> Date: Fri, 31 Mar 2023 22:13:44 +0200 > From: Mark Kettenis <mark.kette...@xs4all.nl> > > > Date: Fri, 31 Mar 2023 20:07:51 +0000 > > From: Klemens Nanni <k...@openbsd.org> > > > > On Fri, Mar 31, 2023 at 08:12:20PM +0200, Mark Kettenis wrote: > > > The name BM1000 only seems to come up in unofficial device trees. > > > There is a preliminary data sheet for a BE-M1000 SoC out there? Is > > > that the proper name? > > > > Right, the CPU and SoC are called "Baikal-M1" and "BE-M1000", respectively. > > > > > If the host bridge is the only PCIe device device on the SoC, consider > > > naming it just "BE-M1000" (or BM1000 if that actually is the right > > > name). That is what I did for the Rockchip PCIe host bridges. > > > > Makes sense, thanks. > > > > OK? > > You can't have a dash in the #define. I think standard practive is to > simply omit the dash, so... > > > Index: pcidevs > > =================================================================== > > RCS file: /cvs/src/sys/dev/pci/pcidevs,v > > retrieving revision 1.2027 > > diff -u -p -r1.2027 pcidevs > > --- pcidevs 31 Mar 2023 08:16:30 -0000 1.2027 > > +++ pcidevs 31 Mar 2023 20:01:29 -0000 > > @@ -344,6 +344,7 @@ vendor SYMPHONY2 0x1c1c Symphony Labs > > vendor SKHYNIX 0x1c5c SK hynix > > vendor ADATA 0x1cc1 ADATA Technology > > vendor UMIS 0x1cc4 Union Memory > > +vendor BAIKAL 0x1d39 Baikal Electronics > > vendor AQUANTIA 0x1d6a Aquantia > > vendor ROCKCHIP 0x1d87 Rockchip > > vendor TEKRAM2 0x1de1 Tekram > > @@ -2421,6 +2422,9 @@ product AVM T1 0x1200 PRI T1 > > ISDN > > > > /* AWT products */ > > product AWT RT2890 0x1059 RT2890 > > + > > +/* Baikal products */ > > +product BAIKAL BE-M1000 0x8060 BE-M1000 > > product BAIKAL BEM1000 0x8060 BE-M1000
or maybe: product BAIKAL BE_M1000 0x8060 BE-M1000 as we have cases where we replace the dash with an underscore as well. Either way, ok kettenis@