[PATCH v5 2/2] PCI: mediatek-gen3: Add 2-lanes mode support for Airoha AN7581
Christian Marangi
ansuelsmth at gmail.com
Sat Aug 15 04:04:45 PDT 2026
On Sat, Aug 15, 2026 at 08:16:56AM +0200, Manivannan Sadhasivam wrote:
> On Thu, Aug 13, 2026 at 09:30:54PM +0200, Christian Marangi wrote:
> > On Thu, Aug 13, 2026 at 12:08:20PM -0500, Bjorn Helgaas wrote:
> > > On Thu, Aug 06, 2026 at 06:53:32PM +0200, Christian Marangi wrote:
> > > > The Airoha AN7581 SoC supports configuring the first PCIe0 lane to 2-lanes
> > > > mode (x2 link) by bonding it with the second PCIe lane (PCIe1). This is
> > > > done by configuring the PCIe MUX in the SCU register.
> > > >
> > > > To correctly configure PCIe0 in x2 link, define in DT the following
> > > > additional properties:
> > > >
> > > > - additional reg, 'sec-pcie-mac' for the secondary PCIe.
> > > > - PERSTOUT reset for both main and secondary PCIE0, called 'perstout' and
> > > > 'sec-perstout'
> > > > - 'airoha,scu' property to correctly configure the SCU register for the
> > > > PCIe MUX
> > > > - 'num-lanes' set to '2' to enable PCIe0 in x2 link
> > > >
> > > > In such configuration the EQ preset are configured to the same values.
> > > >
> > > > To permit correct configuration of the PCIe link, additional logic is added
> > > > to assert and deassert the PERSTOUT resets. Support of these additional
> > > > reset was introduced in Airoha clk driver with commit
> > > > 6712f48eb3a1 ("clk: en7523: add support for dedicated PCIe PERSTOUT reset")
> > > > and on backporting of this commit also the clk driver change will be
> > > > needed.
> > >
> > > I don't see 6712f48eb3a1 in upstream or linux-next. Is this SHA1
> > > going to valid if we merge this for v7.3?
> >
> > I picked that sha1 from the git pull for the staging clk tree.
> >
>
> Ah, I missed it. In this case the SHA should be dropped and maybe we can refer
> the lore link of the patch?
>
Just yesterday it was picked and it's present in clk-next [1].
The commit didn't change as merge commit are used so the sha1 is valid.
I think we can keep it?
[1] https://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git/commit/drivers/clk/clk-en7523.c?h=clk-next&id=6712f48eb3a15842925171441505fa2e7699c57e
--
Ansuel
More information about the Linux-mediatek
mailing list