[PATCH 0/4] staging: mt7621-pci: Handle minor issues

Sergio Paracuellos sergio.paracuellos at gmail.com
Sun Jun 23 05:58:51 UTC 2019

Hi Brett,

On Sun, Jun 23, 2019 at 4:15 AM Brett Neumeier <bneumeier at gmail.com> wrote:
> On Fri, Jun 21, 2019 at 7:35 AM Greg Ungerer <gerg at kernel.org> wrote:
> >
> > Hi Sergio,
> >
> > On 21/6/19 4:15 pm, Sergio Paracuellos wrote:
> > > This patch series properly handle minor issues in this driver. These are:
> > > * Disable pcie port clock on pci dirver instead of doing it in the phy
> > >    driver. The pci driver is the correct place to do this.
> > > * Add a missing call to phy_exit function to properly handle the function
> > >    'mt7621_pcie_init_port' error path.
> > > * Move driver to init in a later stage using 'module_init' instead of using
> > >    'arch_initcall'.
> > >
> > > Patches are only compile-tested. It would be awasome to be tested before applied
> > > them (mainly the change to 'module_init' stuff).
> I have similar results to Greg -- on GnuBee PC1 and PC2, six boot
> attempts each on a kernel built from staging-next, I have four hangs
> and eight successful boots. The hanging patterns are similar to
> previous results. If the full boot logs would be helpful let me know,
> I can provide them.

Thanks for letting me know. One thing we can try is check init order
in v4.20 kernel. Can you please try to compile pci driver for the
kernel v4.20 tag changing
driver's last line 'arch_initcall' into 'module_init'? Just to know if
at that working driver putting all the stuff in a later stage stills
work as expected.

Full dmesg's of this v4.20 wih the change would be helpful.

Best regards,
    Sergio Paracuellos

> --
> Brett Neumeier (bneumeier at gmail.com)

More information about the devel mailing list