Re: The MSI Driver Guide HOWTO

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

 



On Fri, Mar 10, 2023 at 3:32 PM <gael.seibert@xxxxxx> wrote:
> On 10/03/2023 19:41:02, Bjorn Helgaas wrote:
> > On Fri, Mar 10, 2023 at 11:23:14AM +0100, gael.seibert@xxxxxx wrote:
> > > On 09/03/2023 23:55:03, Bjorn Helgaas wrote:
> > > > On Thu, Mar 09, 2023 at 10:57:51AM +0100, rec wrote:
> > > > > On 09/03/2023 00:03:04, Bjorn Helgaas wrote:
> > > > > > On Tue, Mar 07, 2023 at 12:22:44PM +0100, rec wrote:
> > > > > > > Like asked in :
> > https://www.kernel.org/doc/html/latest/PCI/msi-howto.html#disabling-msis-globally
> > > > >
> > > > > > Thanks for the report!  I assume this means your system has
> > problems
> > > > > > with MSIs, and booting with "pci=nomsi" makes it work better?
> > > > >
> > > > > You are welcome,
> > > > > The system doesn't boot completely without the "pci=nomsi"
> > option.
> > > >
> > > > What exactly do you mean by "it doesn't boot completely"?  I
> > compared
> > > > the two dmesg logs, and I see that the "with MSI" log also has the
> > > > "single" parameter, so it will only boot to single-user mode.
> > >
> > > It does it mean than either the boot stop or the system halt,
> > power-off
> > > before it can be possible to connect tty console or display manager.
> >
> > Wow.  I'm not sure what would cause a sudden halt or power-off like
> > that.  Is there any indication on the console when this happens?  Can
> > you try adding the following to your kernel boot parameters to see if
> > you can catch anything via a photo or video (you may have to adjust
> > the boot_delay to make things readable):
> >
> >   nosmp ignore_loglevel lpj=lpj=7000000 boot_delay=100
>
> It will be possible that is a fan problem with a cpu temperature.
> (Probably)
> I attach a video to the boot.

Thanks for this.  I should have asked at the very beginning whether
there are any older kernels that work correctly without "pci=nomsi".
If there is such an older kernel, we can try to figure out what change
broke it.  Otherwise, I'm running out of ideas.

> > I'm curious about the Ricoh thing because I don't see an obvious MSI
> > connection.  Can you collect the output of "sudo lspci -vv"?  The
> > lspci output in your initial email wasn't collected as root, so it
> > doesn't include information about Capabilities (including MSI).
>
> Output of #lspci --vv attached

Thanks!  I was hoping something from lspci would connect with
ricoh_mmc_fixup_rl5c476(), where we get the "proprietary Ricoh MMC
controller" message, e.g., if that function looked at the MSI
Capability or something.  But 00:0d has four functions and none of
them has an MSI Capability.  And 00:0d.0 has nothing we know about at
the offsets the function uses:

  00:0d.0 FireWire (IEEE 1394): Ricoh Co Ltd R5C832
    Capabilities: [dc] Power Management version 2

Bjorn




[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