[parisc-linux] How retrieving HPA/SPA with the device node

[parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 591 bytes --]

	Hi all.
I'm still looking for the HPA/SPA of the SCSI/IDE controler of a disk.
I've found, and able to write the system tree (like ioscan). The starting
point is "root". Now my problem is to find which controler for which disk.
The pointer of the disk (/dev/sd??) is the node. Example:
/dev/sda2 ->> sda2 -> 0x802 -> major:0x8 -> driver "sd" with minor:0x02
Now how can I link 0x802 to hardware path "2/0/1"
So where is the link between both ????? Do you have any idea ?????

	Thanks.

-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Matthew Wilcox]

On Tue, Apr 02, 2002 at 04:08:16PM +0200, Bruno Vidal wrote:
> 	Hi all.
> I'm still looking for the HPA/SPA of the SCSI/IDE controler of a disk.
> I've found, and able to write the system tree (like ioscan). The starting
> point is "root". Now my problem is to find which controler for which disk.
> The pointer of the disk (/dev/sd??) is the node. Example:
> /dev/sda2 ->> sda2 -> 0x802 -> major:0x8 -> driver "sd" with minor:0x02
> Now how can I link 0x802 to hardware path "2/0/1"
> So where is the link between both ????? Do you have any idea ?????

Hmm.  If you can get to the pci_dev or parisc_device for a scsi disc,
we have functions for returning the hwpath:

print_pci_hwpath
get_pci_node_path
print_pa_hwpath
get_node_path

I'm not sure how to get from `sda2' to the pci_dev of the controlling device.

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 845 bytes --]

It is my pb, I don't know how to retrieve the pci_dev or parisc_device
from the `sda2' (the device file). I'm pretty sure that the solution
can comes from the `sd' drivers (btw which is the name for the
ide driver: ide ???). Because when using it, you have to give the device
nodes, so it should translate it somewhere to a pci_dev or parisc_device.

Thanks to people who know how to handle this.

Matthew Wilcox wrote:
> 
> I'm not sure how to get from `sda2' to the pci_dev of the controlling device.
> 
> --
> Revolutions do not require corporate support.
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Grant Grundler]

Bruno Vidal wrote:
> The pointer of the disk (/dev/sd??) is the node. Example:
> /dev/sda2 ->> sda2 -> 0x802 -> major:0x8 -> driver "sd" with minor:0x02
> Now how can I link 0x802 to hardware path "2/0/1"
> So where is the link between both ????? Do you have any idea ?????

Well, there is a link. It's just not as obvious as the iotree.

drivers/scsi/sd.c:sd_open() finds the device:

	target = DEVICE_NR(inode->i_rdev);
	if (target >= sd_template.dev_max || !rscsi_disks[target].device)
		return -ENXIO;  /* No such device */

And sd.h has the definition for Scsi_Disk:
typedef struct scsi_disk {
	unsigned capacity;      /* size in blocks */
	Scsi_Device *device; 
	unsigned char ready;    /* flag ready for FLOPTICAL */
	unsigned char write_prot;       /* flag write_protect for rmvable dev */
	unsigned char sector_bit_size;  /* sector_size = 2 to the  bit size power */
	unsigned char sector_bit_shift;         /* power of 2 sectors per FS block */
	unsigned has_part_table:1;      /* has partition table */
} Scsi_Disk; 

Scsi_Device is *private*. It contains the ptr to Scsi_Host (defined
in hosts.h), an instance of a SCSI controller driver. And finally,
Scsi_Host has a pointer to pci_dev.

Since several of the data structures are intended to be private,
I suggest adding a function to sd.c (or scsi.c) that finds and returns
the pci_dev pointer:

struct pci_dev *sd_get_pci_dev(struct inode *inode)
{
	int target = DEVICE_NR(inode->i_rdev);
	Scsi_Device scsidev = rscsi_disks[target].device;

	if (target >= sd_template.dev_max || !scsidev)
		return NULL;  /* No such device */

	return (scsi_dev->host->pci_dev);
}


It would be nice if SCSI were part of the IO tree but I can understand
reasons for not doing it. In particular, SCSI over FC doesn't lend
itself to an IO Tree arrangement since it's really a "net". Put a SAN,
additional FC controllers from the same host, and a few thousand
LUNs out there and the IO tree can become an ugly mess to.

hth,
grant

ps. I had no clue how this worked until I tracked this down starting
    at sd_open().

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Grant Grundler]

Grant Grundler wrote:
> Since several of the data structures are intended to be private,
> I suggest adding a function to sd.c (or scsi.c) that finds and returns
> the pci_dev pointer:

Well, looks like we *almost* have what we want in scsi_ioctl.c.
scsi_ioctl_get_pci() does about the same thing.
Maybe the right answer is to write another ioctl call that
does what you want. I don't know.

grant

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 931 bytes --]

	Thanks a lot.
It solved my pb. Now I think I'll add a function in scsi.c.
Because I need the pointer to pci_dev. And I don't think it
can be usefull at user level. 

	Many thanks for your answers.

Grant Grundler wrote:
> 
> Grant Grundler wrote:
> > Since several of the data structures are intended to be private,
> > I suggest adding a function to sd.c (or scsi.c) that finds and returns
> > the pci_dev pointer:
> 
> Well, looks like we *almost* have what we want in scsi_ioctl.c.
> scsi_ioctl_get_pci() does about the same thing.
> Maybe the right answer is to write another ioctl call that
> does what you want. I don't know.
> 
> grant
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux


-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 3108 bytes --]

	Hi
I've finally found enought time to continu my long way to the HPA.
So now I've build a function in sd.c in order to retrieve the scsi_host.
But as it has been said, the pointer pci_dev in the structure scsi_host
is not always pointing to a pci_dev. For my poor 712/100, it point
to NULL :-( (it is normal, my scsi host is not a PCI one). So I'm still
stuck at this point, now I'm able to retrieve the scsi_host and scsi_dev
structures, but I'm still unable to go the parisc_device. Any idea ?

	Thanks.

Grant Grundler wrote:
> 
> Bruno Vidal wrote:
> > The pointer of the disk (/dev/sd??) is the node. Example:
> > /dev/sda2 ->> sda2 -> 0x802 -> major:0x8 -> driver "sd" with minor:0x02
> > Now how can I link 0x802 to hardware path "2/0/1"
> > So where is the link between both ????? Do you have any idea ?????
> 
> Well, there is a link. It's just not as obvious as the iotree.
> 
> drivers/scsi/sd.c:sd_open() finds the device:
> 
>         target = DEVICE_NR(inode->i_rdev);
>         if (target >= sd_template.dev_max || !rscsi_disks[target].device)
>                 return -ENXIO;  /* No such device */
> 
> And sd.h has the definition for Scsi_Disk:
> typedef struct scsi_disk {
>         unsigned capacity;      /* size in blocks */
>         Scsi_Device *device;
>         unsigned char ready;    /* flag ready for FLOPTICAL */
>         unsigned char write_prot;       /* flag write_protect for rmvable dev */
>         unsigned char sector_bit_size;  /* sector_size = 2 to the  bit size power */
>         unsigned char sector_bit_shift;         /* power of 2 sectors per FS block */
>         unsigned has_part_table:1;      /* has partition table */
> } Scsi_Disk;
> 
> Scsi_Device is *private*. It contains the ptr to Scsi_Host (defined
> in hosts.h), an instance of a SCSI controller driver. And finally,
> Scsi_Host has a pointer to pci_dev.
> 
> Since several of the data structures are intended to be private,
> I suggest adding a function to sd.c (or scsi.c) that finds and returns
> the pci_dev pointer:
> 
> struct pci_dev *sd_get_pci_dev(struct inode *inode)
> {
>         int target = DEVICE_NR(inode->i_rdev);
>         Scsi_Device scsidev = rscsi_disks[target].device;
> 
>         if (target >= sd_template.dev_max || !scsidev)
>                 return NULL;  /* No such device */
> 
>         return (scsi_dev->host->pci_dev);
> }
> 
> It would be nice if SCSI were part of the IO tree but I can understand
> reasons for not doing it. In particular, SCSI over FC doesn't lend
> itself to an IO Tree arrangement since it's really a "net". Put a SAN,
> additional FC controllers from the same host, and a few thousand
> LUNs out there and the IO tree can become an ugly mess to.
> 
> hth,
> grant
> 
> ps. I had no clue how this worked until I tracked this down starting
>     at sd_open().
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux

-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Grant Grundler]

Bruno Vidal wrote:
> But as it has been said, the pointer pci_dev in the structure scsi_host
> is not always pointing to a pci_dev. For my poor 712/100, it point
> to NULL :-( (it is normal, my scsi host is not a PCI one).

The best idea we came up with was to require the dump partition be
on boot device and pull the HPA from page zero. Matthew Wilcox was
mentioning that this problem will be solved in 2.5 kernel
since it will get an IO tree similar to HPUX's.

TBH, I'm personally not interested in chasing down solutions
for boxes that don't have PCI.

hth,
grant

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 542 bytes --]

	Hi
Can someone can explain what is the difference between the
"hpa" of a scsi card (pci or not) and the "base" in the structure Scsi_Host ?
....
    /* These parameters should be set by the detect routine */
    unsigned long base;
    unsigned long io_port;
    unsigned char n_io_port;
    unsigned char dma_channel;
    unsigned int  irq; 
....

I'm pretty sure that it is the same, can someone say if it is wrong or not ?

	Thanks.

-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Matthew Wilcox]

On Mon, Jun 03, 2002 at 01:04:10PM +0200, Bruno Vidal wrote:
> 	Hi
> Can someone can explain what is the difference between the
> "hpa" of a scsi card (pci or not) and the "base" in the structure Scsi_Host ?
> ....
>     /* These parameters should be set by the detect routine */
>     unsigned long base;
>     unsigned long io_port;
>     unsigned char n_io_port;
>     unsigned char dma_channel;
>     unsigned int  irq; 
> ....
> 
> I'm pretty sure that it is the same, can someone say if it is wrong or not ?

I'm no expert in the SCSI layer, but it could be that base is the base
of the _chip_ and hpa is the base of the _device_.  For example, the
hpa of the serial device is 0xffd05000 but the chip itself is at 0xffd05800.
According to drivers/scsi/lasi700.c:

        unsigned long base = dev->hpa + LASI_SCSI_CORE_OFFSET;
[...]
        hostdata->base = base;

so base != hpa.  Sorry...

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Grant Grundler]

Bruno Vidal wrote:
> Can someone can explain what is the difference between the
> "hpa" of a scsi card (pci or not) and the "base" in the structure Scsi_Host ?

hpa and base are "handles" for different views of HW.
"hpa" doesn't have to be a physical address (though it normally is).
On V-class, the hpa is bogus and just a reference to particular IO device.
base could be "adjusted" (like the LASI example) or "F-extended"
for 64-bit or whatever the platform code thinks the driver needs
to reference the device properly.

But in practice, for the platforms we support to date, I think you
can compare them (or compare most of the bits) to determine
if it's the same device.

grant

Re: [parisc-linux] How retrieving HPA/SPA with the device node

[Bruno Vidal]

[-- Attachment #1: Type: text/plain, Size: 1581 bytes --]

	Hi
Thanks for the answer, it's clear now. The fact is I'm able 
to retrieve the hpa for a PCI device "only". But for the lazy 
example, I can retrieve the base, but not the hpa. So it would 
be really nice if I can write something usable also on 
712/710/etc.... and not only for PCI system. So I think I'll
forget it at this time, and work only for PCI subsystem (I can still
work on a function that test the kind of scsi card, and do the
reverse job to calculate the hpa from the base, but I'm not sure that
it is the best solution).

	Thanks a lot.

Grant Grundler wrote:
> 
> Bruno Vidal wrote:
> > Can someone can explain what is the difference between the
> > "hpa" of a scsi card (pci or not) and the "base" in the structure Scsi_Host ?
> 
> hpa and base are "handles" for different views of HW.
> "hpa" doesn't have to be a physical address (though it normally is).
> On V-class, the hpa is bogus and just a reference to particular IO device.
> base could be "adjusted" (like the LASI example) or "F-extended"
> for 64-bit or whatever the platform code thinks the driver needs
> to reference the device properly.
> 
> But in practice, for the platforms we support to date, I think you
> can compare them (or compare most of the bits) to determine
> if it's the same device.
> 
> grant
> 
> _______________________________________________
> parisc-linux mailing list
> parisc-linux@lists.parisc-linux.org
> http://lists.parisc-linux.org/mailman/listinfo/parisc-linux


-- 
	Vidal Bruno, (770-4271)
        SSD-HA Team, HP-UX & LINUX Support
	bruno_vidal@admin.france.hp.com

[-- Attachment #2: Card for Bruno Vidal --]
[-- Type: text/x-vcard, Size: 386 bytes --]

begin:vcard 
n:Bruno;VIDAL
tel;fax:01-69-82-60-14
tel;work:01-69-29-42-71
x-mozilla-html:TRUE
url:www.france.hp.com
org:Solution Center
version:2.1
email;internet:bruno_vidal@hp.com
title:Expert Logiciel Environnement Haute Disponibilité
adr;quoted-printable:;;HP France=0D=0A;Z.A de Courtaboeuf;1 Av. du Canada;91947 Les Ulis cedex;France
x-mozilla-cpt:;26208
fn:VIDAL Bruno
end:vcard