Date: Wed, 23 Sep 2015 09:26:38 -0700 From: Warner Losh <imp@bsdimp.com> To: Mehmet Erol Sanliturk <m.e.sanliturk@gmail.com> Cc: Hans Petter Selasky <hp@selasky.org>, =?UTF-8?Q?Maxime_Soul=C3=A9?= <btik-fbsd@scoubidou.com>, freebsd-usb@freebsd.org Subject: Re: Patch to allow dynamic USB quirks at boot Message-ID: <CANCZdfqm1%2BCtou2PXtrDWDFomrswa_4bXng5NcjDmh6trgdqjA@mail.gmail.com> In-Reply-To: <CAOgwaMt7zOCDVKMUNq-iuM7ir2sG36NmJrzVgcQJwhOfDBksng@mail.gmail.com> References: <55FF282D.2040003@scoubidou.com> <55FFA8EA.3030308@selasky.org> <56007395.1010006@scoubidou.com> <CAOgwaMuthKvwDJDpGXe-AuwfGtwSHfN9oDQomRbhK1d5qKwGig@mail.gmail.com> <5600FC1E.2060206@selasky.org> <CAOgwaMtzrw5ntWitLeV-HnLOL2L7pBx8QnLL9SOLT2fByVy4OA@mail.gmail.com> <7E6FCC35-5CA5-4B07-A1F1-41B73EA5B0FC@bsdimp.com> <CAOgwaMt7zOCDVKMUNq-iuM7ir2sG36NmJrzVgcQJwhOfDBksng@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Nearly all the PCI drivers have tables of vendor/device IDs. They all have slightly different forms, and none of them use a common macro / format like USB. So each driver needs a line or two to describe the table so we can extract the data from the .ko. Some drivers don't have lists, and those will need to be dealt with on a case-by-case basis. I also have another set of work to do PNP translations at run time. The bus will lie about the PNP IDs of the device so that drivers can treat new hardware just like old hardware. The obvious (and wrong) approach of forcing drivers to "take" devices only works for the most trivial drivers as nearly all of them have a table that tells about the "quirks" of the hardware since not all of that can be determined from registers alone. Warner On Tue, Sep 22, 2015 at 5:38 PM, Mehmet Erol Sanliturk < m.e.sanliturk@gmail.com> wrote: > > > On Tue, Sep 22, 2015 at 2:38 PM, Warner Losh <imp@bsdimp.com> wrote: > >> I need help with PCI driver marking. Grab my patches and we can talk. >> >> Basically, it marks all the .ko modules in a similar way that we do >> inter-module >> dependencies with plug and play info. kldxref then takes this data and >> stuffs >> it in loader.hints. I also need to write something that will parse this >> data and >> either generate a devd.conf file (easy to do from kldxref sources) or >> queues >> the load drive in the kernel (kinda hard). Bonus points for similar code >> in >> /boot/loader for any storage device that=E2=80=99s found. >> >> Warner >> >> >> > On Sep 22, 2015, at 1:12 AM, Mehmet Erol Sanliturk < >> m.e.sanliturk@gmail.com> wrote: >> > >> > >> > >> > On Mon, Sep 21, 2015 at 11:58 PM, Hans Petter Selasky <hp@selasky.org> >> wrote: >> > On 09/22/15 00:58, Mehmet Erol Sanliturk wrote: >> > On Mon, Sep 21, 2015 at 2:16 PM, Maxime Soul=C3=A9<btik-fbsd@scoubidou= .com> >> > wrote: >> > >> > One of my wishes is to move ALL of the device definition information >> for ( >> > NIC , USB , and other devices ) into files where for each device ( a >> unique >> > file to dedicated itself containing information presently defined as >> preset >> > in internal arrays with one additional information ( at least ) which >> > driver will be used for the device ) will be prepared . >> > >> > >> > Hi Mehmet, >> > >> > Warner Losh is working on this currently, CC'ed. See: >> > >> > https://reviews.freebsd.org/D3458 >> > >> > Maybe you want to help out testing? >> > >> > --HPS >> > >> > >> > >> > If I can help in any way , I will do it . >> > >> > >> > Mehmet Erol Sanliturk >> > >> > >> >> > > > What does "PCI driver marking" mean ? > > > Mehmet Erol Sanliturk > > > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfqm1%2BCtou2PXtrDWDFomrswa_4bXng5NcjDmh6trgdqjA>