hdaps-devel

[Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Tomáš J. <to...@no...>]

Hello,

I recently started using hdapsd on my ThinkPad T420 and I am experiencing this
problem originally reported in 2006 -- when the heads are parked, the disk
refuses to go into standby.  The disk I have describes itself as

    scsi 0:0:0:0: Direct-Access     ATA      ST9500420AS      0003 PQ: 0 ANSI: 5

I'm running kernel 3.3.8, but the code in question seems to be untouched since
2008, so that's largely irrelevant.  I suspected that perhaps CHECK POWER MODE
isn't enough to reenable standby timer, so Ie patched to code to check the
output of that command and issue an IDLE IMMEDIATE (without UNLOAD) command if
the disk is not in standby, and it does fix the issue. The code is here:

    https://github.com/liskin/patches/blob/master/hacks/linux-3.3_hdaps-unprevent-standby.patch

Am I the only one experiencing this problem or do we perhaps want this hack in
the mainline?  It should be easily reproducible using

    hdparm -S 1 /dev/sda && echo 1000 >/sys/block/sda/device/unload_heads

If nobody else accesses the disk within the next 5 seconds and it doesn't spin
down, then you have that problem too. :-)

(Feel free to forward this e-mail to anyone who might have an opinion, thanks.)

Regards,
-- 
Tomáš Janoušek, a.k.a. Liskni_si, http://work.lisk.in/

Re: [Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Erik A. <er...@vo...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

T420 with hdaps from tp-smapi, disk spinning down.
Kernel is vanilla 3.4.25.

greetings,
Erik


On 24.02.2013 18:57, Tomáš Janoušek wrote:
> Hello,
> 
> I recently started using hdapsd on my ThinkPad T420 and I am
> experiencing this problem originally reported in 2006 -- when the
> heads are parked, the disk refuses to go into standby.  The disk I
> have describes itself as
> 
> scsi 0:0:0:0: Direct-Access     ATA      ST9500420AS      0003 PQ:
> 0 ANSI: 5
> 
> I'm running kernel 3.3.8, but the code in question seems to be
> untouched since 2008, so that's largely irrelevant.  I suspected
> that perhaps CHECK POWER MODE isn't enough to reenable standby
> timer, so Ie patched to code to check the output of that command
> and issue an IDLE IMMEDIATE (without UNLOAD) command if the disk is
> not in standby, and it does fix the issue. The code is here:
> 
> https://github.com/liskin/patches/blob/master/hacks/linux-3.3_hdaps-unprevent-standby.patch
>
>  Am I the only one experiencing this problem or do we perhaps want
> this hack in the mainline?  It should be easily reproducible using
> 
> hdparm -S 1 /dev/sda && echo 1000
> >/sys/block/sda/device/unload_heads
> 
> If nobody else accesses the disk within the next 5 seconds and it
> doesn't spin down, then you have that problem too. :-)
> 
> (Feel free to forward this e-mail to anyone who might have an
> opinion, thanks.)
> 
> Regards,

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAlEqZe8ACgkQ8NqlQQxmej5CwwCgqfLLJGe0LWT/gph4bJuSMLzy
MD0An2hEFC8Zrpv1W4rWtgFnJmCydrwx
=ej4c
-----END PGP SIGNATURE-----

Re: [Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Tomáš J. <to...@no...>]

Hi,

On Sun, Feb 24, 2013 at 08:11:45PM +0100, Erik Andresen wrote:
> T420 with hdaps from tp-smapi, disk spinning down.
> Kernel is vanilla 3.4.25.

Thanks for your reply.  It is quite strange, though.  I tried 3.2 from Debian,
vanilla 3.3.8 and vanilla 3.8.0 and I have the problem on all of them (except
the one I patched).

Can you confirm that you have the same hard drive and firmware version,
please?  Here, smartctl --all /dev/sda reports:

Model Family:     Seagate Momentus 7200.4
Device Model:     ST9500420AS
Firmware Version: 0003LVM1

That is, I think, the newest firmware available according to Lenovo software
updates.

Regards,
-- 
Tomáš Janoušek, a.k.a. Liskni_si, http://work.lisk.in/

Re: [Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Erik A. <er...@vo...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

This is one of the earliest T420:
Device Model:     HITACHI HTS725050A9A364
Firmware Version: PC4ZC70F

We might have more then two different disks out there..

Please also note that I'm using the hdaps module from tp-smapi.
The vanilla hdaps module doesn't support my system.

greetings,
Erik


On 26.02.2013 20:30, Tomáš Janoušek wrote:
> Hi,
> 
> On Sun, Feb 24, 2013 at 08:11:45PM +0100, Erik Andresen wrote:
>> T420 with hdaps from tp-smapi, disk spinning down. Kernel is
>> vanilla 3.4.25.
> 
> Thanks for your reply.  It is quite strange, though.  I tried 3.2
> from Debian, vanilla 3.3.8 and vanilla 3.8.0 and I have the problem
> on all of them (except the one I patched).
> 
> Can you confirm that you have the same hard drive and firmware
> version, please?  Here, smartctl --all /dev/sda reports:
> 
> Model Family:     Seagate Momentus 7200.4 Device Model:
> ST9500420AS Firmware Version: 0003LVM1
> 
> That is, I think, the newest firmware available according to Lenovo
> software updates.
> 
> Regards,

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAlEtEmYACgkQ8NqlQQxmej53TQCZAXClV2ETYRnUA61zYwOXejKh
pKMAnjTZxJxhz5Uu6wZ82POdsnnQC/PL
=6eHM
-----END PGP SIGNATURE-----

Re: [Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Tomáš J. <to...@no...>]

Hi,

On Tue, Feb 26, 2013 at 08:52:08PM +0100, Erik Andresen wrote:
> This is one of the earliest T420:
> Device Model:     HITACHI HTS725050A9A364
> Firmware Version: PC4ZC70F
> 
> We might have more then two different disks out there..

Yeah, I think this explains it.  Different manufacturer, different firmware,
different behaviour of CHECK POWER MODE ata command. :-)

> Please also note that I'm using the hdaps module from tp-smapi.
> The vanilla hdaps module doesn't support my system.

Yeah, me too, but that's irrelevant.  The command I posted reproduces the
problem without either hdaps module loaded.

Anyway, thanks for your info.

-- 
Tomáš Janoušek, a.k.a. Liskni_si, http://work.lisk.in/

Re: [Hdaps-devel] HDAPS protect patch interferes with disk standby, again

[Elias O. <eo...@ne...>]

(2013-02-26) Tomáš Janoušek <tomi-YoqI/XI...@pu...> wrote:
> Hi,
> On Tue, Feb 26, 2013 at 08:52:08PM +0100, Erik Andresen wrote:
>> This is one of the earliest T420:
>> Device Model:     HITACHI HTS725050A9A364
>> Firmware Version: PC4ZC70F
>> 
>> We might have more then two different disks out there..
> 
> Yeah, I think this explains it.  Different manufacturer, different firmware,
> different behaviour of CHECK POWER MODE ata command. :-)

It does. The patch which eventually made it into mainline had been
tested on a Hitachi drive and worked as expected. I can reproduce your
problem on a Western Digital model though.

Double checking the relevant ATA spec, it turns out that CHECK POWER
MODE not restarting the standby timer might actually be the right
thing(tm) to do. Anyway, I'll try to get someone's attention on
linux-ide regarding the matter at hand.

Thanks for the patch,

Elias