Bitfield usage in FADT table?

Bitfield usage in FADT table?

[maxime bizon]

Hello,

I'm using kernel 2.6.0-test9 on a i386 notebook.

I forced a new FADT table in the kernel and put the following bytes at
offset 0x70:  0xf1 0xa2 0x35 0x78.

Then I made some printk at kernel init and here is the output:

begin dump of fadt:
[...]
6e	0
6f	0
70	f1
71	a2
72	35
73	78
74	0
75	0
[...]
bit wb_invd is 1
bit wb_invd_flush is 0
bit proc_c1 is 0
bit plvl2_up is 0
bit pwr_button is 1
bit sleep_button is 1
bit fixed_rTC is 1
bit rtcs4 is 1
bit tmr_val_ext is 0
bit dock_cap is 1
bit reset_reg_sup is 0
bit sealed_case is 0
bit headless is 0
bit cpu_sw_sleep is 1

gcc-2.95 and gcc-3.3 give the same thing.


Is it really what it is supposed to be?

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 05:15:28PM +0100, maxime bizon wrote:
> 
> Hello,
> 
> I'm using kernel 2.6.0-test9 on a i386 notebook.
> 
> I forced a new FADT table in the kernel and put the following bytes at
> offset 0x70:  0xf1 0xa2 0x35 0x78.
> 
> Then I made some printk at kernel init and here is the output:
> 
> begin dump of fadt:
> [...]
> 6e	0
> 6f	0
> 70	f1
> 71	a2
> 72	35
> 73	78
> 74	0
> 75	0
> [...]
> bit wb_invd is 1
> bit wb_invd_flush is 0
> bit proc_c1 is 0
> bit plvl2_up is 0
> bit pwr_button is 1
> bit sleep_button is 1
> bit fixed_rTC is 1
> bit rtcs4 is 1
> bit tmr_val_ext is 0
> bit dock_cap is 1
> bit reset_reg_sup is 0
> bit sealed_case is 0
> bit headless is 0
> bit cpu_sw_sleep is 1
> 
> gcc-2.95 and gcc-3.3 give the same thing.
> 
> 
> Is it really what it is supposed to be?

To be what?  What you want to do, exactly?

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[maximE bizon]

On Tue, Nov 11, 2003 at 06:02:45PM +0100, Ducrot Bruno wrote:

> > begin dump of fadt:
> > [...]
> > 6e	0
> > 6f	0
> > 70	f1
> > 71	a2
> > 72	35
> > 73	78
> > 74	0
> > 75	0
> > [...]
> > bit wb_invd is 1
> > bit wb_invd_flush is 0
> > bit proc_c1 is 0
> > bit plvl2_up is 0
> > bit pwr_button is 1
> > bit sleep_button is 1
> > bit fixed_rTC is 1
> > bit rtcs4 is 1
> > bit tmr_val_ext is 0
> > bit dock_cap is 1
> > bit reset_reg_sup is 0
> > bit sealed_case is 0
> > bit headless is 0
> > bit cpu_sw_sleep is 1
> > 

> > Is it really what it is supposed to be?
> 
> To be what?

According to the ACPI specs:

value		bit offset
=====		==========
wb_invd		0
wb_invd_flush	1
proc_c1		2
plvl2_up	3
pwr_button	4
sleep_button	5
[...]

So, for instance, sleep_button should be 0, no ?

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Randy.Dunlap]

On Tue, 11 Nov 2003 17:15:28 +0100 maxime bizon <maxime.bizon-ZrNlpQwM5lQ@public.gmane.org> wrote:

| 
| Hello,
| 
| I'm using kernel 2.6.0-test9 on a i386 notebook.
| 
| I forced a new FADT table in the kernel and put the following bytes at
| offset 0x70:  0xf1 0xa2 0x35 0x78.
| 
| Then I made some printk at kernel init and here is the output:
| 
| begin dump of fadt:
| [...]
| 6e	0
| 6f	0
| 70	f1
| 71	a2
| 72	35
| 73	78
| 74	0
| 75	0
| [...]
| bit wb_invd is 1
| bit wb_invd_flush is 0
| bit proc_c1 is 0
| bit plvl2_up is 0
| bit pwr_button is 1
| bit sleep_button is 1
| bit fixed_rTC is 1
| bit rtcs4 is 1
| bit tmr_val_ext is 0
| bit dock_cap is 1
| bit reset_reg_sup is 0
| bit sealed_case is 0
| bit headless is 0
| bit cpu_sw_sleep is 1
| 
| gcc-2.95 and gcc-3.3 give the same thing.
| 
| Is it really what it is supposed to be?

It can be confusing, can't it?

Yes, I think that they are correct.
Look at the bit field as a u32:  it's value is 0x7835a2f1.

Now look at the bit numbers in that u32 value and assign bit field
names to them and it will match your list above.

--
~Randy
MOTD:  Always include version info.


simple bitfields program output:

[rddunlap-GKM3wtwTCu0@public.gmane.org src]$ ./bitfields
bits as u32 = 0x7835a2f1
bytes (hex):  f1 a2 35 78
individual bits:
wb_invd:        1
wb_invd_flush:  0
proc_c1:        0
plvl2_up:       0
pwr_button:     1
sleep_button:   1
fixed_rTC:      1
rtcs4:          1
tmr_val_ext:    0
dock_cap:       1
reset_reg_sup:  0
sealed_case:    0
headless:       0
cpu_sw_sleep:   1
reserved6:    0x1e0d6


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[maximE bizon]

On Tue, Nov 11, 2003 at 09:23:16AM -0800, Randy.Dunlap wrote:

> It can be confusing, can't it?
> 
> Yes, I think that they are correct.
> Look at the bit field as a u32:  it's value is 0x7835a2f1.
> 
> Now look at the bit numbers in that u32 value and assign bit field
> names to them and it will match your list above.

Yes, my mistake.

Thanks.

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 06:18:45PM +0100, maximE bizon wrote:
> 
> On Tue, Nov 11, 2003 at 06:02:45PM +0100, Ducrot Bruno wrote:
> 
> > > begin dump of fadt:
> > > [...]
> > > 6e	0
> > > 6f	0
> > > 70	f1
> > > 71	a2
> > > 72	35
> > > 73	78
> > > 74	0
> > > 75	0
> > > [...]
> > > bit wb_invd is 1
> > > bit wb_invd_flush is 0
> > > bit proc_c1 is 0
> > > bit plvl2_up is 0
> > > bit pwr_button is 1
> > > bit sleep_button is 1
> > > bit fixed_rTC is 1
> > > bit rtcs4 is 1
> > > bit tmr_val_ext is 0
> > > bit dock_cap is 1
> > > bit reset_reg_sup is 0
> > > bit sealed_case is 0
> > > bit headless is 0
> > > bit cpu_sw_sleep is 1
> > > 
> 
> > > Is it really what it is supposed to be?
> > 
> > To be what?
> 
> According to the ACPI specs:
> 
> value		bit offset
> =====		==========
> wb_invd		0
> wb_invd_flush	1
> proc_c1		2
> plvl2_up	3
> pwr_button	4
> sleep_button	5
> [...]
> 
> So, for instance, sleep_button should be 0, no ?

No.  This is a double word, so, with endianess issues, you
actually have overwriten with
0x7825a2f1

and bit 5 of 0x7825a2f1 is 1.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 06:18:45PM +0100, maximE bizon wrote:
> 
> On Tue, Nov 11, 2003 at 06:02:45PM +0100, Ducrot Bruno wrote:
> 
> > > begin dump of fadt:
> > > [...]
> > > 6e	0
> > > 6f	0
> > > 70	f1
> > > 71	a2
> > > 72	35
> > > 73	78
> > > 74	0
> > > 75	0
> > > [...]

You also forgot to answer to my second question:

What you want to do, exactly?


-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[maximE bizon]

On Tue, Nov 11, 2003 at 06:57:30PM +0100, Ducrot Bruno wrote:

> You also forgot to answer to my second question:
> 
> What you want to do, exactly?

I own a Latitude X200 notebook, with kernel 2.6.0-test9 installed, and
S3 doesn't work.  It was posted here a few days  ago.  S3 does suspend
the computer but it is impossible to turn it on again.

I  noticed that  my power  button  is detected  as a  FF because  FADT
pwr_button flag is not set, though the DSDT defines one.

I suspected the S3 problem to come from this, so I tried to correct my
FADT, but I forgot that the bitfield was on u32...

Now with  the correct fix, I  can see an  acpi event when I  press the
power button, but the system still won't come up after an S3.

Any idea ?

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 07:25:22PM +0100, maximE bizon wrote:
> 
> On Tue, Nov 11, 2003 at 06:57:30PM +0100, Ducrot Bruno wrote:
> 
> > You also forgot to answer to my second question:
> > 
> > What you want to do, exactly?
> 
> I own a Latitude X200 notebook, with kernel 2.6.0-test9 installed, and
> S3 doesn't work.  It was posted here a few days  ago.  S3 does suspend
> the computer but it is impossible to turn it on again.

S3 seems to be broken in that kernel, anyway.

> I  noticed that  my power  button  is detected  as a  FF because  FADT
> pwr_button flag is not set, though the DSDT defines one.
> 
> I suspected the S3 problem to come from this, so I tried to correct my
> FADT, but I forgot that the bitfield was on u32...

Why not just hacking the source, instead of overwritten the FADT, for
only one flag?

> 
> Now with  the correct fix, I  can see an  acpi event when I  press the
> power button, but the system still won't come up after an S3.
> 
> Any idea ?
> 

perhaps you can try a latest mm serie?

Cheers,

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[maxime bizon]

On Tue, Nov 11, 2003 at 07:57:17PM +0100, Ducrot Bruno wrote:

> S3 seems to be broken in that kernel, anyway.

I tried every release since 2.5.70.


> Why not  just hacking the  source, instead of overwritten  the FADT,
> for only one flag?

Had it worked, it would have been more convenient to provide a patched
FADT to other X200 users.


> perhaps you can try a latest mm serie?

I just tried, that doesn't change anything.

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 09:26:13PM +0100, maxime bizon wrote:
> 
> On Tue, Nov 11, 2003 at 07:57:17PM +0100, Ducrot Bruno wrote:
> 
> > S3 seems to be broken in that kernel, anyway.
> 
> I tried every release since 2.5.70.
> 

Ahem, I got success with 2.5.7x, or so.

I think also that I got success with 2.6.0-testX, but I don't remember
when it broke (should be for X=5 IIRC) even without the 'faulty' drivers.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[maximE bizon]

On Tue, Nov 11, 2003 at 09:34:33PM +0100, Ducrot Bruno wrote:

> Ahem, I got success with 2.5.7x, or so.
> 
> I think also that I got success with 2.6.0-testX, but I don't remember
> when it broke (should be for X=5 IIRC) even without the 'faulty' drivers.

Did it work for an X200 or for another laptop?

-- 
Maxime


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Ducrot Bruno]

On Tue, Nov 11, 2003 at 09:46:27PM +0100, maximE bizon wrote:
> 
> On Tue, Nov 11, 2003 at 09:34:33PM +0100, Ducrot Bruno wrote:
> 
> > Ahem, I got success with 2.5.7x, or so.
> > 
> > I think also that I got success with 2.6.0-testX, but I don't remember
> > when it broke (should be for X=5 IIRC) even without the 'faulty' drivers.
> 
> Did it work for an X200 or for another laptop?
> 

For another laptop.  Well, I supposed you had understood.  Sorry for
confusion.

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/

Re: Bitfield usage in FADT table?

[Karol Kozimor]

Thus wrote Ducrot Bruno:
> > I own a Latitude X200 notebook, with kernel 2.6.0-test9 installed, and
> > S3 doesn't work.  It was posted here a few days  ago.  S3 does suspend
> > the computer but it is impossible to turn it on again.
> S3 seems to be broken in that kernel, anyway.

Strange, it worked fine on my system, keyboard and USB problems aside.
Best regards,

-- 
Karol 'sziwan' Kozimor
sziwan-DETuoxkZsSqrDJvtcaxF/A@public.gmane.org


-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003,
16-19 November in Las Vegas. Learn firsthand the latest
developments in Apache, PHP, Perl, XML, Java, MySQL,
WebDAV, and more! http://www.apachecon.com/