jetson-tx1 pcie2sata connect hdd disk error with ahci msi as interrupt

Hi Pushkar90,
This looks like a different issue.
Can we have some more information on this?
-> Did you make sure that the SMPS being used there works correctly? We have seen issues with some of the SMPS (used to supply power to drive) that are not supplying power properly to drives
-> Have you connected any driver to port-0 of the multiplier? It seems this is mandatory
-> Have you made sure that the drive works fine? (by connecting to an x86 system??)
-> Have you tried connecting the drive directly without use of a port multiplier?

Yes, using the same setup swapping out the x1 for an x86 machine works perfectly.

driver? Do u mean drive? if so, yes there is a working drive on port 0 as well as having drives on all the ports.

yes, the exact same harware setup works perfectly on an x86 machine.

no i havent tried that, as i dont have the required cable on hand that’d go from the HBA to a regular sata. But Why would this matter?

I think it definitely is a related issue because the HBA uses a Marvell 9215 controller like the OP.
And i’m also getting similar errors. heres some more of the relevant dmesg logs:

mc-err: (0) csr_afir: EMEM decode error on PDE or PTE entry
[   24.644772] mc-err:   status = 0x6000000e; addr = 0x00000000
[   24.644775] mc-err:   secure: no, access-type: read, SMMU fault: nr-nw-s
[   30.074768] mc-err: Too many MC errors; throttling prints
[   59.694762] ata14: softreset failed (1st FIS failed)
[   59.694775] smmu_dump_pagetable(): fault_address=0x0000000000000000 pa=0xffffffffffffffff bytes=ffffffffffffffff #pte=0 in L2

Ok… this new log makes it look like a Jetson-TX1 specific issue.
Can you please attach the full log?

Yeah sorry, should’ve posted more of the log.
Its gonna be huge. But sure:
Nope, can’t do it. I’m hitting some kind of limit here, so i’ll post a link to pastebin instead:
http://pastebin.com/NB8v8qq3

any ideas on what I could try to fix it?

can you revert back previous generation PCIe drivers ? our custom carrier boards dead with latest Jetpack 2.3
Thanks for fucking up guys.

Hi raak,
Can you tell us the version where it was working fine? (24.1 / 23.2 Etc…)

No Vidyas, our boards are dead after we flashed with 24.1. Every thing was fine with ubuntu 14 and previous Jetpack.

smmu_dump_pagetable(): fault_address=0x0000000000000000 pa=0xffffffffffffffff bytes=ffffffffffffffff #pte=0 in L2
May  6 10:37:59 tegra-ubuntu kernel: [17879.112384] mc-err: (0) csw_afiw: EMEM decode error on PDE or PTE entry
May  6 10:37:59 tegra-ubuntu kernel: [17879.119036] mc-err:   status = 0x60010031; addr = 0x00000000
May  6 10:37:59 tegra-ubuntu kernel: [17879.124697] mc-err:   secure: no, access-type: write, SMMU fault: nr-nw-s

I mean seriously guys.