Re: hard drive errors on acpi s3 resume

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

Please don't top post.

> Just out of curiosity, what issues do you find with the coding
> style?

I hope that means you want to clean it up and apply it ;-).

+#ifdef CONFIG_ACPI
+#include <linux/acpi.h>
+#define DBG(x...) printk(x)

Include in the middle of file? Renaming of printk to confuse reader?

+       int i, tmp;
+       acpi_integer addr;
...
+       addr = i;

...oops, something is very wrong in acpi land.

+#define GTM_LEN (sizeof(u32) * 5)
...
+       u32     gtm[GTM_LEN/sizeof(u32)]; /* info from _GTM */

Eh? Better names would be nice, too.

+       for (i = 0; i < MAX_DEVICES; i ++)

Please no space between i and ++.

+static struct acpi_ide_stat *ide_get_acpi_state(acpi_handle handle)

The function is strange, btw. Linear search, and it basically returns
pointer to zeros. I'd expect zeroing to be in caller or something like that.

+int acpi_ide_suspend(struct device *dev)
+{
+       acpi_handle handle, parent_handle;
+       struct acpi_ide_stat *stat;
+       acpi_status     status;

At least alignment should be consistent locally.

+static int acpi_ide_gtf(acpi_handle handle, ide_drive_t *drive)
+{
+       struct acpi_buffer      output = {ACPI_ALLOCATE_BUFFER, NULL};
+       ide_task_t      args;
+       int index = 0;
+       unsigned char *data;
+       union acpi_object       *package = NULL;
+       acpi_status status;

...even more horible example of above.

+               /* submit command request */
+//             printk("return value %d\n", ide_raw_taskfile(drive, &args, NULL));
+               index += 7;

Someone should learn to use delete key and version control...

+       if (stat->channel_handled == 0) {
+               stat->handle = NULL;
+               goto gtf;
+       }
+DBG("Start STM\n");
+       if (acpi_ide_stm(stat))
+               return -ENODEV;
+       stat->channel_handled = 0;
+gtf:
+       return acpi_ide_gtf(handle, drive);
+}

What about "else" instead of ugly goto. And that DBG should really go
away.

-
+#ifdef CONFIG_ACPI
+       acpi_ide_suspend(dev);
+#endif

-
+#ifdef CONFIG_ACPI
+       acpi_ide_resume(dev);
+#endif
        memset(&rq, 0, sizeof(rq));
        memset(&rqpm, 0, sizeof(rqpm));
        memset(&args, 0, sizeof(args));

We normally allways provide acpi_ide_resume, but make them nop in
!ACPI case. That means #ifdefs are avoided in callers.

...should be enough...
							Pavel
-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

RE: hard drive errors on acpi s3 resume

[Moore, Robert]

Yep on all.
Bob

> +       acpi_handle handle, parent_handle;

I don't like multiple declarations in one statement.

Bob

> -----Original Message-----
> From: Pavel Machek [mailto:pavel-+ZI9xUNit7I@public.gmane.org]
> Sent: Tuesday, November 08, 2005 1:12 AM
> To: Moore, Robert
> Cc: ACPI mailing list
> Subject: Re: [ACPI] hard drive errors on acpi s3 resume
> 
> Hi!
> 
> Please don't top post.
> 
> > Just out of curiosity, what issues do you find with the coding
> > style?
> 
> I hope that means you want to clean it up and apply it ;-).
> 
> +#ifdef CONFIG_ACPI
> +#include <linux/acpi.h>
> +#define DBG(x...) printk(x)
> 
> Include in the middle of file? Renaming of printk to confuse reader?
> 
> +       int i, tmp;
> +       acpi_integer addr;
> ...
> +       addr = i;
> 
> ...oops, something is very wrong in acpi land.
> 
> +#define GTM_LEN (sizeof(u32) * 5)
> ...
> +       u32     gtm[GTM_LEN/sizeof(u32)]; /* info from _GTM */
> 
> Eh? Better names would be nice, too.
> 
> +       for (i = 0; i < MAX_DEVICES; i ++)
> 
> Please no space between i and ++.
> 
> +static struct acpi_ide_stat *ide_get_acpi_state(acpi_handle handle)
> 
> The function is strange, btw. Linear search, and it basically returns
> pointer to zeros. I'd expect zeroing to be in caller or something like
> that.
> 
> +int acpi_ide_suspend(struct device *dev)
> +{
> +       acpi_handle handle, parent_handle;
> +       struct acpi_ide_stat *stat;
> +       acpi_status     status;
> 
> At least alignment should be consistent locally.
> 
> +static int acpi_ide_gtf(acpi_handle handle, ide_drive_t *drive)
> +{
> +       struct acpi_buffer      output = {ACPI_ALLOCATE_BUFFER, NULL};
> +       ide_task_t      args;
> +       int index = 0;
> +       unsigned char *data;
> +       union acpi_object       *package = NULL;
> +       acpi_status status;
> 
> ...even more horible example of above.
> 
> +               /* submit command request */
> +//             printk("return value %d\n", ide_raw_taskfile(drive,
&args,
> NULL));
> +               index += 7;
> 
> Someone should learn to use delete key and version control...
> 
> +       if (stat->channel_handled == 0) {
> +               stat->handle = NULL;
> +               goto gtf;
> +       }
> +DBG("Start STM\n");
> +       if (acpi_ide_stm(stat))
> +               return -ENODEV;
> +       stat->channel_handled = 0;
> +gtf:
> +       return acpi_ide_gtf(handle, drive);
> +}
> 
> What about "else" instead of ugly goto. And that DBG should really go
> away.
> 
> -
> +#ifdef CONFIG_ACPI
> +       acpi_ide_suspend(dev);
> +#endif
> 
> -
> +#ifdef CONFIG_ACPI
> +       acpi_ide_resume(dev);
> +#endif
>         memset(&rq, 0, sizeof(rq));
>         memset(&rqpm, 0, sizeof(rqpm));
>         memset(&args, 0, sizeof(args));
> 
> We normally allways provide acpi_ide_resume, but make them nop in
> !ACPI case. That means #ifdefs are avoided in callers.
> 
> ...should be enough...
> 							Pavel
> --
> Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

RE: hard drive errors on acpi s3 resume

[Li, Shaohua]

Hi,
>
>Please don't top post.
>
>> Just out of curiosity, what issues do you find with the coding
>> style?
>
>I hope that means you want to clean it up and apply it ;-).
>
BTW, is this really worth submitting to upstream? It doesn't support
_GTF. And it appears useless in my test systems.

Thanks,
Shaohua


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> >Please don't top post.
> >
> >> Just out of curiosity, what issues do you find with the coding
> >> style?
> >
> >I hope that means you want to clean it up and apply it ;-).
> >
> BTW, is this really worth submitting to upstream? It doesn't support
> _GTF. And it appears useless in my test systems.

My test systems work well enough without that. If there's a system
where it helps, I'd submit it.

[Could we loudly printk in not-yet-implemented _GTF to at least see
how many BIOSen need it? That could be good enough reason to apply it,
too.]

								Pavel
-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Shaohua Li]

On Tue, 2005-11-08 at 10:30 +0100, Pavel Machek wrote:
> Hi!
> 
> > >Please don't top post.
> > >
> > >> Just out of curiosity, what issues do you find with the coding
> > >> style?
> > >
> > >I hope that means you want to clean it up and apply it ;-).
> > >
> > BTW, is this really worth submitting to upstream? It doesn't support
> > _GTF. And it appears useless in my test systems.
> 
> My test systems work well enough without that. If there's a system
> where it helps, I'd submit it.
> 
> [Could we loudly printk in not-yet-implemented _GTF to at least see
> how many BIOSen need it? That could be good enough reason to apply it,
> too.]
I guess I should clean it up and let people in bugzilla try. If there
are success stories, I'll report back.

Thanks,
Shaohua 



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Matthew Garrett]

On Tue, Nov 08, 2005 at 05:27:42PM +0800, Li, Shaohua wrote:

> BTW, is this really worth submitting to upstream? It doesn't support
> _GTF. And it appears useless in my test systems.

It fixes a lot of systems. With _GTF, it may fix more.

-- 
Matthew Garrett | mjg59-1xO5oi07KQx4cg9Nei1l7Q@public.gmane.org


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> > BTW, is this really worth submitting to upstream? It doesn't support
> > _GTF. And it appears useless in my test systems.
> 
> It fixes a lot of systems. With _GTF, it may fix more.

If you have one of those systems, can you clean the patch up and
submit it?
							Pavel
-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Shaohua Li]

On Tue, 2005-11-08 at 19:06 +0000, Matthew Garrett wrote:
> On Tue, Nov 08, 2005 at 05:27:42PM +0800, Li, Shaohua wrote:
> 
> > BTW, is this really worth submitting to upstream? It doesn't support
> > _GTF. And it appears useless in my test systems.
> 
> It fixes a lot of systems. With _GTF, it may fix more.
I indeed have a system with ide issue. The patch can't work for it, I
guess it needs _GTF.
In the bugzilla reports, ide failure is the second priority bugs (video
issue is the first). I guess we also need the similar thing for sata.
One of my test system suffers from sata failure after resume. I guess
it's the same reason.

Thanks,
Shaohua



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Alan Cox]

On Mer, 2005-11-09 at 08:52 +0800, Shaohua Li wrote:
> In the bugzilla reports, ide failure is the second priority bugs (video
> issue is the first). I guess we also need the similar thing for sata.
> One of my test system suffers from sata failure after resume. I guess
> it's the same reason.

SATA power management is still mostly a set of add on patches not in the
mainstream tree. Some may need _GTF but fortunately the interfaces to do
raw polled taskfiles are clean and present in libata



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[sdfa]

wow, i didn't expect so many responses, thanks alot!



about the patch:

i'll try that patch and tell you guys how it goes.



about checking the dsdt:

i dont have much experience with dsdts, so i dont know how to check them, edit them, definitely wouldnt know what interesting behavior might constitute. but this guy here: http://linux.highsphere.net/howtos/b100b.php has almost the exact same laptop as i do(the only difference is screensize, i believe) and he didn't seem to have any problems with his dsdt, though he doesnt seem to have this hard drive issue(or isnt aware of it, he does mention it taking about 40 seconds to resume from s3).



about spinning the drive down before s3: 

when i do a hdparm -y /dev/hda, the drive immediately spins back up, probably due to some logger daemon. i've tried killing syslogd and klogd, but this still happens.



about it being an ide driver problem:

if it is an ide driver issue, can anybody point me in the right direction? maybe an ide driver mailing list or something? i'm believe i'm using the PIIX driver(i've got an ich4-m ide controller), but i havent been able to find a development website or anything for it. the help in menuconfig said that PIIX was for ich4 ide controllers, so i was hoping to find some intel-sponsored ich4-m driver development, but no luck.

_______________________________________________
Join Excite! - http://www.excite.com
The most personalized portal on the Web!




-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> about spinning the drive down before s3: 
> 
> when i do a hdparm -y /dev/hda, the drive immediately spins back up, probably due to some logger daemon. i've tried killing syslogd and klogd, but this still happens.
> 
>

Try init=bin/bash with ext2 (not ext3).

						Pavel

-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Alan Cox]

On Llu, 2005-11-07 at 22:06 -0500, sdfa wrote:
> about it being an ide driver problem:
> 
> if it is an ide driver issue, can anybody point me in the right direction? maybe an ide driver mailing list or something? i'm believe i'm using the PIIX driver(i've got an ich4-m ide controller), but i havent been able to find a development website or anything for it. the help in menuconfig said that PIIX was for ich4 ide controllers, so i was hoping to find some intel-sponsored ich4-m driver development, but no luck.

linux-ide-u79uwXL29TY76Z2rM5mHXA@public.gmane.org



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

RE: hard drive errors on acpi s3 resume

[Li, Shaohua]

Hi,
>-----Original Message-----
>From: acpi-devel-admin-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org [mailto:acpi-devel-
>admin-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org] On Behalf Of Matthew Garrett
>Sent: Monday, November 07, 2005 9:23 AM
>To: Alan Cox
>Cc: Pavel Machek; sdfa; acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
>Subject: Re: [ACPI] hard drive errors on acpi s3 resume
>
>On Sun, Nov 06, 2005 at 08:48:59PM +0000, Alan Cox wrote:
>> On Sul, 2005-11-06 at 17:27 +0100, Pavel Machek wrote:
>> > Also I'm afraid you'll have to ask this on ide lists; this is not
an
>> > ACPI but a driver problem.
>>
>> Its probably also an ACPI problem. ACPI provides disk suspend/resume
>> methods but nobody as hooked them to anything. Basically you
shouldn't
>> mix ACPI S3 and IDE disks in Linux until this is done.
>
>There's a patch that floated on here a while ago - we're shipping it in
>Ubuntu. Hang on...
>
>http://www.codon.org.uk/~mjg59/tmp/external-drivers-ide_sleep.dpatch is
>what we're shipping - it seems to work in most cases, but not all. It
>implements the _GTM and _STM methods, but doesn't do anything with _GTF
>since the IDE layer hasn't been resumed at the point where this code
>executes, so there's no existing mechanism to execute the taskfile
>commands returned. It may help in this case.
>
>(I've no idea why this hasn't been submitted upstream yet - it's not my
>patch...)
yes, this is the debug patch from me. As you said, it does nothing with
_GTF, so I didn't submit it.

Alan, is there any way we can execute the taskfile commands before the
IDE device resume?

Thanks,
Shaohua


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

hard drive errors on acpi s3 resume

[sdfa]

i've been trying to get s3 working on my sony vaio, but i'm having some problems concerning the hard drive during resume. after a resume i see these messages in dmesg:



hda: lost interrupt

 evevent-0299: *** Error: No installed handler for fixed event [00000002]

Restarting tasks... done

hda: dma_timer_expiry: dma status == 0x21

hda: DMA timeout error

hda: dma timeout error: status=0x50 { DriveReady SeekComplete }

ide: failed opcode was: unknown

hda: status error: status=0x50 { DriveReady SeekComplete }

ide: failed opcode was: unknown

hda: no DRQ after issuing MULTWRITE_EXT

hda: status error: status=0x50 { DriveReady SeekComplete }

ide: failed opcode was: unknown

hda: no DRQ after issuing MULTWRITE_EXT

hda: lost interrupt

hda: status error: status=0x50 { DriveReady SeekComplete }

ide: failed opcode was: unknown

hda: no DRQ after issuing MULTWRITE_EXT

hda: status error: status=0x50 { DriveReady SeekComplete }

ide: failed opcode was: unknown

hda: no DRQ after issuing MULTWRITE_EXT

ide0: reset: success



sometimes i also see this message right before all of this, i'm not sure if it's related:

acpi_bus_0200 [77] bus_set_power      : Device is not power manageable



the drive eventually spins up and the resume is a success, but it takes like two minutes(too long) and i'm afraid it may be damaging my hard drive.



the drive is an ide ata drive with an ich4-m controller.

_______________________________________________
Join Excite! - http://www.excite.com
The most personalized portal on the Web!




-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> i've been trying to get s3 working on my sony vaio, but i'm having
>some problems concerning the hard drive during resume. after a resume
>i see these messages in dmesg:

...
> 
> the drive eventually spins up and the resume is a success, but it takes like two minutes(too long) and i'm afraid it may be damaging my hard drive.


Please fit your mails to 80 columns.

Also I'm afraid you'll have to ask this on ide lists; this is not an
ACPI but a driver problem.

[Are you doing suspend in middle of some disk operation?]
								Pavel

-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[DJA]

Pavel Machek wrote:
> Hi!
> 
> 
>>i've been trying to get s3 working on my sony vaio, but i'm having
>>some problems concerning the hard drive during resume. after a resume
>>i see these messages in dmesg:
> 
> 
> ...
> 
>>the drive eventually spins up and the resume is a success, but it takes like two minutes(too long) and i'm afraid it may be damaging my hard drive.
> 
> 
> 
> Please fit your mails to 80 columns.
> 
> 								Pavel

His post looked fine in my MUA (Thunderbird). Your MUA is responsible 
for line-wrapping, not the sender's (assuming the email is text and not 
HTML).

-- 
    Best Regards,
       ~DJA.


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

> >Hi!
> >
> >>i've been trying to get s3 working on my sony vaio, but i'm having
> >>some problems concerning the hard drive during resume. after a resume
> >>i see these messages in dmesg:
> >...
> >
> >>the drive eventually spins up and the resume is a success, but it takes 
> >>like two minutes(too long) and i'm afraid it may be damaging my hard 
> >>drive.
> >
> >Please fit your mails to 80 columns.
> 
> His post looked fine in my MUA (Thunderbird). Your MUA is responsible 
> for line-wrapping, not the sender's (assuming the email is text and not 
> HTML).

No, thats not how it should work on the lists.
-- 
64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Alan Cox]

On Sul, 2005-11-06 at 17:27 +0100, Pavel Machek wrote:
> Also I'm afraid you'll have to ask this on ide lists; this is not an
> ACPI but a driver problem.

Its probably also an ACPI problem. ACPI provides disk suspend/resume
methods but nobody as hooked them to anything. Basically you shouldn't
mix ACPI S3 and IDE disks in Linux until this is done.

Usually it'll work but if it doesn't then you may want to join in
writing the needed code.

Alan



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> > Also I'm afraid you'll have to ask this on ide lists; this is not an
> > ACPI but a driver problem.
> 
> Its probably also an ACPI problem. ACPI provides disk suspend/resume
> methods but nobody as hooked them to anything. Basically you shouldn't
> mix ACPI S3 and IDE disks in Linux until this is done.

Okay, I keep forgetting about this one.

slinky, can you take a look if your DSDT does something interesting in
IDE suspend/resume methods?

> Usually it'll work but if it doesn't then you may want to join in
> writing the needed code.

Could we work around it by hdparm -y /dev/hda just before S3? There
should be nothing to save about stopped harddrive ;-).
								Pavel
-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Matthew Garrett]

On Sun, Nov 06, 2005 at 08:48:59PM +0000, Alan Cox wrote:
> On Sul, 2005-11-06 at 17:27 +0100, Pavel Machek wrote:
> > Also I'm afraid you'll have to ask this on ide lists; this is not an
> > ACPI but a driver problem.
> 
> Its probably also an ACPI problem. ACPI provides disk suspend/resume
> methods but nobody as hooked them to anything. Basically you shouldn't
> mix ACPI S3 and IDE disks in Linux until this is done.

There's a patch that floated on here a while ago - we're shipping it in 
Ubuntu. Hang on...

http://www.codon.org.uk/~mjg59/tmp/external-drivers-ide_sleep.dpatch is 
what we're shipping - it seems to work in most cases, but not all. It 
implements the _GTM and _STM methods, but doesn't do anything with _GTF 
since the IDE layer hasn't been resumed at the point where this code 
executes, so there's no existing mechanism to execute the taskfile 
commands returned. It may help in this case.

(I've no idea why this hasn't been submitted upstream yet - it's not my 
patch...)
-- 
Matthew Garrett | mjg59-1xO5oi07KQx4cg9Nei1l7Q@public.gmane.org


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php

Re: hard drive errors on acpi s3 resume

[Pavel Machek]

Hi!

> > > Also I'm afraid you'll have to ask this on ide lists; this is not an
> > > ACPI but a driver problem.
> > 
> > Its probably also an ACPI problem. ACPI provides disk suspend/resume
> > methods but nobody as hooked them to anything. Basically you shouldn't
> > mix ACPI S3 and IDE disks in Linux until this is done.
> 
> There's a patch that floated on here a while ago - we're shipping it in 
> Ubuntu. Hang on...
> 
> http://www.codon.org.uk/~mjg59/tmp/external-drivers-ide_sleep.dpatch is 
> what we're shipping - it seems to work in most cases, but not all. It 
> implements the _GTM and _STM methods, but doesn't do anything with _GTF 
> since the IDE layer hasn't been resumed at the point where this code 
> executes, so there's no existing mechanism to execute the taskfile 
> commands returned. It may help in this case.

Its coding style leaves something to be desired, but otherwise it
looks okay. Unfortunately, I do not need it on my machines, so someone
else has to test/submit this one.

								Pavel
-- 
Thanks, Sharp!


-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php