LinuxLists.cc - Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

Hi Elvis,

On Sun, 09 Dec 2007 16:12:25 -0500, Elvis Pranskevichus wrote:
> I have exactly the same problem here on a Gigabyte GA-965G-DS3 motherboard
> based box:

Same motherboard as Mike has.

> it87: Found IT8718F chip at 0x290, revision 1
> it87: in3 is VCC (+5V)
> it87 it87.656: Failed to request region 0x290-0x297
> it87: probe of it87.656 failed with error -16
>
> git bisecting revealed the offending commit:
>
> a7839e960675b54: PNP: increase the maximum number of resources
>
> Happened between rc3 and rc4.
>
> > Either way, the overlapping areas smell like a BIOS bug, meaning that
> > you should look for an updated BIOS for your system first.
> >
> >> http://www.mikeserv.com/temp/config-2.6.24-rc4.txt
> >
>
> This indeed looks like a broken ACPI BIOS since the aforementioned commit
> touches only the PNP ACPI driver. I'm not sure how to work around this,
> though. Ideas?

Complaining to Gigabyte seems to be the best approach.

In the meantime, I guess that booting with pnpacpi=off should fix your
problem. But it might break something else; I'm not sure what the PNP
ACPI driver is good for in the first place.

--
Jean Delvare

2007-12-09 23:35:01

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sun, 9 Dec 2007 23:42:15 +0100
Jean Delvare <[email protected]> wrote:
> In the meantime, I guess that booting with pnpacpi=off should fix
> your problem. But it might break something else; I'm not sure what
> the PNP ACPI driver is good for in the first place.

Ahh, thanks guys. Yes, that did indeed work around the problem. My
sensors work again, and I don't appear to have any ill effects. At
least my devices are all working, and resource assignments (irq, io)
look ok.

I'll put an append in my lilo.conf for pnpacpi=off for now.

/proc/ioports with pnpacpi=off:
0000-001f : dma1
0020-0021 : pic1
0040-0043 : timer0
0050-0053 : timer1
0060-006f : keyboard
0070-0077 : rtc
0080-008f : dma page reg
00a0-00a1 : pic2
00c0-00df : dma2
00f0-00ff : fpu
0290-0297 : it87
0290-0297 : it87
03c0-03df : vga+
0400-047f : 0000:00:1f.0
0400-0403 : ACPI PM1a_EVT_BLK
0404-0405 : ACPI PM1a_CNT_BLK
0408-040b : ACPI PM_TMR
0410-0415 : ACPI CPU throttle
0428-042f : ACPI GPE0_BLK
0480-04bf : 0000:00:1f.0
0500-051f : 0000:00:1f.3
0500-051f : i801_smbus
0cf8-0cff : PCI conf1
7000-7fff : PCI Bus #02
8000-8fff : PCI Bus #04
9000-9fff : PCI Bus #03
a000-afff : PCI Bus #01
a000-a07f : 0000:01:00.0
b000-bfff : PCI Bus #05
b000-b0ff : 0000:05:02.0
b000-b0ff : 8139too
c000-c01f : 0000:00:1a.1
c400-c41f : 0000:00:1d.0
c800-c81f : 0000:00:1d.1
cc00-cc1f : 0000:00:1d.2
d000-d01f : 0000:00:1a.0
d400-d407 : 0000:00:1f.2
d800-d803 : 0000:00:1f.2
dc00-dc07 : 0000:00:1f.2
e000-e003 : 0000:00:1f.2
e400-e41f : 0000:00:1f.2

Mike Houston

2007-12-10 00:20:18

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sun, 9 Dec 2007 23:42:15 +0100
Jean Delvare <[email protected]> wrote:

> On Sun, 09 Dec 2007 16:12:25 -0500, Elvis Pranskevichus wrote:

> > This indeed looks like a broken ACPI BIOS since the
> > aforementioned commit touches only the PNP ACPI driver. I'm not
> > sure how to work around this, though. Ideas?
>
> Complaining to Gigabyte seems to be the best approach.

I just happen to have a Windows Vista installation on this box as
well, and I just thought to check. Sorry, I wish I'd have thought of
it sooner but I don't go there often. You folks might be interested
to know that Windows appears to have the same silly problem with the
i/o resources (from Device Manager):

[000000290 - 000000294] Motherboard resources
[000000290 - 00000029F] Motherboard resources

I don't have anything that reads sensors in Windows though, so I
couldn't tell you if it could access that it87 chip or not.

So this pretty much confirms that it's a motherboard/bios issue.

Mike Houston

2007-12-10 01:32:56

by Ed Sweetman

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

Mike Houston wrote:
> On Sun, 9 Dec 2007 23:42:15 +0100
> Jean Delvare <[email protected]> wrote:
>
>
>> On Sun, 09 Dec 2007 16:12:25 -0500, Elvis Pranskevichus wrote:
>>
>
>
>>> This indeed looks like a broken ACPI BIOS since the
>>> aforementioned commit touches only the PNP ACPI driver. I'm not
>>> sure how to work around this, though. Ideas?
>>>
>> Complaining to Gigabyte seems to be the best approach.
>>
>
> I just happen to have a Windows Vista installation on this box as
> well, and I just thought to check. Sorry, I wish I'd have thought of
> it sooner but I don't go there often. You folks might be interested
> to know that Windows appears to have the same silly problem with the
> i/o resources (from Device Manager):
>
> [000000290 - 000000294] Motherboard resources
> [000000290 - 00000029F] Motherboard resources
>
> I don't have anything that reads sensors in Windows though, so I
> couldn't tell you if it could access that it87 chip or not.
>
> So this pretty much confirms that it's a motherboard/bios issue.
>
> Mike Houston
>
>

I'm seeing this exact problem on an Asus Nforce4 based board. Prior to
moving to 2.6.24-rc4 it worked just fine. No additional acpi options
were selected in kernel config.

So add Asus A8N-E to the list of broken pnpacpi

2007-12-10 02:32:17

by Shaohua Li

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sun, 2007-12-09 at 23:04 +0100, Adrian Bunk wrote:
> On Sun, Dec 09, 2007 at 04:12:25PM -0500, Elvis Pranskevichus wrote:
> > Jean Delvare wrote:
> >
> > > Hi Mike,
> > >
> > > On Sat, 8 Dec 2007 21:22:34 -0500, Mike Houston wrote:
> > >> On Sun, 9 Dec 2007 01:05:54 +0100
> > >> Adrian Bunk <[email protected]> wrote:
> > >>
> > >> > On Tue, Dec 04, 2007 at 09:51:54PM -0500, Mike Houston wrote:
> > >> > > I finally got around to testing Linux 2.6.24 (2.6.24-rc4) and
> > >> > > found that the it87 driver fails to probe and consequently, my
> > >> > > sensors no longer work. This was fine with Linux 2.6.23.8 (the
> > >> > > last kernel I was using)
> > >> > >
> > >> > > The necessary modules load, but:
> > >> > >
> > >> > > it87: Found IT8718F chip at 0x290, revision 2
> > >> > > it87: in3 is VCC (+5V)
> > >> > > it87 it87.656: Failed to request region 0x290-0x297
> > >> > > it87: probe of it87.656 failed with error -16
> > >> > >
> > >> > > Coretemp still works.
> > >> > >
> > >> > > It appears it has something to do with the ioport range being
> > >> > > reserved for some reason:
> > >> > >
> > >> > > system 00:01: ioport range 0x290-0x29f has been reserved
> > >>
> > >> >
> > >> > Thanks for your report.
> > >> >
> > >> > Please also provide:
> > >> > - dmesg from 2.6.23.8
> > >> > - The output of "cat /proc/ioports" for both kernels
> > >>
> > >> Thanks Adrian, here is the information you have requested, for
> > >> both kernels (I have 2.6.23.9 now though where it87 still works)
> > >>
> > >> Linux 2.6.23.9:
> > >> http://www.mikeserv.com/temp/proc_ioports-2.6.23.9.txt
> > >> http://www.mikeserv.com/temp/dmesg-2.6.23.9.txt
> > >> http://www.mikeserv.com/temp/config-2.6.23.9.txt
> > >>
> > >> Linux 2.6.24-rc4:
> > >> http://www.mikeserv.com/temp/proc_ioports-2.6.24-rc4.txt
> > >> http://www.mikeserv.com/temp/dmesg-2.6.24-rc4.txt
> > >
> > > This one shows:
> > >
> > > system 00:01: ioport range 0x290-0x29f has been reserved
> > > (...)
> > > system 00:01: ioport range 0x290-0x294 has been reserved
> > >
> > > This is clearly not correct as both areas overlap. The second
> > > reservation is responsible for the it87 breakage, because it conflicts
> > > with what the it87 driver later attempts to request (0x290-0x297). The
> > > first is wrong as well (the IT87xxF environment controller I/O area is
> > > 8 port wide, not 16) but shouldn't be a problem in practice.
> > >
> > > These port reservations weren't happening in 2.6.23.9 according to your
> > > dmesg output for that kernel. I don't know what changed in this area
> > > since 2.6.23.9, maybe Bjorn or Adam (Cc'd) can tell.
> > >
> >
> > Hi,
> >
> > I have exactly the same problem here on a Gigabyte GA-965G-DS3 motherboard
> > based box:
> >
> > it87: Found IT8718F chip at 0x290, revision 1
> > it87: in3 is VCC (+5V)
> > it87 it87.656: Failed to request region 0x290-0x297
> > it87: probe of it87.656 failed with error -16
> >
> > git bisecting revealed the offending commit:
> >
> > a7839e960675b54: PNP: increase the maximum number of resources
> >
> > Happened between rc3 and rc4.
>
> Thanks for doing the work of bisecting!
>
> > > Either way, the overlapping areas smell like a BIOS bug, meaning that
> > > you should look for an updated BIOS for your system first.
> > >
> > >> http://www.mikeserv.com/temp/config-2.6.24-rc4.txt
> > >
> >
> > This indeed looks like a broken ACPI BIOS since the aforementioned commit
> > touches only the PNP ACPI driver. I'm not sure how to work around this,
> > though. Ideas?
>
> People responsible for this commit + ACPI maintainer added to Cc.
This should exist in previous kernel (before we remove acpi motherboard
driver) too. Basically it's a broken BIOS. Could below patch work around
it?

Thanks,
Shaohua

Index: linux/drivers/pnp/system.c
===================================================================
--- linux.orig/drivers/pnp/system.c 2007-12-10 10:17:46.000000000 +0800
+++ linux/drivers/pnp/system.c 2007-12-10 10:24:42.000000000 +0800
@@ -22,7 +22,7 @@ static const struct pnp_device_id pnp_de
{"", 0}
};

-static void reserve_range(struct pnp_dev *dev, resource_size_t start,
+static struct resource* reserve_range(struct pnp_dev *dev, resource_size_t start,
resource_size_t end, int port)
{
char *regionid;
@@ -31,16 +31,14 @@ static void reserve_range(struct pnp_dev

regionid = kmalloc(16, GFP_KERNEL);
if (!regionid)
- return;
+ return NULL;

snprintf(regionid, 16, "pnp %s", pnpid);
if (port)
res = request_region(start, end - start + 1, regionid);
else
res = request_mem_region(start, end - start + 1, regionid);
- if (res)
- res->flags &= ~IORESOURCE_BUSY;
- else
+ if (!res)
kfree(regionid);

/*
@@ -52,12 +50,17 @@ static void reserve_range(struct pnp_dev
port ? "ioport" : "iomem",
(unsigned long long) start, (unsigned long long) end,
res ? "has been" : "could not be");
+ return res;
}

static void reserve_resources_of_dev(struct pnp_dev *dev)
{
int i;
+ struct resource **res;

+ res = kzalloc(sizeof(struct resource *) * PNP_MAX_PORT, GFP_KERNEL);
+ if (!res)
+ return;
for (i = 0; i < PNP_MAX_PORT; i++) {
if (!pnp_port_valid(dev, i))
continue;
@@ -76,17 +79,28 @@ static void reserve_resources_of_dev(str
if (pnp_port_end(dev, i) < pnp_port_start(dev, i))
continue; /* invalid */

- reserve_range(dev, pnp_port_start(dev, i),
+ res[i] = reserve_range(dev, pnp_port_start(dev, i),
pnp_port_end(dev, i), 1);
}
+ for (i = 0; i < PNP_MAX_PORT; i++)
+ if (res[i])
+ res[i]->flags &= ~IORESOURCE_BUSY;
+ kfree(res);

+ res = kzalloc(sizeof(struct resource *) * PNP_MAX_MEM, GFP_KERNEL);
+ if (!res)
+ return;
for (i = 0; i < PNP_MAX_MEM; i++) {
if (!pnp_mem_valid(dev, i))
continue;

- reserve_range(dev, pnp_mem_start(dev, i),
+ res[i] = reserve_range(dev, pnp_mem_start(dev, i),
pnp_mem_end(dev, i), 0);
}
+ for (i = 0; i < PNP_MAX_MEM; i++)
+ if (res[i])
+ res[i]->flags &= ~IORESOURCE_BUSY;
+ kfree(res);
}

static int system_pnp_probe(struct pnp_dev *dev,

2007-12-10 03:16:32

by Elvis Pranskevichus

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sunday December 9 2007 09:31:27 pm Shaohua Li wrote:
> On Sun, 2007-12-09 at 23:04 +0100, Adrian Bunk wrote:
> > On Sun, Dec 09, 2007 at 04:12:25PM -0500, Elvis Pranskevichus wrote:
> > > Jean Delvare wrote:
> > > > Hi Mike,
> > > >
> > > > On Sat, 8 Dec 2007 21:22:34 -0500, Mike Houston wrote:
> > > >> On Sun, 9 Dec 2007 01:05:54 +0100
> > > >>
> > > >> Adrian Bunk <[email protected]> wrote:
> > > >> > On Tue, Dec 04, 2007 at 09:51:54PM -0500, Mike Houston wrote:
> > > >> > > I finally got around to testing Linux 2.6.24 (2.6.24-rc4) and
> > > >> > > found that the it87 driver fails to probe and consequently, my
> > > >> > > sensors no longer work. This was fine with Linux 2.6.23.8 (the
> > > >> > > last kernel I was using)
> > > >> > >
> > > >> > > The necessary modules load, but:
> > > >> > >
> > > >> > > it87: Found IT8718F chip at 0x290, revision 2
> > > >> > > it87: in3 is VCC (+5V)
> > > >> > > it87 it87.656: Failed to request region 0x290-0x297
> > > >> > > it87: probe of it87.656 failed with error -16
> > > >> > >
> > > >> > > Coretemp still works.
> > > >> > >
> > > >> > > It appears it has something to do with the ioport range being
> > > >> > > reserved for some reason:
> > > >> > >
> > > >> > > system 00:01: ioport range 0x290-0x29f has been reserved
> > > >> >
> > > >> > Thanks for your report.
> > > >> >
> > > >> > Please also provide:
> > > >> > - dmesg from 2.6.23.8
> > > >> > - The output of "cat /proc/ioports" for both kernels
> > > >>
> > > >> Thanks Adrian, here is the information you have requested, for
> > > >> both kernels (I have 2.6.23.9 now though where it87 still works)
> > > >>
> > > >> Linux 2.6.23.9:
> > > >> http://www.mikeserv.com/temp/proc_ioports-2.6.23.9.txt
> > > >> http://www.mikeserv.com/temp/dmesg-2.6.23.9.txt
> > > >> http://www.mikeserv.com/temp/config-2.6.23.9.txt
> > > >>
> > > >> Linux 2.6.24-rc4:
> > > >> http://www.mikeserv.com/temp/proc_ioports-2.6.24-rc4.txt
> > > >> http://www.mikeserv.com/temp/dmesg-2.6.24-rc4.txt
> > > >
> > > > This one shows:
> > > >
> > > > system 00:01: ioport range 0x290-0x29f has been reserved
> > > > (...)
> > > > system 00:01: ioport range 0x290-0x294 has been reserved
> > > >
> > > > This is clearly not correct as both areas overlap. The second
> > > > reservation is responsible for the it87 breakage, because it
> > > > conflicts with what the it87 driver later attempts to request
> > > > (0x290-0x297). The first is wrong as well (the IT87xxF environment
> > > > controller I/O area is 8 port wide, not 16) but shouldn't be a
> > > > problem in practice.
> > > >
> > > > These port reservations weren't happening in 2.6.23.9 according to
> > > > your dmesg output for that kernel. I don't know what changed in this
> > > > area since 2.6.23.9, maybe Bjorn or Adam (Cc'd) can tell.
> > >
> > > Hi,
> > >
> > > I have exactly the same problem here on a Gigabyte GA-965G-DS3
> > > motherboard based box:
> > >
> > > it87: Found IT8718F chip at 0x290, revision 1
> > > it87: in3 is VCC (+5V)
> > > it87 it87.656: Failed to request region 0x290-0x297
> > > it87: probe of it87.656 failed with error -16
> > >
> > > git bisecting revealed the offending commit:
> > >
> > > a7839e960675b54: PNP: increase the maximum number of resources
> > >
> > > Happened between rc3 and rc4.
> >
> > Thanks for doing the work of bisecting!
> >
> > > > Either way, the overlapping areas smell like a BIOS bug, meaning that
> > > > you should look for an updated BIOS for your system first.
> > > >
> > > >> http://www.mikeserv.com/temp/config-2.6.24-rc4.txt
> > >
> > > This indeed looks like a broken ACPI BIOS since the aforementioned
> > > commit touches only the PNP ACPI driver. I'm not sure how to work
> > > around this, though. Ideas?
> >
> > People responsible for this commit + ACPI maintainer added to Cc.
>
> This should exist in previous kernel (before we remove acpi motherboard
> driver) too. Basically it's a broken BIOS. Could below patch work around
> it?
>
> Thanks,
> Shaohua
>
> Index: linux/drivers/pnp/system.c
> ===================================================================
> --- linux.orig/drivers/pnp/system.c 2007-12-10 10:17:46.000000000 +0800
> +++ linux/drivers/pnp/system.c 2007-12-10 10:24:42.000000000 +0800
> @@ -22,7 +22,7 @@ static const struct pnp_device_id pnp_de
> {"", 0}
> };
>
> -static void reserve_range(struct pnp_dev *dev, resource_size_t start,
> +static struct resource* reserve_range(struct pnp_dev *dev, resource_size_t
> start, resource_size_t end, int port)
> {
> char *regionid;
> @@ -31,16 +31,14 @@ static void reserve_range(struct pnp_dev
>
> regionid = kmalloc(16, GFP_KERNEL);
> if (!regionid)
> - return;
> + return NULL;
>
> snprintf(regionid, 16, "pnp %s", pnpid);
> if (port)
> res = request_region(start, end - start + 1, regionid);
> else
> res = request_mem_region(start, end - start + 1, regionid);
> - if (res)
> - res->flags &= ~IORESOURCE_BUSY;
> - else
> + if (!res)
> kfree(regionid);
>
> /*
> @@ -52,12 +50,17 @@ static void reserve_range(struct pnp_dev
> port ? "ioport" : "iomem",
> (unsigned long long) start, (unsigned long long) end,
> res ? "has been" : "could not be");
> + return res;
> }
>
> static void reserve_resources_of_dev(struct pnp_dev *dev)
> {
> int i;
> + struct resource **res;
>
> + res = kzalloc(sizeof(struct resource *) * PNP_MAX_PORT, GFP_KERNEL);
> + if (!res)
> + return;
> for (i = 0; i < PNP_MAX_PORT; i++) {
> if (!pnp_port_valid(dev, i))
> continue;
> @@ -76,17 +79,28 @@ static void reserve_resources_of_dev(str
> if (pnp_port_end(dev, i) < pnp_port_start(dev, i))
> continue; /* invalid */
>
> - reserve_range(dev, pnp_port_start(dev, i),
> + res[i] = reserve_range(dev, pnp_port_start(dev, i),
> pnp_port_end(dev, i), 1);
> }
> + for (i = 0; i < PNP_MAX_PORT; i++)
> + if (res[i])
> + res[i]->flags &= ~IORESOURCE_BUSY;
> + kfree(res);
>
> + res = kzalloc(sizeof(struct resource *) * PNP_MAX_MEM, GFP_KERNEL);
> + if (!res)
> + return;
> for (i = 0; i < PNP_MAX_MEM; i++) {
> if (!pnp_mem_valid(dev, i))
> continue;
>
> - reserve_range(dev, pnp_mem_start(dev, i),
> + res[i] = reserve_range(dev, pnp_mem_start(dev, i),
> pnp_mem_end(dev, i), 0);
> }
> + for (i = 0; i < PNP_MAX_MEM; i++)
> + if (res[i])
> + res[i]->flags &= ~IORESOURCE_BUSY;
> + kfree(res);
> }
>
> static int system_pnp_probe(struct pnp_dev *dev,

Hi Shaohua,

Your patch works for me. /proc/ioports looks sane now:

0290-029f : pnp 00:01
0290-0297 : it87
0290-0297 : it87

And it87 doesn't complain anymore.

Thanks!
--
Elvis

2007-12-10 04:02:25

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Mon, 10 Dec 2007 10:31:27 +0800
Shaohua Li <[email protected]> wrote:
> This should exist in previous kernel (before we remove acpi
> motherboard driver) too. Basically it's a broken BIOS. Could below
> patch work around it?
>
> Thanks,
> Shaohua
>
> Index: linux/drivers/pnp/system.c
> ===================================================================
> --- linux.orig/drivers/pnp/system.c 2007-12-10
> 10:17:46.000000000 +0800 +++ linux/drivers/pnp/system.c

Thanks Shaohua, I tested this as well and it appears to have worked
around the issue for me.

Now, in dmesg, I get:

system 00:01: ioport range 0x290-0x29f has been reserved
(...)
system 00:01: ioport range 0x290-0x294 could not be reserved

In /proc/ioports I see:

0290-029f : pnp 00:01
0290-0297 : it87
0290-0297 : it87

The it87 sensor now works without disabling acpipnp

Mike Houston

2007-12-10 14:56:04

[permalink] [raw]

Subject: Re: 2.6.24-rc4 hwmon it87 probe fails

Hi Ed,

On Sun, 09 Dec 2007 20:32:41 -0500, Ed Sweetman wrote:
> I'm seeing this exact problem on an Asus Nforce4 based board. Prior to
> moving to 2.6.24-rc4 it worked just fine. No additional acpi options
> were selected in kernel config.
>
> So add Asus A8N-E to the list of broken pnpacpi

For the records, can you please paste the offending lines (from the
boot logs or /proc/ioports)?

Then please give a try to Shaohua's patch and report:
http://lkml.org/lkml/diff/2007/12/9/186/1

Thanks,
--
Jean Delvare

2007-12-17 02:11:19

by Shaohua Li

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sun, 2007-12-09 at 23:02 -0500, Mike Houston wrote:
> On Mon, 10 Dec 2007 10:31:27 +0800
> Shaohua Li <[email protected]> wrote:
> > This should exist in previous kernel (before we remove acpi
> > motherboard driver) too. Basically it's a broken BIOS. Could below
> > patch work around it?
> >
> > Thanks,
> > Shaohua
> >
> > Index: linux/drivers/pnp/system.c
> > ===================================================================
> > --- linux.orig/drivers/pnp/system.c 2007-12-10
> > 10:17:46.000000000 +0800 +++ linux/drivers/pnp/system.c
>
> Thanks Shaohua, I tested this as well and it appears to have worked
> around the issue for me.
>
> Now, in dmesg, I get:
>
> system 00:01: ioport range 0x290-0x29f has been reserved
> (...)
> system 00:01: ioport range 0x290-0x294 could not be reserved
>
> In /proc/ioports I see:
>
> 0290-029f : pnp 00:01
> 0290-0297 : it87
> 0290-0297 : it87
Unfortunately this can't solve all such issues.

Adam & Bjorn,
Could we just reserve IO ports >= 0x1000 in pnp system driver? The
purpose of the driver is to avoid resource conflict with PCI device, and
PCI device can't user io port < 0x1000.

Thanks,
Shaohua

2007-12-17 17:15:00

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sunday 16 December 2007 06:59:39 pm Shaohua Li wrote:
> On Sun, 2007-12-09 at 23:02 -0500, Mike Houston wrote:
> > On Mon, 10 Dec 2007 10:31:27 +0800
> > Shaohua Li <[email protected]> wrote:
> > > This should exist in previous kernel (before we remove acpi
> > > motherboard driver) too. Basically it's a broken BIOS. Could below
> > > patch work around it?
> > >
> > > Thanks,
> > > Shaohua
> > >
> > > Index: linux/drivers/pnp/system.c
> > > ===================================================================
> > > --- linux.orig/drivers/pnp/system.c 2007-12-10
> > > 10:17:46.000000000 +0800 +++ linux/drivers/pnp/system.c
> >
> > Thanks Shaohua, I tested this as well and it appears to have worked
> > around the issue for me.
> >
> > Now, in dmesg, I get:
> >
> > system 00:01: ioport range 0x290-0x29f has been reserved
> > (...)
> > system 00:01: ioport range 0x290-0x294 could not be reserved
> >
> > In /proc/ioports I see:
> >
> > 0290-029f : pnp 00:01
> > 0290-0297 : it87
> > 0290-0297 : it87
> Unfortunately this can't solve all such issues.
>
> Adam & Bjorn,
> Could we just reserve IO ports >= 0x1000 in pnp system driver? The
> purpose of the driver is to avoid resource conflict with PCI device, and
> PCI device can't user io port < 0x1000.

The purpose of the PNP system driver is to avoid conflicts with
*all* devices. And if the PNP core were a little smarter, we
wouldn't need the system driver at all. We don't have one for
PCI -- the PCI core manages resources for all PCI devices, even
ones that have no driver.

Why is 0x1000 a magic number? drivers/acpi/motherboard.c used to
ignore IO port ranges that ended below PCIBIOS_MIN_IO (== 0x1000 for
most architectures). I don't think Linux will assign IO ports below
PCIBIOS_MIN_IO to a PCI device, but the BIOS could, and I've seen
CardBus devices below PCIBIOS_MIN_IO.

I think having drivers/pnp/system.c ignore resources below
PCIBIOS_MIN_IO would be a hack that happens to cover up problems
like this without understanding the real cause.

Bjorn

2007-12-18 17:59:30

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

Hi Bjorn,

On Mon, 17 Dec 2007 10:14:43 -0700, Bjorn Helgaas wrote:
> On Sunday 16 December 2007 06:59:39 pm Shaohua Li wrote:
> > On Sun, 2007-12-09 at 23:02 -0500, Mike Houston wrote:
> > > On Mon, 10 Dec 2007 10:31:27 +0800
> > > Shaohua Li <[email protected]> wrote:
> > > > This should exist in previous kernel (before we remove acpi
> > > > motherboard driver) too. Basically it's a broken BIOS. Could below
> > > > patch work around it?
> > > >
> > > > Thanks,
> > > > Shaohua
> > > >
> > > > Index: linux/drivers/pnp/system.c
> > > > ===================================================================
> > > > --- linux.orig/drivers/pnp/system.c 2007-12-10
> > > > 10:17:46.000000000 +0800 +++ linux/drivers/pnp/system.c
> > >
> > > Thanks Shaohua, I tested this as well and it appears to have worked
> > > around the issue for me.
> > >
> > > Now, in dmesg, I get:
> > >
> > > system 00:01: ioport range 0x290-0x29f has been reserved
> > > (...)
> > > system 00:01: ioport range 0x290-0x294 could not be reserved
> > >
> > > In /proc/ioports I see:
> > >
> > > 0290-029f : pnp 00:01
> > > 0290-0297 : it87
> > > 0290-0297 : it87
> >
> > Unfortunately this can't solve all such issues.

I don't see any issue here, but anyway it doesn't really matter as the
proposed patch is not sufficient for all affected motherboards.

> >
> > Adam & Bjorn,
> > Could we just reserve IO ports >= 0x1000 in pnp system driver? The
> > purpose of the driver is to avoid resource conflict with PCI device, and
> > PCI device can't user io port < 0x1000.
>
> The purpose of the PNP system driver is to avoid conflicts with
> *all* devices. And if the PNP core were a little smarter, we
> wouldn't need the system driver at all. We don't have one for
> PCI -- the PCI core manages resources for all PCI devices, even
> ones that have no driver.
>
> Why is 0x1000 a magic number? drivers/acpi/motherboard.c used to
> ignore IO port ranges that ended below PCIBIOS_MIN_IO (== 0x1000 for
> most architectures). I don't think Linux will assign IO ports below
> PCIBIOS_MIN_IO to a PCI device, but the BIOS could, and I've seen
> CardBus devices below PCIBIOS_MIN_IO.
>
> I think having drivers/pnp/system.c ignore resources below
> PCIBIOS_MIN_IO would be a hack that happens to cover up problems
> like this without understanding the real cause.

The real cause is pretty clear here: broken BIOS. In an ideal world we
would ask the manufacturer for a fixed BIOS and they would give that to
us, unfortunately my experience is that it won't happen. So, unless we
accept that idea that some users won't be able to use some of their
devices because of broken resource enumeration in their BIOS, we will
have to find a workaround, whatever it is.

My initial idea was to identify the faulty motherboard using DMI and to
force pnpacpi=off on the faulty motherboards. If this is considered too
aggressive, maybe we can just reject resource declarations that
intersect (but don't match) 0x290-0x297 for these motherboards. Either
way, we have to do something, and we have to do it quickly. 2.6.24
final isn't too far away, and more importantly, the patch that revealed
the problem has been backported to 2.6.23.10 so people are experiencing
regressions already.

--
Jean Delvare

2007-12-19 23:53:50

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sunday 09 December 2007 09:02:11 pm Mike Houston wrote:
> On Mon, 10 Dec 2007 10:31:27 +0800
> Shaohua Li <[email protected]> wrote:
> > This should exist in previous kernel (before we remove acpi
> > motherboard driver) too. Basically it's a broken BIOS. Could below
> > patch work around it?
> >
> > Thanks,
> > Shaohua
> >
> > Index: linux/drivers/pnp/system.c
> > ===================================================================
> > --- linux.orig/drivers/pnp/system.c 2007-12-10
> > 10:17:46.000000000 +0800 +++ linux/drivers/pnp/system.c
>
> Thanks Shaohua, I tested this as well and it appears to have worked
> around the issue for me.
>
> Now, in dmesg, I get:
>
> system 00:01: ioport range 0x290-0x29f has been reserved
> (...)
> system 00:01: ioport range 0x290-0x294 could not be reserved
>
> In /proc/ioports I see:
>
> 0290-029f : pnp 00:01
> 0290-0297 : it87
> 0290-0297 : it87
>
> The it87 sensor now works without disabling acpipnp

This whole thing is too ugly for words.

First of all, it87 only accesses two I/O ports: IT87_ADDR_REG_OFFSET
and IT87_DATA_REG_OFFSET, which are at offsets 5 and 6 from the base.
So if it87 reserved only those ports, e.g., 0x295-0x296, I think things
would work fine (or at least, as well as they used to).

Secondly, it87 ignores any information from the BIOS and grubs around
for a superio chip at a magic address and tries to figure out what's
behind it. I know this is because we don't know the "real" way to
read those sensors, but it still sucks, and it still means we're likely
to have conflicts with ACPI usage of the device. And no, I don't have
anything useful like an ACPI driver to offer :-)

Bjorn

2007-12-20 00:20:37

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Tuesday 18 December 2007 10:59:18 am Jean Delvare wrote:
> The real cause is pretty clear here: broken BIOS. In an ideal world we
> would ask the manufacturer for a fixed BIOS and they would give that to
> us, unfortunately my experience is that it won't happen. So, unless we
> accept that idea that some users won't be able to use some of their
> devices because of broken resource enumeration in their BIOS, we will
> have to find a workaround, whatever it is.

I suspect the manufacturers would say "Oh, the sensors? The BIOS
isn't broken, you're just supposed to use WMI or some (undocumented)
ACPI device to get at those."

I know next to nothing about WMI, and there are probably IP issues
in that area. But I'd rather spend effort in figuring out the "right"
way to do this. Until we do it the same way as Windows, we can pile
on hacks till the cows come home, and we'll still have issues.

I added Carlos to the cc: list because he's been doing a lot of
interesting-looking work on WMI.

> My initial idea was to identify the faulty motherboard using DMI and to
> force pnpacpi=off on the faulty motherboards. If this is considered too
> aggressive, maybe we can just reject resource declarations that
> intersect (but don't match) 0x290-0x297 for these motherboards. Either
> way, we have to do something, and we have to do it quickly. 2.6.24
> final isn't too far away, and more importantly, the patch that revealed
> the problem has been backported to 2.6.23.10 so people are experiencing
> regressions already.

Windows apparently doesn't reject those resource declarations, so I'm
a bit hesitant to do it in Linux. That tends to cover up problems,
and then it becomes very difficult to remove the band-aid and figure
out a correct fix in the future.

Bjorn

2007-12-20 00:47:18

by Carlos Corbacho

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Thursday 20 December 2007 00:20:21 Bjorn Helgaas wrote:
> I suspect the manufacturers would say "Oh, the sensors? The BIOS
> isn't broken, you're just supposed to use WMI or some (undocumented)
> ACPI device to get at those."

It's quite possible - can we have DSDTs for the boards in question so we can
quickly check if this is a possibility? (Basically, to see if they have
PNP0C14 devices - if they don't, then I'm afraid it's nothing to do with
WMI).

-Carlos
--
E-Mail: [email protected]
Web: strangeworlds.co.uk
GPG Key ID: 0x23EE722D

2007-12-20 02:14:08

by Elvis Pranskevichus

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Wednesday December 19 2007 07:45:14 pm Carlos Corbacho wrote:
> On Thursday 20 December 2007 00:20:21 Bjorn Helgaas wrote:
> > I suspect the manufacturers would say "Oh, the sensors? The BIOS
> > isn't broken, you're just supposed to use WMI or some (undocumented)
> > ACPI device to get at those."
>
> It's quite possible - can we have DSDTs for the boards in question so we
> can quickly check if this is a possibility? (Basically, to see if they have
> PNP0C14 devices - if they don't, then I'm afraid it's nothing to do with
> WMI).
>
> -Carlos

Hi Carlos,

I've attached the DSDT for Gigabyte GA-965G-DS3 (rev 1.0, bios rev. F9) to
bugzilla entry #9514:

http://bugzilla.kernel.org/attachment.cgi?id=14132

Thanks,
--
Elvis

2007-12-20 02:18:55

by Carlos Corbacho

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Thursday 20 December 2007 02:13:22 Elvis Pranskevichus wrote:
> Hi Carlos,
>
> I've attached the DSDT for Gigabyte GA-965G-DS3 (rev 1.0, bios rev. F9) to
> bugzilla entry #9514:
>
> http://bugzilla.kernel.org/attachment.cgi?id=14132

A quick look over the DSDT shows that there is no ACPI-WMI mapper device
(PNP0C14) on this board; so WMI is, at least in this case, not the solution
to this bug.

-Carlos
--
E-Mail: [email protected]
Web: strangeworlds.co.uk
GPG Key ID: 0x23EE722D

2007-12-21 19:00:46

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Tuesday 18 December 2007 10:59:18 am Jean Delvare wrote:
> My initial idea was to identify the faulty motherboard using DMI and to
> force pnpacpi=off on the faulty motherboards. If this is considered too
> aggressive, maybe we can just reject resource declarations that
> intersect (but don't match) 0x290-0x297 for these motherboards. Either
> way, we have to do something, and we have to do it quickly. 2.6.24
> final isn't too far away, and more importantly, the patch that revealed
> the problem has been backported to 2.6.23.10 so people are experiencing
> regressions already.

What do you think of something like the following patch? If we do
this, I don't think we'd need to force pnpacpi=off or change the
way PNP reserves resources.

I'll be on vacation until about January 2, so I won't be able to
do much with this until then.

[patch] it87: request only Environment Controller ports

The IT8705F and related parts are Super I/O controllers that contain
many separate devices.

Some BIOSes describe IT8705F I/O port usage under a motherboard device
(PNP0C02) with overlapping regions, e.g., 0x290-0x29f and 0x290-0x294.

The it87 driver supports only the Environment Controller, which requires
only two ISA ports, but it used to request an eight-port range. If that
range exceeds a range reported by the BIOS, as 0x290-0x297 would, the
request fails, and the it87 driver cannot claim the device.

This patch makes the it87 driver request only the two ports used for the
Environment Controller device.

Systems where this problem has been reported:
Gigabyte GA-K8N Ultra 9
Gigabyte M56S-S3
Gigabyte GA-965G-DS3

Kernel bug reports:
http://bugzilla.kernel.org/show_bug.cgi?id=9514
http://lkml.org/lkml/2007/12/4/466

Related change:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=a7839e960675b549f06209d18283d5cee2ce9261

The patch above increases the number of PNP port resources we support.
Prior to this patch, we ignored some port resources, which masked the
it87 problem.

Signed-off-by: Bjorn Helgaas <[email protected]>

Index: work4/drivers/hwmon/it87.c
===================================================================
--- work4.orig/drivers/hwmon/it87.c 2007-12-21 10:38:46.000000000 -0700
+++ work4/drivers/hwmon/it87.c 2007-12-21 11:43:50.000000000 -0700
@@ -2,6 +2,14 @@
it87.c - Part of lm_sensors, Linux kernel modules for hardware
monitoring.

+ The IT8705F is an LPC-based Super I/O part that contains UARTs, a
+ parallel port, an IR port, a MIDI port, a floppy controller, etc., in
+ addition to an Environment Controller (Enhanced Hardware Monitor and
+ Fan Controller)
+
+ This driver supports only the Environment Controller in the IT8705F and
+ similar parts. The other devices are supported by different drivers.
+
Supports: IT8705F Super I/O chip w/LPC interface
IT8712F Super I/O chip w/LPC interface
IT8716F Super I/O chip w/LPC interface
@@ -118,9 +126,15 @@
/* Length of ISA address segment */
#define IT87_EXTENT 8

-/* Where are the ISA address/data registers relative to the base address */
-#define IT87_ADDR_REG_OFFSET 5
-#define IT87_DATA_REG_OFFSET 6
+/* Length of ISA address segment for Environmental Controller */
+#define IT87_EC_EXTENT 2
+
+/* Offset of EC registers from ISA base address */
+#define IT87_EC_OFFSET 5
+
+/* Where are the ISA address/data registers relative to the EC base address */
+#define IT87_ADDR_REG_OFFSET 0
+#define IT87_DATA_REG_OFFSET 1

/*----- The IT87 registers -----*/

@@ -968,10 +982,10 @@
};

res = platform_get_resource(pdev, IORESOURCE_IO, 0);
- if (!request_region(res->start, IT87_EXTENT, DRVNAME)) {
+ if (!request_region(res->start, IT87_EC_EXTENT, DRVNAME)) {
dev_err(dev, "Failed to request region 0x%lx-0x%lx\n",
(unsigned long)res->start,
- (unsigned long)(res->start + IT87_EXTENT - 1));
+ (unsigned long)(res->start + IT87_EC_EXTENT - 1));
err = -EBUSY;
goto ERROR0;
}
@@ -1124,7 +1138,7 @@
platform_set_drvdata(pdev, NULL);
kfree(data);
ERROR1:
- release_region(res->start, IT87_EXTENT);
+ release_region(res->start, IT87_EC_EXTENT);
ERROR0:
return err;
}
@@ -1137,7 +1151,7 @@
sysfs_remove_group(&pdev->dev.kobj, &it87_group);
sysfs_remove_group(&pdev->dev.kobj, &it87_group_opt);

- release_region(data->addr, IT87_EXTENT);
+ release_region(data->addr, IT87_EC_EXTENT);
platform_set_drvdata(pdev, NULL);
kfree(data);

@@ -1402,8 +1416,8 @@
const struct it87_sio_data *sio_data)
{
struct resource res = {
- .start = address ,
- .end = address + IT87_EXTENT - 1,
+ .start = address + IT87_EC_OFFSET,
+ .end = address + IT87_EC_OFFSET + IT87_EC_EXTENT - 1,
.name = DRVNAME,
.flags = IORESOURCE_IO,
};

2007-12-21 19:50:33

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Fri, 21 Dec 2007 12:00:30 -0700
Bjorn Helgaas <[email protected]> wrote:

> What do you think of something like the following patch? If we do
> this, I don't think we'd need to force pnpacpi=off or change the
> way PNP reserves resources.
>
> I'll be on vacation until about January 2, so I won't be able to
> do much with this until then.
>
>
>
> [patch] it87: request only Environment Controller ports

I'm not qualified to give an opinion on what's proper for the
kernel, so all I can say is that I've applied it (to fresh 2.6.24-rc6
source without Shaohua's pnpacpi patch) and it works nicely for me.

>From /proc/ioports:

0290-029f : pnp 00:01
0290-0294 : pnp 00:01
0295-0296 : it87
0295-0296 : it87

>From dmesg:

system 00:01: ioport range 0x290-0x29f has been reserved
...
system 00:01: ioport range 0x290-0x294 has been reserved

My sensors work.

Thanks for this patch,

Mike Houston

2007-12-22 11:58:58

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

Hi Bjorn,

Le 21/12/2007, "Bjorn Helgaas" <[email protected]> a ?crit:

>On Tuesday 18 December 2007 10:59:18 am Jean Delvare wrote:
>> My initial idea was to identify the faulty motherboard using DMI and to
>> force pnpacpi=off on the faulty motherboards. If this is considered too
>> aggressive, maybe we can just reject resource declarations that
>> intersect (but don't match) 0x290-0x297 for these motherboards. Either
>> way, we have to do something, and we have to do it quickly. 2.6.24
>> final isn't too far away, and more importantly, the patch that revealed
>> the problem has been backported to 2.6.23.10 so people are experiencing
>> regressions already.
>
>What do you think of something like the following patch? If we do
>this, I don't think we'd need to force pnpacpi=off or change the
>way PNP reserves resources.
>
>I'll be on vacation until about January 2, so I won't be able to
>do much with this until then.
>
>
>
>[patch] it87: request only Environment Controller ports
>
>The IT8705F and related parts are Super I/O controllers that contain
>many separate devices.
>
>Some BIOSes describe IT8705F I/O port usage under a motherboard device
>(PNP0C02) with overlapping regions, e.g., 0x290-0x29f and 0x290-0x294.
>
>The it87 driver supports only the Environment Controller, which requires
>only two ISA ports, but it used to request an eight-port range. If that
>range exceeds a range reported by the BIOS, as 0x290-0x297 would, the
>request fails, and the it87 driver cannot claim the device.
>
>This patch makes the it87 driver request only the two ports used for the
>Environment Controller device.

The problem is that the IT87xxF chips do decode 4 ports (recent chips,
0x294-0x297) or 8 ports (older chips, 0x290-0x297), not 2 as the
datasheets say. The ITE Super-I/O chips differ from the Winbond
Super-I/O chips in this respect. The it87 driver only needs to access
ports 0x295 and 0x296, true, but the device itself decodes more
addresses than that. So, with this proposed patch, ports which are busy
will be shown as being free. This pretty much voids the point of
resource declarations, doesn't it? This might not cause too much
trouble in practice, but to me this still looks like the wrong way to go.

The resource declarations made by these motherboard BIOSes are totally
bogus. 0x290-0x29f is much larger than what the chip decodes.
0x290-0x294 is a subset that doesn't make any sense. The GA-K8N Ultra 9
is even funnier with 0x295-0x314, which again doesn't correspond to
anything real. All these happen to not intersect with 0x295-0x296 but I
wouldn't count on it. If Gigabyte (and possibly others) care that
little about these declarations, pretty much anything could be seen. So
while your proposed workaround happens to fix the problem at hand, it is
not really correct technically, and could break again soon.

I'd rather fix the problem at its source, or, as fixing it as the source
isn't very realistic in this case, as near of the source as possible.
That would mean DMI-based quirks for the affected motherboards, that
would discard or adjust the bogus resource declarations. I have not yet
looked at what the code would look like, as I am not familiar with the
pnp system driver, I don't know what it would take to implement my
proposal.

I also don't fully understand what pnpacpi is useful for. I have heard
about PCI drivers that might request resources that the motherboard
doesn't want them to touch, but I don't know the details, I also
don't know if it is a theoretical issue or if it really happens on some
systems, and I don't know if there are other uses for pnpacpi. I'm
asking because I want to know if forcing pnpacpi=off on the faulty
motherboards would be good enough or if we really need finer-grained
quirks (assuming that we go the DMI-based quirks route at all.) The
former would obviously be easier.

>
>Systems where this problem has been reported:
> Gigabyte GA-K8N Ultra 9
> Gigabyte M56S-S3
> Gigabyte GA-965G-DS3
> (...)

Thanks,
--
Jean Delvare

2007-12-23 03:42:35

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Saturday 22 December 2007 4:21:41 am Jean Delvare wrote:
> >This patch makes the it87 driver request only the two ports used for the
> >Environment Controller device.
>
> The problem is that the IT87xxF chips do decode 4 ports (recent chips,
> 0x294-0x297) or 8 ports (older chips, 0x290-0x297), not 2 as the
> datasheets say. The ITE Super-I/O chips differ from the Winbond
> Super-I/O chips in this respect. The it87 driver only needs to access
> ports 0x295 and 0x296, true, but the device itself decodes more
> addresses than that. So, with this proposed patch, ports which are busy
> will be shown as being free. This pretty much voids the point of
> resource declarations, doesn't it? This might not cause too much
> trouble in practice, but to me this still looks like the wrong way to go.

Yes, all the ports decoded by the chip should certainly be reserved,
but I think the entire range should be reserved at a higher level,
like the PNP core, and the driver should reserve only the ports it
uses. Then the entire range is reserved even if there is no driver
or the driver is not loaded. That's the approach we use for PCI,
e.g.,

e8100000-e81fffff : 0000:00:08.0 <-- reserved by PCI core
e8100000-e8102fff : CS46xx_BA1_data0 <-- reserved by driver
e8110000-e81137ff : CS46xx_BA1_data1
e8120000-e8126fff : CS46xx_BA1_pmem
e8130000-e81300ff : CS46xx_BA1_reg

> The resource declarations made by these motherboard BIOSes are totally
> bogus. 0x290-0x29f is much larger than what the chip decodes.
> 0x290-0x294 is a subset that doesn't make any sense. The GA-K8N Ultra 9
> is even funnier with 0x295-0x314, which again doesn't correspond to
> anything real.

I agree those declarations are probably wrong. But at least they're
larger than required, so they should be safe.

> All these happen to not intersect with 0x295-0x296 but I
> wouldn't count on it. If Gigabyte (and possibly others) care that
> little about these declarations, pretty much anything could be seen. So
> while your proposed workaround happens to fix the problem at hand, it is
> not really correct technically, and could break again soon.
>
> I'd rather fix the problem at its source, or, as fixing it as the source
> isn't very realistic in this case, as near of the source as possible.
> That would mean DMI-based quirks for the affected motherboards, that
> would discard or adjust the bogus resource declarations.

Well, I think the driver change *is* correct, assuming that the
entire range can be reserved at a higher level. In this case,
it is, via a PNP0C02 device.

I think a DMI-based quirk to fix the PNP0C02 resources would be
a good approach, but we shouldn't do that until those resources
cause some other problem.

> I also don't fully understand what pnpacpi is useful for. I have heard
> about PCI drivers that might request resources that the motherboard
> doesn't want them to touch, but I don't know the details, I also
> don't know if it is a theoretical issue or if it really happens on some
> systems, and I don't know if there are other uses for pnpacpi.

It is a real issue. It's probably not very common because many ACPI
devices use "legacy" resources like ports below 0x1000, and the kernel
has heuristics to avoid placing other devices there. But I have seen
the kernel place a CardBus device on top of an ACPI device, so it does
happen.

Apart from preventing resource conflicts, PNP has framework for
driver/device binding, suspend/resume, etc. Without PNPACPI,
we'll fall back to PNPBIOS, but my guess is that nowadays, the
ACPI namespace is better-tested than PNPBIOS. And I wouldn't
be surprised if PNPBIOS disappears from new boxes.

> I'm
> asking because I want to know if forcing pnpacpi=off on the faulty
> motherboards would be good enough or if we really need finer-grained
> quirks (assuming that we go the DMI-based quirks route at all.) The
> former would obviously be easier.

I'm sure these boxes still have PNPBIOS, so pnpacpi=off would
probably work fine. But it feels like a bigger hammer than
necessary for this problem.

Bjorn

2007-12-23 09:29:17

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

Hi Bjorn,

Le 23/12/2007, Bjorn Helgaas a ?crit:
>On Saturday 22 December 2007 4:21:41 am Jean Delvare wrote:
>> >This patch makes the it87 driver request only the two ports used for the
>> >Environment Controller device.
>>
>> The problem is that the IT87xxF chips do decode 4 ports (recent chips,
>> 0x294-0x297) or 8 ports (older chips, 0x290-0x297), not 2 as the
>> datasheets say. The ITE Super-I/O chips differ from the Winbond
>> Super-I/O chips in this respect. The it87 driver only needs to access
>> ports 0x295 and 0x296, true, but the device itself decodes more
>> addresses than that. So, with this proposed patch, ports which are busy
>> will be shown as being free. This pretty much voids the point of
>> resource declarations, doesn't it? This might not cause too much
>> trouble in practice, but to me this still looks like the wrong way to go.
>
>Yes, all the ports decoded by the chip should certainly be reserved,
>but I think the entire range should be reserved at a higher level,
>like the PNP core, and the driver should reserve only the ports it
>uses. Then the entire range is reserved even if there is no driver
>or the driver is not loaded.

The problem is that the it87 driver is used on a variety of motherboards,
some where the hardware monitoring device I/O ports are reserved by the
BIOS, some where they aren't. How am I supposed to deal with this?

>That's the approach we use for PCI, e.g.,
>
> e8100000-e81fffff : 0000:00:08.0 <-- reserved by PCI core
> e8100000-e8102fff : CS46xx_BA1_data0 <-- reserved by driver
> e8110000-e81137ff : CS46xx_BA1_data1
> e8120000-e8126fff : CS46xx_BA1_pmem
> e8130000-e81300ff : CS46xx_BA1_reg

PCI is an entirely different beast. With PCI you know the PCI device ID
that is associated with the resources, and for a given device, the
resources are always declared (if standard BARs are used) or never
declared; there's no "may be". So it's much easier to handle the
resources properly.

>> The resource declarations made by these motherboard BIOSes are totally
>> bogus. 0x290-0x29f is much larger than what the chip decodes.
>> 0x290-0x294 is a subset that doesn't make any sense. The GA-K8N Ultra 9
>> is even funnier with 0x295-0x314, which again doesn't correspond to
>> anything real.
>
>I agree those declarations are probably wrong. But at least they're
>larger than required, so they should be safe.

That's not really safe, no. They may overlap with other device resources
and prevent those drivers from loading.

>> All these happen to not intersect with 0x295-0x296 but I
>> wouldn't count on it. If Gigabyte (and possibly others) care that
>> little about these declarations, pretty much anything could be seen. So
>> while your proposed workaround happens to fix the problem at hand, it is
>> not really correct technically, and could break again soon.
>>
>> I'd rather fix the problem at its source, or, as fixing it as the source
>> isn't very realistic in this case, as near of the source as possible.
>> That would mean DMI-based quirks for the affected motherboards, that
>> would discard or adjust the bogus resource declarations.
>
>Well, I think the driver change *is* correct, assuming that the
>entire range can be reserved at a higher level. In this case,
>it is, via a PNP0C02 device.

As I wrote above, the problem is that you can't assume that. Some
motherboards do declare the range at PNP level but some don't. That's
the reason why the it87 driver (and most other hwmon drivers for
Super-I/O devices) do declare the I/O resource again.

Another problem is how do I connect this specific I/O port range of the
PNP0C02 device with the it87 driver? I am by far no PNP expert but it
looks to me like this PNP device covers more than one actual device, and
I/O port ranges do not have labels nor identifiers that would let me
find the one that corresponds to the IT87xxF device (if it exists at
all.)

I'm all for integrating the it87 driver into the PNP subsystem if it is
going to solve problems, but I just don't know how it works. I you do
some work in this direction, I'll be happy to help with reviews and
testing.

>I think a DMI-based quirk to fix the PNP0C02 resources would be
>a good approach, but we shouldn't do that until those resources
>cause some other problem.

Well these resources already do cause problems, otherwise we wouldn't be
having this discussion ;)

>(...)
>I'm sure these boxes still have PNPBIOS, so pnpacpi=off would
>probably work fine. But it feels like a bigger hammer than
>necessary for this problem.

OK.

--
Jean Delvare

2007-12-23 23:17:01

[permalink] [raw]

Subject: Re: [lm-sensors] 2.6.24-rc4 hwmon it87 probe fails

On Sunday 23 December 2007 2:28:05 am Jean Delvare wrote:
> Le 23/12/2007, Bjorn Helgaas a ?crit:
> >On Saturday 22 December 2007 4:21:41 am Jean Delvare wrote:
> >> >This patch makes the it87 driver request only the two ports used for
> >> > the Environment Controller device.
> >>
> >> The problem is that the IT87xxF chips do decode 4 ports (recent chips,
> >> 0x294-0x297) or 8 ports (older chips, 0x290-0x297), not 2 as the
> >> datasheets say. The ITE Super-I/O chips differ from the Winbond
> >> Super-I/O chips in this respect. The it87 driver only needs to access
> >> ports 0x295 and 0x296, true, but the device itself decodes more
> >> addresses than that. So, with this proposed patch, ports which are busy
> >> will be shown as being free. This pretty much voids the point of
> >> resource declarations, doesn't it? This might not cause too much
> >> trouble in practice, but to me this still looks like the wrong way to
> >> go.
> >
> >Yes, all the ports decoded by the chip should certainly be reserved,
> >but I think the entire range should be reserved at a higher level,
> >like the PNP core, and the driver should reserve only the ports it
> >uses. Then the entire range is reserved even if there is no driver
> >or the driver is not loaded.
>
> The problem is that the it87 driver is used on a variety of motherboards,
> some where the hardware monitoring device I/O ports are reserved by the
> BIOS, some where they aren't. How am I supposed to deal with this?

I assume you mean some boards describe those ports in ACPI, and some
don't? I think the ideal solution would be to figure out how the
BIOS writers intended the device to be used, and do that. But the
documentation may be lacking, and it degenerates into an OEM-specific
mess. Second-best seems like a PNP quirk that pokes enough to determine
that a Super I/O device is present, and then reserves resources for
it. Then we avoid the collision even if it87 isn't present.

> >That's the approach we use for PCI, e.g.,
> >
> > e8100000-e81fffff : 0000:00:08.0 <-- reserved by PCI core
> > e8100000-e8102fff : CS46xx_BA1_data0 <-- reserved by driver
> > e8110000-e81137ff : CS46xx_BA1_data1
> > e8120000-e8126fff : CS46xx_BA1_pmem
> > e8130000-e81300ff : CS46xx_BA1_reg
>
> PCI is an entirely different beast. With PCI you know the PCI device ID
> that is associated with the resources, and for a given device, the
> resources are always declared (if standard BARs are used) or never
> declared; there's no "may be". So it's much easier to handle the
> resources properly.

That's the way PNP is supposed to work, too (more about this below).

> >> The resource declarations made by these motherboard BIOSes are totally
> >> bogus. 0x290-0x29f is much larger than what the chip decodes.
> >> 0x290-0x294 is a subset that doesn't make any sense. The GA-K8N Ultra 9
> >> is even funnier with 0x295-0x314, which again doesn't correspond to
> >> anything real.
> >
> >I agree those declarations are probably wrong. But at least they're
> >larger than required, so they should be safe.
>
> That's not really safe, no. They may overlap with other device resources
> and prevent those drivers from loading.

True. If that happens, I think we definitely need some kind of DMI-
based quirk to fix the resources. My points are (a) the quirk needs
to be at the PNP level; it can't be in a driver, and (b) I'm lazy and
would wait until a problem happens before implementing it, because I
know so little about PNP and the specific board, and by waiting, there's
a chance I'll learn enough to avoid a mistake :-)

> >> All these happen to not intersect with 0x295-0x296 but I
> >> wouldn't count on it. If Gigabyte (and possibly others) care that
> >> little about these declarations, pretty much anything could be seen. So
> >> while your proposed workaround happens to fix the problem at hand, it is
> >> not really correct technically, and could break again soon.
> >>
> >> I'd rather fix the problem at its source, or, as fixing it as the source
> >> isn't very realistic in this case, as near of the source as possible.
> >> That would mean DMI-based quirks for the affected motherboards, that
> >> would discard or adjust the bogus resource declarations.
> >
> >Well, I think the driver change *is* correct, assuming that the
> >entire range can be reserved at a higher level. In this case,
> >it is, via a PNP0C02 device.
>
> As I wrote above, the problem is that you can't assume that. Some
> motherboards do declare the range at PNP level but some don't. That's
> the reason why the it87 driver (and most other hwmon drivers for
> Super-I/O devices) do declare the I/O resource again.

The overlapping device problem is a subsystem problem, not a
driver problem. We can't solve it in the driver because we can't
depend on the it87 driver being loaded.

> Another problem is how do I connect this specific I/O port range of the
> PNP0C02 device with the it87 driver? I am by far no PNP expert but it
> looks to me like this PNP device covers more than one actual device, and
> I/O port ranges do not have labels nor identifiers that would let me
> find the one that corresponds to the IT87xxF device (if it exists at
> all.)

The general rule is that you have a PNP device with an identifier
like PNP0500 that means "16550 UART" and some resources underneath
it. The PNP ID ("PNP0500") tells the OS which driver to bind to the
device, and the resources tell the driver where the device is. The
serial driver in drivers/serial/8250_pnp.c is a good example of the
"normal" way PNP drivers work.

it87 doesn't fit the model very well because usually the BIOS
doesn't describe the device explicitly. I think the expectation
is that the OS will get sensor readings some other way, such as
by evaluating ACPI "_TMP" methods, or by using some OEM-specific
ACPI device.

It's very irritating when ACPI is used to take some device that
would otherwise be nicely generic and machine-independent, and
wrap it up into some abstract device that requires an OEM-specific
driver, but I think that's what's happening here. I suspect it
has to do with the fact that the BIOS wants to retain some
control over the device so it can deal with things like thermal
emergencies, even if the OS driver is missing or broken.

> I'm all for integrating the it87 driver into the PNP subsystem if it is
> going to solve problems, but I just don't know how it works. I you do
> some work in this direction, I'll be happy to help with reviews and
> testing.

You don't see how it works because the BIOS writers have deliberately
obscured things (though no doubt they had good reasons).

Bjorn

2007-12-25 21:31:51

[permalink] [raw]

Subject: Re: 2.6.24-rc4 hwmon it87 probe fails

Hi Bjorn,

Le 23/12/2007, "Bjorn Helgaas" <[email protected]> a ?crit:
>On Sunday 23 December 2007 2:28:05 am Jean Delvare wrote:
>> The problem is that the it87 driver is used on a variety of motherboards,
>> some where the hardware monitoring device I/O ports are reserved by the
>> BIOS, some where they aren't. How am I supposed to deal with this?
>
>I assume you mean some boards describe those ports in ACPI, and some
>don't?

Yes, that's what I meant.

>(...) I think the ideal solution would be to figure out how the
>BIOS writers intended the device to be used, and do that. But the
>documentation may be lacking, and it degenerates into an OEM-specific
>mess. Second-best seems like a PNP quirk that pokes enough to determine
>that a Super I/O device is present, and then reserves resources for
>it. Then we avoid the collision even if it87 isn't present.

I'm not sure what you mean here exactly. We have two different cases:
either the ACPI code did declare the PNP device, or it did not. How do
you unify both? And what "collision" are you talking about?

>> PCI is an entirely different beast. With PCI you know the PCI device ID
>> that is associated with the resources, and for a given device, the
>> resources are always declared (if standard BARs are used) or never
>> declared; there's no "may be". So it's much easier to handle the
>> resources properly.
>
>That's the way PNP is supposed to work, too (more about this below).

Not really. PNP devices may or may not be declared. PCI devices are
always declared. This is a fundamental difference. I'm also not sure
about the PNP IDs, see below.

>> That's not really safe, no. They may overlap with other device resources
>> and prevent those drivers from loading.
>
>True. If that happens, I think we definitely need some kind of DMI-
>based quirk to fix the resources. My points are (a) the quirk needs
>to be at the PNP level; it can't be in a driver, and (b) I'm lazy and
>would wait until a problem happens before implementing it, because I
>know so little about PNP and the specific board, and by waiting, there's
>a chance I'll learn enough to avoid a mistake :-)

But we _do_ have a problem to solve! That's what started this thread.
And we need to solve it before 2.6.24 is released. I'd rather (try to)
do the right thing now rather than going in one direction today and
doing something different next month.

I do agree that the quirks should be done at the PNP level and not in the
it87 driver.

>> Another problem is how do I connect this specific I/O port range of the
>> PNP0C02 device with the it87 driver? I am by far no PNP expert but it
>> looks to me like this PNP device covers more than one actual device, and
>> I/O port ranges do not have labels nor identifiers that would let me
>> find the one that corresponds to the IT87xxF device (if it exists at
>> all.)
>
>The general rule is that you have a PNP device with an identifier
>like PNP0500 that means "16550 UART" and some resources underneath
>it. The PNP ID ("PNP0500") tells the OS which driver to bind to the
>device, and the resources tell the driver where the device is. The
>serial driver in drivers/serial/8250_pnp.c is a good example of the
>"normal" way PNP drivers work.
>
>it87 doesn't fit the model very well because usually the BIOS
>doesn't describe the device explicitly. I think the expectation
>is that the OS will get sensor readings some other way, such as
>by evaluating ACPI "_TMP" methods, or by using some OEM-specific
>ACPI device.

If the IT87xxF device is declared, will it always the same PNP ID on
every system? I fear that the answer is no, in which case I see no
simple way to implement the it87 driver as a PNP driver.

>It's very irritating when ACPI is used to take some device that
>would otherwise be nicely generic and machine-independent, and
>wrap it up into some abstract device that requires an OEM-specific
>driver, but I think that's what's happening here. I suspect it
>has to do with the fact that the BIOS wants to retain some
>control over the device so it can deal with things like thermal
>emergencies, even if the OS driver is missing or broken.

Yes, very irritating. Especially when the ACPI way is not only
system-specific, but also very poor in terms of features compared to
what a native driver can do. This is the main reason why we still write
native drivers for hardware monitoring drivers despite the conflicts
with ACPI that arise on some systems.

--
Jean Delvare

2008-01-02 18:31:21

[permalink] [raw]

Subject: Re: 2.6.24-rc4 hwmon it87 probe fails

On Tuesday 25 December 2007 02:31:26 pm Jean Delvare wrote:
> Le 23/12/2007, "Bjorn Helgaas" <[email protected]> a ?crit:
> >On Sunday 23 December 2007 2:28:05 am Jean Delvare wrote:
> >> The problem is that the it87 driver is used on a variety of motherboards,
> >> some where the hardware monitoring device I/O ports are reserved by the
> >> BIOS, some where they aren't. How am I supposed to deal with this?
> >
> >I assume you mean some boards describe those ports in ACPI, and some
> >don't?
>
> Yes, that's what I meant.
>
> >(...) I think the ideal solution would be to figure out how the
> >BIOS writers intended the device to be used, and do that. But the
> >documentation may be lacking, and it degenerates into an OEM-specific
> >mess. Second-best seems like a PNP quirk that pokes enough to determine
> >that a Super I/O device is present, and then reserves resources for
> >it. Then we avoid the collision even if it87 isn't present.
>
> I'm not sure what you mean here exactly. We have two different cases:
> either the ACPI code did declare the PNP device, or it did not. How do
> you unify both?

Even if the BIOS does not declare an IT87xxF as an independent device,
there may be AML that uses the chip internally. For example, the
BIOS could declare a thermal zone with a _TMP method, and the _TMP
method could use the IT87xxF to read the temperature. In that case,
the BIOS would still need to declare the IT87xxF resources so the OS
doesn't place another device on top of it. An easy way to do this
is with a PNP0C02 "motherboard" device, which just says "these
resources are in use, but the OS needn't attach a driver to this
device."

> And what "collision" are you talking about?

I meant the problem where we place two devices at the same address,
causing one to stop working. I wrote "even if it87 isn't present,"
which is ambiguous -- I meant that we need something that works even
when the it87 _driver_ isn't present.

> >> PCI is an entirely different beast. With PCI you know the PCI device ID
> >> that is associated with the resources, and for a given device, the
> >> resources are always declared (if standard BARs are used) or never
> >> declared; there's no "may be". So it's much easier to handle the
> >> resources properly.
> >
> >That's the way PNP is supposed to work, too (more about this below).
>
> Not really. PNP devices may or may not be declared. PCI devices are
> always declared. This is a fundamental difference. I'm also not sure
> about the PNP IDs, see below.

Well, it's true that PNP does not need to describe ALL devices in the
system. However, it should describe all devices to which the OS should
bind a driver. The BIOS writer has the discretion to hide devices from
the OS by omitting them from the ACPI namespace. If he does that, his
assumption is that the OS will ignore the device (but of course, he's
thinking about a Plug-and-Play-compliant OS like Windows).

> >> That's not really safe, no. They may overlap with other device resources
> >> and prevent those drivers from loading.
> >
> >True. If that happens, I think we definitely need some kind of DMI-
> >based quirk to fix the resources. My points are (a) the quirk needs
> >to be at the PNP level; it can't be in a driver, and (b) I'm lazy and
> >would wait until a problem happens before implementing it, because I
> >know so little about PNP and the specific board, and by waiting, there's
> >a chance I'll learn enough to avoid a mistake :-)
>
> But we _do_ have a problem to solve! That's what started this thread.
> And we need to solve it before 2.6.24 is released. I'd rather (try to)
> do the right thing now rather than going in one direction today and
> doing something different next month.

I think my patch to request only the ports actually used by the it87
driver is sufficient to solve the current problem. If we find that
a BIOS neglects to reserve some of the ports decoded by the IT87xxF,
we can add a PNP quirk to handle that. But I'm not aware of any
problems like that yet.

> >it87 doesn't fit the model very well because usually the BIOS
> >doesn't describe the device explicitly. I think the expectation
> >is that the OS will get sensor readings some other way, such as
> >by evaluating ACPI "_TMP" methods, or by using some OEM-specific
> >ACPI device.
>
> If the IT87xxF device is declared, will it always the same PNP ID on
> every system? I fear that the answer is no, in which case I see no
> simple way to implement the it87 driver as a PNP driver.

PNP drivers can claim multiple PNP IDs, e.g., drivers/serial/8250_pnp.c
recognizes many IDs for 16550-compatible UARTs. But I doubt that
any BIOS actually describes an IT87xxF explicitly.

Bjorn

2008-01-12 09:49:33