AER: Malformed TLP recovery deadlock with NVMe drives

Bjorn Helgaas helgaas at kernel.org
Tue May 8 09:58:49 PDT 2018


On Tue, May 08, 2018 at 01:21:38AM +0100, okaya at codeaurora.org wrote:
> On 2018-05-08 00:57, Alex_Gagniuc at Dellteam.com wrote:
> > On 5/7/2018 5:46 PM, okaya at codeaurora.org wrote:
> > [snip]
> > > > If it were easy, somebody would have patched it by now ;)
> > > 
> > > Can you file a bugzilla CC me, keith and bjorn and attach all of your
> > > logs?
> > 
> > Sure. Which bugzilla?
> > 
> 
> https://bugzilla.kernel.org
> 
> Drivers -> pci

https://bugzilla.kernel.org/show_bug.cgi?id=199657

Thanks for the bugzilla.  Can you add the "lspci -vv" output (the
"lspci -tvnn" output doesn't have the details about individual
devices) and also the specifics about how you inject the error (I
assume you're using setpci to update MPS, but it would help others
reproduce it to see the exact command and the lspci output so they can
adjust to their local topology).

I don't know how you collected the dmesg logs, but they have a bunch
of gunk in them (looks like maybe colorizing escape sequences?).  I
always think it's nice to have the plain ASCII complete log from
beginning of boot, e.g., from "dmesg" command or similar.  But I
probably won't have time to look at them myself, and maybe what's
there is enough.



More information about the Linux-nvme mailing list