Date: Fri, 08 Apr 2016 09:11:15 -0700 From: John Baldwin <jhb@freebsd.org> To: Stanislav Galabov <sgalabov@freebsd.org> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r297717 - head/sys/mips/mediatek Message-ID: <2014592.cEATiW0yl7@ralph.baldwin.cx> In-Reply-To: <201604081526.u38FQnvC062652@repo.freebsd.org> References: <201604081526.u38FQnvC062652@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday, April 08, 2016 03:26:49 PM Stanislav Galabov wrote: > Author: sgalabov > Date: Fri Apr 8 15:26:49 2016 > New Revision: 297717 > URL: https://svnweb.freebsd.org/changeset/base/297717 > > Log: > Introduce Mediatek/Ralink PCIe support > > This revision introduces PCIe support for the relevant Mediatek/Ralink > SoCs. > Currently the PCIe support is not converted to INTRNG, this may be a > task for the future. > > Approved by: adrian (mentor) > Sponsored by: Smartcom - Bulgaria AD > Differential Revision: https://reviews.freebsd.org/D5886 I would _really_ like to get away from having N different Foo-PCI bridge drivers that all do an initial walk of the PCI hierarchy configuring BARs and PCI-PCI bridge I/O windows. With NEW_PCIB, the PCI-PCI bridge driver is capable of figuring out windows on its own, and the PCI bus is also capable of allocating resource for BARs. Leveraging that means removing N copies of this approach and having one place to manage BARs, etc. Can you please try removing all of your I/O window and BAR setup and seeing if the PCI bus works ok? If not, we should fix it so that it does, but fixing the PCI bus means we only have to fix one place, not N. (And yes, I want to axe the code from the various arm places that do this as well and want to avoid adding even more copies of it.) -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2014592.cEATiW0yl7>