the same transfer speed issues with NCQ. They were too arrogant to admit the issue and very reluctant to release new firmware. We were informed that there will be no fix even though the 7200.11 family was just released. We were forced to RMA 4x 320GB and 4x 500GB of them at once and switch to the legendary WD RE3(all those still work flawlessly from 2008!). This was just prior seagate admitted the rest firmware issues of this family... After that mess, new firmware has been released gradually for most seagate HDD families!

As you can guess I am not surprised that you have an issue with such a disk on an AMD board. Compared to current disks it is very power hungry and not suitable for your board. What is worse, you are using an out of warranty HDD here.

Your SATA II HDD is attached on a SATA 3.x AMD chip. This HDD cannot even reach SATA I speeds(average well below 100MB/s). If your disk operates in 1.5 Gbps you will not have noticeable performance degradation. You will just lose maximum burst speed. Some time ago I saw this: "https://marc.info/?t=150843788300002&r=1&w=2", but I have not fallen on it as I have never used SATA hotplug. You have not mentioned hotplug here but it might be related as the libata EH is invoked!

I still have a 320GB 7200.10 ST3320620AS with 3.AAK firmware and will try to measure it with SATA I speed on my AMD KABINI box. We were given a newer firmware for this one while the seagate firmware finder site says it has no firmware! The new firmware is 3.AAM for this GALAXY HDD but the speed problem on AMD AHCI was not completely resolved but it made a difference so we have kept this HDD. We had a couple 320GB too, with 3.AAE firmware but seagate supplied no new firmware so they were RMAed.

As I said above, those disks have had speed issues with NCQ. Let's start from that. Please boot with NCQ disabled and see if it helps...

Since your HDD is old, please run:
sudo smartctl /dev/sd@ -x
sudo hdparm -I /dev/sd@
sudo hdparm --Istdout /dev/sd@
dmesg
and upload their output here(replace @ with your drive letter) so that I can view its current condition and give proper advice.


AMD SATA hosts have various issues with old HDDs:

On SB7x0 and SB950 all SATA I WD ****JD series went crazy when DIPM was enabled. Its heads parked then unparked and that was repeated endlessly. This has been fixed by Tejun after this issue was reported to the kernel...

Also SB950 will not even detect some MAXTOR HDDs unless they have the latest firmware! Again that firmware does "NOT" exist...

             reply	other threads:[~2017-11-22 11:09 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-22 11:02 sonofagun [this message]
2017-11-22 15:39 ` AM4 B350 chipset Sata/IDE problem Mark Hounschell
  -- strict thread matches above, loose matches on Subject: below --
2017-11-23  6:40 sonofagun
2017-11-27 20:50 ` Mark Hounschell
2017-11-21 16:01 Mark Hounschell

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20171122110254.ADB344E0090@mta-1.openmailbox.og \
    --to=sonofagun@openmailbox.org \
    --cc=linux-ide@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox