Marvell eSATA - 88SE6121 on Asus M3A79-T Deluxe - not in AHCI
Phil Reynolds
motherboard, an Asus M3A79-T Deluxe, but have found that it is not in
AHCI mode and therefore not hotplug capable. There are no settings for
the Marvell 88SE6121 controller available, other than simply whether
it is enabled or not.
I gather from reading elsewhere that the kernel can put this
controller into AHCI - I found a patch but found it was already
present in the kernel in Debian lenny's backports. It is their version
2.6.32-bpo.3 - not entirely sure which exact source version this is
built from. However, this patch does not work on this motherboard -
either that or I have to make it active somehow, and the method of
doing so is not obvious.
Of course, I could be missing something simple - if so, please accept
my apologies and point me in the right direction.
However, if further information or testing is required, I am more than
happy to assist in any way I can.
I am not on the linux-kernel mailing list at present, so please CC any
replies to the address I sent this from.
Thanks in advance for any assistance.
--
Phil Reynolds
email: phil-...@tinsleyviaduct.com
Not subscribed to lkml - please CC replies to me
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Phil Turmel
It seems no one noticed your question. I've CC'd linux-scsi, in case the
experts have additional/better advice.
On 05/09/2010 10:13 AM, Phil Reynolds wrote:
> I have recently had cause to start using the eSATA port on my
> motherboard, an Asus M3A79-T Deluxe, but have found that it is not in
> AHCI mode and therefore not hotplug capable. There are no settings for
> the Marvell 88SE6121 controller available, other than simply whether it
> is enabled or not.
>
> I gather from reading elsewhere that the kernel can put this controller
> into AHCI - I found a patch but found it was already present in the
> kernel in Debian lenny's backports. It is their version 2.6.32-bpo.3 -
> not entirely sure which exact source version this is built from.
> However, this patch does not work on this motherboard - either that or I
> have to make it active somehow, and the method of doing so is not obvious.
[snip /]
I have a similar chipset in my office server: 88SE6145. If yours can run
in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel command
line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
ahci built-in, and no initramfs, so I put the option in my grub.conf.)
My Marvell chipset has both SATA and PATA ports... the PATA ports are
non-functional when the ahci driver runs the chipset. Your mileage may
vary.
HTH,
Phil Turmel
Phil Reynolds
> I have a similar chipset in my office server: 88SE6145. If yours can run
> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel command
> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
Tried in grub.conf already, no success... will try the modprobe.conf
way in case of differing mileage.
> My Marvell chipset has both SATA and PATA ports... the PATA ports are
> non-functional when the ahci driver runs the chipset. Your mileage may
> vary.
I gather the PATA port on my board is on the Southbridge instead - but
I don't use it anyway.
--
Phil Reynolds
email: phil-...@tinsleyviaduct.com
Not subscribed to lkml - please CC replies to me
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
--
Phil Reynolds
> I have a similar chipset in my office server: 88SE6145. If yours can run
> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel command
> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
Neither works for me unfortunately.
--
Phil Reynolds
email: phil-...@tinsleyviaduct.com
Not subscribed to lkml - please CC replies to me
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
--
Phil Turmel
> Quoting "Phil Turmel" <pturmel...@turmel.org>:
>
>> I have a similar chipset in my office server: 88SE6145. If yours can run
>> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel
>> command
>> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
>> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
>
> Neither works for me unfortunately.
>
Hmmm. If you post your dmesg, lspci, and config, I'll be happy to compare
to my setup. It might be something as simple as the Marvell driver loading
first and claiming the device.
The relevant part of my dmesg looks like this:
[ 4.802703] ahci 0000:06:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
[ 4.806351] ahci 0000:06:00.0: controller can't do NCQ, turning off CAP_NCQ
[ 4.809921] ahci 0000:06:00.0: controller can't do PMP, turning off CAP_PMP
[ 4.813334] ahci 0000:06:00.0: MV_AHCI HACK: port_map 1f -> f
[ 4.816656] ahci 0000:06:00.0: Disabling your PATA port. Use the boot option 'ahci.marvell_enable=0' to avoid this.
[ 4.823333] ahci: SSS flag set, parallel bus scan disabled
[ 4.826765] ahci 0000:06:00.0: AHCI 0001.0000 32 slots 5 ports 3 Gbps 0xf impl RAID mode
[ 4.830240] ahci 0000:06:00.0: flags: 64bit stag slum part
[ 4.833810] ahci 0000:06:00.0: setting latency timer to 64
[ 4.834021] scsi4 : ahci
[ 4.837755] scsi5 : ahci
[ 4.841426] scsi6 : ahci
[ 4.844997] scsi7 : ahci
[ 4.848509] scsi8 : ahci
[ 4.851956] ata5: SATA max UDMA/133 irq_stat 0x02400000, PHY RDY changed irq 17
[ 4.855423] ata6: SATA max UDMA/133 irq_stat 0x02400000, PHY RDY changed irq 17
[ 4.858814] ata7: SATA max UDMA/133 abar m1024@0xfbfffc00 port 0xfbfffe00 irq 17
[ 4.862177] ata8: SATA max UDMA/133 abar m1024@0xfbfffc00 port 0xfbfffe80 irq 17
[ 4.865472] ata9: DUMMY
The Marvell driver never gets a chance to claim my device, as it is disabled in
my kernel config. If you are using modules, you might want to blacklist "sata_mv".
If you are using modules from an initramfs, you might need to update your
initramfs, too.
Phil
John Stoffel
Phil> Quoting "Phil Turmel" <pturmel...@turmel.org>:
>> I have a similar chipset in my office server: 88SE6145. If yours can run
>> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel command
>> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
>> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
Phil> Neither works for me unfortunately.
Can you post your dmesg output from both bootups? Are there any
differences? Also submit the lspci -vvv output as well, just so
people can help you better.
John
Jeff Garzik
sata_mv is not related to this hardware.
pata_marvell works with 6101/6121/6141/6145 PATA ports.
ahci works with 6101/6121/6141/6145 SATA ports, with marvell_enable=1.
Jeff
Phil Turmel
On 05/11/2010 05:09 PM, Jeff Garzik wrote:
> On 05/11/2010 02:28 PM, Phil Turmel wrote:
>> On 05/11/2010 01:21 PM, Phil Reynolds wrote:
>>> Quoting "Phil Turmel"<pturmel...@turmel.org>:
>>>
>>>> I have a similar chipset in my office server: 88SE6145. If yours
>>>> can run
>>>> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel
>>>> command
>>>> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I
>>>> have
>>>> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
>>>
>>> Neither works for me unfortunately.
>>>
[snip /]
> sata_mv is not related to this hardware.
>
> pata_marvell works with 6101/6121/6141/6145 PATA ports.
>
> ahci works with 6101/6121/6141/6145 SATA ports, with marvell_enable=1.
>
> Jeff
I think you might have me (Phil T) confused with the original poster, Phil
Reynolds [1].
I have working SATA hotplug on my 88SE6145 w/ the ahci driver and
marvell_enable=1. The OP, Phil R., has an 88SE6121 that has been running
with the Marvell driver, without hotplug. Neither of us cares about the
PATA ports. Phil R. wants hotplug, but has not successfully enabled
marvell support in the ahci driver.
I've suggested Phil R. post his dmesg, lspci, and config, so we can figure out
what's going on.
Did I miss something else?
Phil
[1] http://marc.info/?l=linux-kernel&m=127341691613147&w=2
Jeff Garzik
> Hi Jeff,
>
> On 05/11/2010 05:09 PM, Jeff Garzik wrote:
>> On 05/11/2010 02:28 PM, Phil Turmel wrote:
>>> On 05/11/2010 01:21 PM, Phil Reynolds wrote:
>>>> Quoting "Phil Turmel"<pturmel...@turmel.org>:
>>>>
>>>>> I have a similar chipset in my office server: 88SE6145. If yours
>>>>> can run
>>>>> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel
>>>>> command
>>>>> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I
>>>>> have
>>>>> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
>>>>
>>>> Neither works for me unfortunately.
>>>>
>
> [snip /]
>
>> sata_mv is not related to this hardware.
>>
>> pata_marvell works with 6101/6121/6141/6145 PATA ports.
>>
>> ahci works with 6101/6121/6141/6145 SATA ports, with marvell_enable=1.
>>
>> Jeff
>
> I think you might have me (Phil T) confused with the original poster, Phil
> Reynolds [1].
No confusion. You suggested blacklisting sata_mv, and I note above that
sata_mv has nothing to do with the hardware in question.
Jeff
Phil Reynolds
>>>>>> "Phil" == Phil Reynolds <phil-...@tinsleyviaduct.com> writes:
>
> Phil> Quoting "Phil Turmel" <pturmel...@turmel.org>:
>>> I have a similar chipset in my office server: 88SE6145. If yours can run
>>> in AHCI mode, you just need "ahci.marvell_enable=1" on your kernel command
>>> line, or "options ahci marvell_enable=1" in your modprobe.conf. (I have
>>> ahci built-in, and no initramfs, so I put the option in my grub.conf.)
>
> Phil> Neither works for me unfortunately.
>
> Can you post your dmesg output from both bootups? Are there any
> differences? Also submit the lspci -vvv output as well, just so
> people can help you better.
After following advice from various people, I am about there - though
it doesn't seem to last many changes, the eSATA port is now hotplug
capable.
I still have issues to solve to make it perform more like USB does,
but they are not related to anything in the kernel or SCSI layer.
Thanks to all involved.
--
Phil Reynolds
email: phil-...@tinsleyviaduct.com
Not subscribed to lkml - please CC replies to me
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.