hotplug works, but newly plug not

hotplug works, but newly plug not

[Ming Zhang]

Hi All

I use the libata patch from
http://home-tj.org/wiki/index.php/Libata-tj-stable with my supermicro MB
( ICH7R chip). The AHCI works as desired in most of the time.

If I have disk(s) plugged in during boot. Then after boot, I kept unplug
and plug many times and it works fine.

but i noticed that if I have the disk(s) unplugged during boot, then
after boot, I plug the disk and nothing shows. if I try to do rescan by
run the rescan lun script, i saw these from dmesg.

Anything clue ? Thanks!

Ming

---------------------------------------
ata4: soft resetting port
ata4: SATA link down (SStatus 0 SControl 0)
ata4: EH complete
ata3: soft resetting port
ata3: SATA link down (SStatus 0 SControl 0)
ata3: EH complete
ata4: soft resetting port
ata4: SATA link down (SStatus 0 SControl 0)
ata4: EH complete
ata3: soft resetting port
ata3: SATA link down (SStatus 0 SControl 0)
ata3: EH complete
ata4: soft resetting port
ata4: SATA link down (SStatus 0 SControl 0)
ata4: EH complete

Re: hotplug works, but newly plug not

[Tejun Heo]

[-- Attachment #1: Type: text/plain, Size: 1379 bytes --]

Ming Zhang wrote:
> Hi All
> 
> I use the libata patch from
> http://home-tj.org/wiki/index.php/Libata-tj-stable with my supermicro MB
> ( ICH7R chip). The AHCI works as desired in most of the time.
> 
> If I have disk(s) plugged in during boot. Then after boot, I kept unplug
> and plug many times and it works fine.
> 
> but i noticed that if I have the disk(s) unplugged during boot, then
> after boot, I plug the disk and nothing shows. if I try to do rescan by
> run the rescan lun script, i saw these from dmesg.
> 
> Anything clue ? Thanks!
> 
> Ming
> 
> ---------------------------------------
> ata4: soft resetting port
> ata4: SATA link down (SStatus 0 SControl 0)
> ata4: EH complete
> ata3: soft resetting port
> ata3: SATA link down (SStatus 0 SControl 0)
> ata3: EH complete
> ata4: soft resetting port
> ata4: SATA link down (SStatus 0 SControl 0)
> ata4: EH complete
> ata3: soft resetting port
> ata3: SATA link down (SStatus 0 SControl 0)
> ata3: EH complete
> ata4: soft resetting port
> ata4: SATA link down (SStatus 0 SControl 0)
> ata4: EH complete

Hello,

That never happened to me on AHCI mode but I experienced similar thing 
with piix mode.  It might be that your BIOS doesn't turn on PCS enable 
bits for empty ports and thus the empty ports cannot be used after boot. 
Can you try the attached patch?  (only compile-tested, be cautious)

-- 
tejun

[-- Attachment #2: patch --]
[-- Type: text/plain, Size: 568 bytes --]

diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c
index 45fd71d..83751dd 100644
--- a/drivers/scsi/ahci.c
+++ b/drivers/scsi/ahci.c
@@ -1306,6 +1306,15 @@ static int ahci_init_one (struct pci_dev
 		have_msi = 0;
 	}
 
+	{
+		u8 orig_pcs, pcs;
+		pci_read_config_byte(pdev, 0x92, &orig_pcs);
+		pci_write_config_byte(pdev, 0x92, 0xf);
+		pci_read_config_byte(pdev, 0x92, &pcs);
+		dev_printk(KERN_INFO, &pdev->dev, "pcs 0x%x->0x%x\n",
+			   orig_pcs, pcs);
+	}
+
 	probe_ent = kmalloc(sizeof(*probe_ent), GFP_KERNEL);
 	if (probe_ent == NULL) {
 		rc = -ENOMEM;

Re: hotplug works, but newly plug not

[Ming Zhang]

On Wed, 2006-05-24 at 13:51 -0700, Tejun Heo wrote:
> Ming Zhang wrote:
> > Hi All
> > 
> > I use the libata patch from
> > http://home-tj.org/wiki/index.php/Libata-tj-stable with my supermicro MB
> > ( ICH7R chip). The AHCI works as desired in most of the time.
> > 
> > If I have disk(s) plugged in during boot. Then after boot, I kept unplug
> > and plug many times and it works fine.
> > 
> > but i noticed that if I have the disk(s) unplugged during boot, then
> > after boot, I plug the disk and nothing shows. if I try to do rescan by
> > run the rescan lun script, i saw these from dmesg.
> > 
> > Anything clue ? Thanks!
> > 
> > Ming
> > 
> > ---------------------------------------
> > ata4: soft resetting port
> > ata4: SATA link down (SStatus 0 SControl 0)
> > ata4: EH complete
> > ata3: soft resetting port
> > ata3: SATA link down (SStatus 0 SControl 0)
> > ata3: EH complete
> > ata4: soft resetting port
> > ata4: SATA link down (SStatus 0 SControl 0)
> > ata4: EH complete
> > ata3: soft resetting port
> > ata3: SATA link down (SStatus 0 SControl 0)
> > ata3: EH complete
> > ata4: soft resetting port
> > ata4: SATA link down (SStatus 0 SControl 0)
> > ata4: EH complete
> 
> Hello,
> 
> That never happened to me on AHCI mode but I experienced similar thing 
> with piix mode.  It might be that your BIOS doesn't turn on PCS enable 

the BIOS has a option to switch between AHCI and PIIX. So I guess the
reason is as you said.

also this is kernel boot log related to this. could u point out the
place for this pcs in AHCI spec? i just printed AHCI and can not find
related information about this PCS. the code operate at offset 0x92 but
i can not find anything about that.


libata version 1.30 loaded.
ahci 0000:00:1f.2: version 1.3
ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
ahci 0000:00:1f.2: pcs 0x11->0x1f
PCI: Setting latency timer of device 0000:00:1f.2 to 64
ahci 0000:00:1f.2: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0x1 impl SATA
mode
ahci 0000:00:1f.2: flags: 64bit ncq pm led clo pio slum part
ata1: SATA max UDMA/133 cmd 0xF881C500 ctl 0x0 bmdma 0x0 irq 19
ata2: SATA max UDMA/133 cmd 0xF881C580 ctl 0x0 bmdma 0x0 irq 19
ata3: SATA max UDMA/133 cmd 0xF881C600 ctl 0x0 bmdma 0x0 irq 19
ata4: SATA max UDMA/133 cmd 0xF881C680 ctl 0x0 bmdma 0x0 irq 19
scsi0 : ahci
ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata1.00: cfg 49:2f00 82:74eb 83:7feb 84:4123 85:74e9 86:3c03 87:4123
88:007f
ata1.00: ATA-7, max UDMA/133, 781422768 sectors: LBA48
ata1.00: configured for UDMA/133
scsi1 : ahci
ata2: SATA link down (SStatus 0 SControl 0)
scsi2 : ahci
ata3: SATA link down (SStatus 0 SControl 0)
scsi3 : ahci
ata4: SATA link down (SStatus 0 SControl 0)
  Vendor: ATA       Model: HDS724040KLSA80   Rev: KFAO
  Type:   Direct-Access                      ANSI SCSI revision: 05
SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
sda: Write Protect is off
sda: Mode Sense: 00 3a 00 00
SCSI device sda: drive cache: write back
SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
sda: Write Protect is off
sda: Mode Sense: 00 3a 00 00
SCSI device sda: drive cache: write back
 sda:
sd 0:0:0:0: Attached scsi disk sda
sd 0:0:0:0: Attached scsi generic sg0 type 0




> bits for empty ports and thus the empty ports cannot be used after boot. 
> Can you try the attached patch?  (only compile-tested, be cautious)
> 
> plain text document attachment (patch)
> diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c
> index 45fd71d..83751dd 100644
> --- a/drivers/scsi/ahci.c
> +++ b/drivers/scsi/ahci.c
> @@ -1306,6 +1306,15 @@ static int ahci_init_one (struct pci_dev
>  		have_msi = 0;
>  	}
>  
> +	{
> +		u8 orig_pcs, pcs;
> +		pci_read_config_byte(pdev, 0x92, &orig_pcs);
> +		pci_write_config_byte(pdev, 0x92, 0xf);
> +		pci_read_config_byte(pdev, 0x92, &pcs);
> +		dev_printk(KERN_INFO, &pdev->dev, "pcs 0x%x->0x%x\n",
> +			   orig_pcs, pcs);
> +	}
> +
>  	probe_ent = kmalloc(sizeof(*probe_ent), GFP_KERNEL);
>  	if (probe_ent == NULL) {
>  		rc = -ENOMEM;

Re: hotplug works, but newly plug not

[Ming Zhang]

my fault. forgot to say that the patch does not change the behavior.
nothing happen after plug a disk to empty slot.

ming


On Wed, 2006-05-24 at 17:09 -0400, Ming Zhang wrote:
> On Wed, 2006-05-24 at 13:51 -0700, Tejun Heo wrote:
> > Ming Zhang wrote:
> > > Hi All
> > > 
> > > I use the libata patch from
> > > http://home-tj.org/wiki/index.php/Libata-tj-stable with my supermicro MB
> > > ( ICH7R chip). The AHCI works as desired in most of the time.
> > > 
> > > If I have disk(s) plugged in during boot. Then after boot, I kept unplug
> > > and plug many times and it works fine.
> > > 
> > > but i noticed that if I have the disk(s) unplugged during boot, then
> > > after boot, I plug the disk and nothing shows. if I try to do rescan by
> > > run the rescan lun script, i saw these from dmesg.
> > > 
> > > Anything clue ? Thanks!
> > > 
> > > Ming
> > > 
> > > ---------------------------------------
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > > ata3: soft resetting port
> > > ata3: SATA link down (SStatus 0 SControl 0)
> > > ata3: EH complete
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > > ata3: soft resetting port
> > > ata3: SATA link down (SStatus 0 SControl 0)
> > > ata3: EH complete
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > 
> > Hello,
> > 
> > That never happened to me on AHCI mode but I experienced similar thing 
> > with piix mode.  It might be that your BIOS doesn't turn on PCS enable 
> 
> the BIOS has a option to switch between AHCI and PIIX. So I guess the
> reason is as you said.
> 
> also this is kernel boot log related to this. could u point out the
> place for this pcs in AHCI spec? i just printed AHCI and can not find
> related information about this PCS. the code operate at offset 0x92 but
> i can not find anything about that.
> 
> 
> libata version 1.30 loaded.
> ahci 0000:00:1f.2: version 1.3
> ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
> ahci 0000:00:1f.2: pcs 0x11->0x1f
> PCI: Setting latency timer of device 0000:00:1f.2 to 64
> ahci 0000:00:1f.2: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0x1 impl SATA
> mode
> ahci 0000:00:1f.2: flags: 64bit ncq pm led clo pio slum part
> ata1: SATA max UDMA/133 cmd 0xF881C500 ctl 0x0 bmdma 0x0 irq 19
> ata2: SATA max UDMA/133 cmd 0xF881C580 ctl 0x0 bmdma 0x0 irq 19
> ata3: SATA max UDMA/133 cmd 0xF881C600 ctl 0x0 bmdma 0x0 irq 19
> ata4: SATA max UDMA/133 cmd 0xF881C680 ctl 0x0 bmdma 0x0 irq 19
> scsi0 : ahci
> ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
> ata1.00: cfg 49:2f00 82:74eb 83:7feb 84:4123 85:74e9 86:3c03 87:4123
> 88:007f
> ata1.00: ATA-7, max UDMA/133, 781422768 sectors: LBA48
> ata1.00: configured for UDMA/133
> scsi1 : ahci
> ata2: SATA link down (SStatus 0 SControl 0)
> scsi2 : ahci
> ata3: SATA link down (SStatus 0 SControl 0)
> scsi3 : ahci
> ata4: SATA link down (SStatus 0 SControl 0)
>   Vendor: ATA       Model: HDS724040KLSA80   Rev: KFAO
>   Type:   Direct-Access                      ANSI SCSI revision: 05
> SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
> sda: Write Protect is off
> sda: Mode Sense: 00 3a 00 00
> SCSI device sda: drive cache: write back
> SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
> sda: Write Protect is off
> sda: Mode Sense: 00 3a 00 00
> SCSI device sda: drive cache: write back
>  sda:
> sd 0:0:0:0: Attached scsi disk sda
> sd 0:0:0:0: Attached scsi generic sg0 type 0
> 
> 
> 
> 
> > bits for empty ports and thus the empty ports cannot be used after boot. 
> > Can you try the attached patch?  (only compile-tested, be cautious)
> > 
> > plain text document attachment (patch)
> > diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c
> > index 45fd71d..83751dd 100644
> > --- a/drivers/scsi/ahci.c
> > +++ b/drivers/scsi/ahci.c
> > @@ -1306,6 +1306,15 @@ static int ahci_init_one (struct pci_dev
> >  		have_msi = 0;
> >  	}
> >  
> > +	{
> > +		u8 orig_pcs, pcs;
> > +		pci_read_config_byte(pdev, 0x92, &orig_pcs);
> > +		pci_write_config_byte(pdev, 0x92, 0xf);
> > +		pci_read_config_byte(pdev, 0x92, &pcs);
> > +		dev_printk(KERN_INFO, &pdev->dev, "pcs 0x%x->0x%x\n",
> > +			   orig_pcs, pcs);
> > +	}
> > +
> >  	probe_ent = kmalloc(sizeof(*probe_ent), GFP_KERNEL);
> >  	if (probe_ent == NULL) {
> >  		rc = -ENOMEM;
> 

Re: hotplug works, but newly plug not

[Tejun Heo]

Ming Zhang wrote:
> my fault. forgot to say that the patch does not change the behavior.
> nothing happen after plug a disk to empty slot.
> 

Can you change 0xf to 0xff of the following line in the patch and see if 
anything changes?

pci_write_config_byte(pdev, 0x92, 0xf);
                                   ^^^^-> 0xff


-- 
tejun

Re: hotplug works, but newly plug not

[Ming Zhang]

nothing changed. no plug detected.

ahci 0000:00:1f.2: version 1.3
ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
ahci 0000:00:1f.2: pcs 0xf->0xf
PCI: Setting latency timer of device 0000:00:1f.2 to 64
ahci 0000:00:1f.2: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0x0 impl SATA
mode
ahci 0000:00:1f.2: flags: 64bit ncq pm led clo pio slum part
ata5: SATA max UDMA/133 cmd 0xF8820500 ctl 0x0 bmdma 0x0 irq 19
ata6: SATA max UDMA/133 cmd 0xF8820580 ctl 0x0 bmdma 0x0 irq 19
ata7: SATA max UDMA/133 cmd 0xF8820600 ctl 0x0 bmdma 0x0 irq 19
ata8: SATA max UDMA/133 cmd 0xF8820680 ctl 0x0 bmdma 0x0 irq 19
scsi4 : ahci
ata5: SATA link down (SStatus 0 SControl 0)
scsi5 : ahci
ata6: SATA link down (SStatus 0 SControl 0)
scsi6 : ahci
ata7: SATA link down (SStatus 0 SControl 0)
scsi7 : ahci
ata8: SATA link down (SStatus 0 SControl 0)


On Wed, 2006-05-24 at 14:25 -0700, Tejun Heo wrote:
> Ming Zhang wrote:
> > my fault. forgot to say that the patch does not change the behavior.
> > nothing happen after plug a disk to empty slot.
> > 
> 
> Can you change 0xf to 0xff of the following line in the patch and see if 
> anything changes?
> 
> pci_write_config_byte(pdev, 0x92, 0xf);
>                                    ^^^^-> 0xff
> 
> 

Re: hotplug works, but newly plug not

[Ming Zhang]

Hi Tejun

I read the whole AHCI but I still could not find out what is the magic
number 0x92. Is this in some Intel doc?

Ming


On Wed, 2006-05-24 at 17:09 -0400, Ming Zhang wrote:
> On Wed, 2006-05-24 at 13:51 -0700, Tejun Heo wrote:
> > Ming Zhang wrote:
> > > Hi All
> > > 
> > > I use the libata patch from
> > > http://home-tj.org/wiki/index.php/Libata-tj-stable with my supermicro MB
> > > ( ICH7R chip). The AHCI works as desired in most of the time.
> > > 
> > > If I have disk(s) plugged in during boot. Then after boot, I kept unplug
> > > and plug many times and it works fine.
> > > 
> > > but i noticed that if I have the disk(s) unplugged during boot, then
> > > after boot, I plug the disk and nothing shows. if I try to do rescan by
> > > run the rescan lun script, i saw these from dmesg.
> > > 
> > > Anything clue ? Thanks!
> > > 
> > > Ming
> > > 
> > > ---------------------------------------
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > > ata3: soft resetting port
> > > ata3: SATA link down (SStatus 0 SControl 0)
> > > ata3: EH complete
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > > ata3: soft resetting port
> > > ata3: SATA link down (SStatus 0 SControl 0)
> > > ata3: EH complete
> > > ata4: soft resetting port
> > > ata4: SATA link down (SStatus 0 SControl 0)
> > > ata4: EH complete
> > 
> > Hello,
> > 
> > That never happened to me on AHCI mode but I experienced similar thing 
> > with piix mode.  It might be that your BIOS doesn't turn on PCS enable 
> 
> the BIOS has a option to switch between AHCI and PIIX. So I guess the
> reason is as you said.
> 
> also this is kernel boot log related to this. could u point out the
> place for this pcs in AHCI spec? i just printed AHCI and can not find
> related information about this PCS. the code operate at offset 0x92 but
> i can not find anything about that.
> 
> 
> libata version 1.30 loaded.
> ahci 0000:00:1f.2: version 1.3
> ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
> ahci 0000:00:1f.2: pcs 0x11->0x1f
> PCI: Setting latency timer of device 0000:00:1f.2 to 64
> ahci 0000:00:1f.2: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0x1 impl SATA
> mode
> ahci 0000:00:1f.2: flags: 64bit ncq pm led clo pio slum part
> ata1: SATA max UDMA/133 cmd 0xF881C500 ctl 0x0 bmdma 0x0 irq 19
> ata2: SATA max UDMA/133 cmd 0xF881C580 ctl 0x0 bmdma 0x0 irq 19
> ata3: SATA max UDMA/133 cmd 0xF881C600 ctl 0x0 bmdma 0x0 irq 19
> ata4: SATA max UDMA/133 cmd 0xF881C680 ctl 0x0 bmdma 0x0 irq 19
> scsi0 : ahci
> ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
> ata1.00: cfg 49:2f00 82:74eb 83:7feb 84:4123 85:74e9 86:3c03 87:4123
> 88:007f
> ata1.00: ATA-7, max UDMA/133, 781422768 sectors: LBA48
> ata1.00: configured for UDMA/133
> scsi1 : ahci
> ata2: SATA link down (SStatus 0 SControl 0)
> scsi2 : ahci
> ata3: SATA link down (SStatus 0 SControl 0)
> scsi3 : ahci
> ata4: SATA link down (SStatus 0 SControl 0)
>   Vendor: ATA       Model: HDS724040KLSA80   Rev: KFAO
>   Type:   Direct-Access                      ANSI SCSI revision: 05
> SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
> sda: Write Protect is off
> sda: Mode Sense: 00 3a 00 00
> SCSI device sda: drive cache: write back
> SCSI device sda: 781422768 512-byte hdwr sectors (400088 MB)
> sda: Write Protect is off
> sda: Mode Sense: 00 3a 00 00
> SCSI device sda: drive cache: write back
>  sda:
> sd 0:0:0:0: Attached scsi disk sda
> sd 0:0:0:0: Attached scsi generic sg0 type 0
> 
> 
> 
> 
> > bits for empty ports and thus the empty ports cannot be used after boot. 
> > Can you try the attached patch?  (only compile-tested, be cautious)
> > 
> > plain text document attachment (patch)
> > diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c
> > index 45fd71d..83751dd 100644
> > --- a/drivers/scsi/ahci.c
> > +++ b/drivers/scsi/ahci.c
> > @@ -1306,6 +1306,15 @@ static int ahci_init_one (struct pci_dev
> >  		have_msi = 0;
> >  	}
> >  
> > +	{
> > +		u8 orig_pcs, pcs;
> > +		pci_read_config_byte(pdev, 0x92, &orig_pcs);
> > +		pci_write_config_byte(pdev, 0x92, 0xf);
> > +		pci_read_config_byte(pdev, 0x92, &pcs);
> > +		dev_printk(KERN_INFO, &pdev->dev, "pcs 0x%x->0x%x\n",
> > +			   orig_pcs, pcs);
> > +	}
> > +
> >  	probe_ent = kmalloc(sizeof(*probe_ent), GFP_KERNEL);
> >  	if (probe_ent == NULL) {
> >  		rc = -ENOMEM;
> 

Re: hotplug works, but newly plug not

[Jeff Garzik]

Ming Zhang wrote:
> I read the whole AHCI but I still could not find out what is the magic
> number 0x92. Is this in some Intel doc?

AHCI spec only documents generic PCI config registers.  Each vendor has 
their own vendor-specific registers.  Intel's are documented publicly on 
the developer website.

	Jeff

Re: hotplug works, but newly plug not

[Ming Zhang]

On Thu, 2006-05-25 at 11:14 -0400, Jeff Garzik wrote:
> Ming Zhang wrote:
> > I read the whole AHCI but I still could not find out what is the magic
> > number 0x92. Is this in some Intel doc?
> 
> AHCI spec only documents generic PCI config registers.  Each vendor has 
> their own vendor-specific registers.  Intel's are documented publicly on 
> the developer website.
> 
> 	Jeff
> 
> 

Thanks! I will grab one now.

Ming

Re: hotplug works, but newly plug not

[Ming Zhang]

I added these 2 lines in ahci_host_init()

                writel(1 << i, mmio + HOST_IRQ_STAT);

                /* set irq mask (enables interrupts) */
                writel(DEF_PORT_IRQ, port_mmio + PORT_IRQ_MASK);

~~~~~~code i added
                tmp = readl(port_mmio + PORT_IRQ_MASK);
                VPRINTK("PORT_IRQ_MASK 0x%x\n", tmp);
~~~~~~~~~~~~~

And I got this dmesg with VPRINTK enabled in libata.h. Does this mean
that IRQ is not enabled on this port at all?

ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
ahci 0000:00:1f.2: pcs 0x8f->0x8f
ahci_host_init: cap 0xc720ff03  port_map 0x0  n_ports 4
ahci_host_init: mmio f8820400  port_mmio f8820500
ahci_setup_port: ENTER, base==0xf8820400, port_idx 0
ahci_setup_port: base now==0xf8820500
ahci_setup_port: EXIT
ahci_host_init: PORT_CMD 0x0
ahci_host_init: PORT_SCR_ERR 0x0
ahci_host_init: PORT_IRQ_STAT 0x0
ahci_host_init: PORT_IRQ_MASK 0x0
ahci_host_init: mmio f8820400  port_mmio f8820580
ahci_setup_port: ENTER, base==0xf8820400, port_idx 1
ahci_setup_port: base now==0xf8820580
ahci_setup_port: EXIT
ahci_host_init: PORT_CMD 0x0
ahci_host_init: PORT_SCR_ERR 0x0
ahci_host_init: PORT_IRQ_STAT 0x0
ahci_host_init: PORT_IRQ_MASK 0x0
...


On Thu, 2006-05-25 at 11:14 -0400, Jeff Garzik wrote:
> Ming Zhang wrote:
> > I read the whole AHCI but I still could not find out what is the magic
> > number 0x92. Is this in some Intel doc?
> 
> AHCI spec only documents generic PCI config registers.  Each vendor has 
> their own vendor-specific registers.  Intel's are documented publicly on 
> the developer website.
> 
> 	Jeff
> 
> 
> 

Re: hotplug works, but newly plug not

[Ming Zhang]

also I downloaded Intel ICH7 Data sheet which contains all pci
configuration registers defines. thx for pointing out this.

also i saw that ahci_host_init() will set low 4 bits to 0xF anyway for
intel chip, and high 4 bits are RO. so the patch set 0xf or 0xFF make no
sense here.

so what can i do now? thx!


Ming



On Thu, 2006-05-25 at 11:14 -0400, Jeff Garzik wrote:
> Ming Zhang wrote:
> > I read the whole AHCI but I still could not find out what is the magic
> > number 0x92. Is this in some Intel doc?
> 
> AHCI spec only documents generic PCI config registers.  Each vendor has 
> their own vendor-specific registers.  Intel's are documented publicly on 
> the developer website.
> 
> 	Jeff
> 
> 
> 

Re: hotplug works, but newly plug not

[Ming Zhang]

This time I boot with 2 disks in slot. So the IRQ_MASK is set
successfully for 2 ports and left as 0 for another 2 ports.

Ming

========================
May 25 13:00:53 localhost kernel: ahci_setup_port: base now==0xf881c500
May 25 13:00:53 localhost kernel: ahci_setup_port: EXIT
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_CMD 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_SCR_ERR 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_STAT 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_MASK 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: mmio f881c400
port_mmio f881c580
May 25 13:00:53 localhost kernel: ahci_setup_port: ENTER,
base==0xf881c400, port_idx 1
May 25 13:00:53 localhost kernel: ahci_setup_port: base now==0xf881c580
May 25 13:00:53 localhost kernel: ahci_setup_port: EXIT
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_CMD 0x6
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_SCR_ERR 0x4050000
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_STAT 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_MASK
0x7840007f
May 25 13:00:53 localhost kernel: ahci_host_init: mmio f881c400
port_mmio f881c600
May 25 13:00:53 localhost kernel: ahci_setup_port: ENTER,
base==0xf881c400, port_idx 2
May 25 13:00:53 localhost kernel: ahci_setup_port: base now==0xf881c600
May 25 13:00:53 localhost kernel: ahci_setup_port: EXIT
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_CMD 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_SCR_ERR 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_STAT 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_MASK 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: mmio f881c400
port_mmio f881c680
May 25 13:00:53 localhost kernel: ahci_setup_port: ENTER,
base==0xf881c400, port_idx 3
May 25 13:00:53 localhost kernel: ahci_setup_port: base now==0xf881c680
May 25 13:00:53 localhost kernel: ahci_setup_port: EXIT
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_CMD 0x6
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_SCR_ERR 0x4050000
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_STAT 0x0
May 25 13:00:53 localhost kernel: ahci_host_init: PORT_IRQ_MASK
0x7840007f
May 25 13:00:53 localhost kernel: ahci_host_init: HOST_CTL 0x80000000
May 25 13:00:53 localhost kernel: ahci_host_init: HOST_CTL 0x80000002


On Thu, 2006-05-25 at 12:04 -0400, Ming Zhang wrote:
> I added these 2 lines in ahci_host_init()
> 
>                 writel(1 << i, mmio + HOST_IRQ_STAT);
> 
>                 /* set irq mask (enables interrupts) */
>                 writel(DEF_PORT_IRQ, port_mmio + PORT_IRQ_MASK);
> 
> ~~~~~~code i added
>                 tmp = readl(port_mmio + PORT_IRQ_MASK);
>                 VPRINTK("PORT_IRQ_MASK 0x%x\n", tmp);
> ~~~~~~~~~~~~~
> 
> And I got this dmesg with VPRINTK enabled in libata.h. Does this mean
> that IRQ is not enabled on this port at all?
> 
> ACPI: PCI Interrupt 0000:00:1f.2[B] -> GSI 19 (level, low) -> IRQ 19
> ahci 0000:00:1f.2: pcs 0x8f->0x8f
> ahci_host_init: cap 0xc720ff03  port_map 0x0  n_ports 4
> ahci_host_init: mmio f8820400  port_mmio f8820500
> ahci_setup_port: ENTER, base==0xf8820400, port_idx 0
> ahci_setup_port: base now==0xf8820500
> ahci_setup_port: EXIT
> ahci_host_init: PORT_CMD 0x0
> ahci_host_init: PORT_SCR_ERR 0x0
> ahci_host_init: PORT_IRQ_STAT 0x0
> ahci_host_init: PORT_IRQ_MASK 0x0
> ahci_host_init: mmio f8820400  port_mmio f8820580
> ahci_setup_port: ENTER, base==0xf8820400, port_idx 1
> ahci_setup_port: base now==0xf8820580
> ahci_setup_port: EXIT
> ahci_host_init: PORT_CMD 0x0
> ahci_host_init: PORT_SCR_ERR 0x0
> ahci_host_init: PORT_IRQ_STAT 0x0
> ahci_host_init: PORT_IRQ_MASK 0x0
> ...
> 
> 
> On Thu, 2006-05-25 at 11:14 -0400, Jeff Garzik wrote:
> > Ming Zhang wrote:
> > > I read the whole AHCI but I still could not find out what is the magic
> > > number 0x92. Is this in some Intel doc?
> > 
> > AHCI spec only documents generic PCI config registers.  Each vendor has 
> > their own vendor-specific registers.  Intel's are documented publicly on 
> > the developer website.
> > 
> > 	Jeff
> > 
> > 
> > 
>