Re: Bug report for ahci-mvebu driver

[Damien Le Moal <damien.lemoal@opensource.wdc.com>]

On 1/23/23 18:10, Hajo Noerenberg wrote:
>> The same router works perfectly when I connect a laptop HDD. The same
>> DAS, including the same eSATA cable, works perfectly when connected
>> to an older Linksys WRT-1900AC v1 with Marvell Armada XP (MV78230),
>> using sata-mv driver. This is the kernel log from the older router:
>> 
> 
> Sorry to interfering in from the side, but I have observed very similar
> behavior with a Marvel 88SE6121 controller (AHCI kernel module): Very
> old SATA-1 HDDs (mostly laptop HDDs in my tests) work flawlessly,
> SATA-2/3 HDDs consistently fail. Limiting SATA speed (libata.force=1.5G
> ...) does not help.
> 
> Interestingly, SATA-2/3 HDDs do work with U-Boot and 'ancient' 3.x
> kernels. There was a suspicion that it was the PCI subsystem (the
> change in kernel 3.16 from kirkwood/pci.c to mvebu-pci). Pali Rohár did
> a great job to investigate this problem very persistently, but in the
> end we did not find a solution.
> 
> Without knowing anything about this area, there was the assumption that
> the problem must be at one of the lower levels close to the hardware.
> 
> You can find all the details and a lot of protocols in Bug
> https://bugzilla.kernel.org/show_bug.cgi?id=216094
> 
> I've been trying to revive my NAS440 for over 10 years with all the
> components (https://github.com/hn/seagate-blackarmor-nas), so I have
> the hardware available and am happy to persevere to try things or
> otherwise help.

Well, the first thing to do would be to test your system using the latest
kernel 6.2-rc5 and see if the problem still exists. If it does, please
report the issue.

-- 
Damien Le Moal
Western Digital Research