RE: Packet Switch enumeration issue on Linux Kernel 4.2

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Bjorn,

I create the bug report.

https://bugzilla.kernel.org/show_bug.cgi?id=189361


Thanks,
JC

-----Original Message-----
From: Bjorn Helgaas [mailto:bhelgaas@xxxxxxxxxx] 
Sent: Wednesday, November 30, 2016 11:19 AM
To: JC Chen/PTP
Subject: Re: Packet Switch enumeration issue on Linux Kernel 4.2

On Tue, Nov 29, 2016 at 8:20 PM, JC Chen/PTP <JC_Chen@xxxxxxxxxxxxx> wrote:
> Hi Bjorn,
>
>
>
> We have two products, one is PCI-to-PCIe reverse bridge (9x111, 
> id=0x12d8e111), and another one is PCIe packet switch (2G808, 
> id=12d88608, it includes one up port and 7 down ports). When I plugged 
> 9x111+2G808+one pcie wifi card on Linux Kernel 3.19, wifi card can 
> work fine. However, if use the same hardware arch., but Linux kernel 
> is upgraded to 4.2 version, wifi card can not work fine. After check, 
> I found that 7 down ports are disappeared, and wifi’s resource is 
> clear (cfg offset 10h is clear). Without resources, wifi card can not work normally.
>
>
>
> From the kernel dump (refer to attached file),
>
> [    0.173421] pci 0000:01:00.0: [12d8:e111] type 01 class 0x060400 => 9x111
> reverse bridge is found
>
> [    0.174195] pci 0000:02:00.0: [12d8:8608] type 01 class 0x060400 => up
> port of 2G808PR is found
>
>
>
> But I can not find 7 down ports. They should be on Bus 3:01.0, 3:04.0, 
> 3:05.0, 3:06.0, 3:07.0, 3:08.0 and 3:09.0. But, I can not see it from 
> kernel dump. Could you help to debug this issue?

Sure.  Can you please try a current kernel, ideally v4.9-rc7, but at least v4.8?

If the problem still exists in a current kernel, can you please open a bug report at http://bugzilla.kernel.org, mark it as a regression, attach the complete dmesg logs from 3.19 and 4.2, and attach the output of "lspci -vv" on 3.19?

Then send me mail with a link to the bug report, and copy as well?

That way we can track the regression, and others can help debug it.

It's likely that we can debug it from that information, but if you want to bisect it to a specific commit that broke it, that would make it even easier.

Bjorn
��.n��������+%������w��{.n�����{���"�)��jg��������ݢj����G�������j:+v���w�m������w�������h�����٥




[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux