Re: [PATCH 01/86] pci: export pci_ids.h

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Mar 30, 2015 at 08:55:22AM +0200, Greg KH wrote:
> On Mon, Mar 30, 2015 at 08:48:44AM +0200, Michael S. Tsirkin wrote:
> > On Sun, Mar 29, 2015 at 10:40:47PM +0200, Greg KH wrote:
> > > On Sun, Mar 29, 2015 at 03:37:01PM +0200, Michael S. Tsirkin wrote:
> > > > The macros in pci_ids.h are pretty useful for userspace
> > > > using the pci sysfs interface.
> > > > At the moment userspace is forced to duplicate these macros
> > > > (e.g. QEMU does this), it is better to expose them in
> > > > /usr/include/linux/pci_ids.h so everyone can just include
> > > > this header.
> > > > 
> > > > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > > > ---
> > > >  include/linux/pci_ids.h      | 2998 +-----------------------------------------
> > > >  include/uapi/linux/pci_ids.h | 2997 +++++++++++++++++++++++++++++++++++++++++
> > > 
> > > No, please use the pci ids file from the upstream pci id database
> > > instead.
> > 
> > 
> > >  We shouldn't be putting these all in one file,
> > 
> > pci.txt says:
> > 	Please add PCI_VENDOR_ID_xxx for vendors
> > You want to change this policy, and get rid of
> > vendor ids in pci_ids.h completely?
> 
> Please read the top of pci_ids.h.  It hasn't had new ids added to it in
> a long time.

OK, looks like pci.txt should be fixed then.

> > Bjorn, what do you think of this?
> > 
> > >  and pulling
> > > them out of drivers isn't ok.
> > 
> > This patchset is not pulling any files out of drivers fwiw.
> 
> It did for the USB gadget driver patch that I commented on.
> 
> > > Userspace shouldn't need to know any of these, use libpci.
> > 
> > Unless I'm mistaken, libpci does not export a header with defines. It
> > has a text file pci.ids, but parsing that when all I want is e.g. locate
> > all intel devices is just too much overhead. No one wants that, so
> > people just duplicate headers.
> 
> Why would userspace need the pci id of anything?

Look at how they are used e.g. by QEMU, seabios, gpxe.
People want to say e.g. "find all network class devices".

>  Again, just use
> libpci, isn't it fast enough?  Don't duplicate existing logic.

This really depends on whether you want something else that
libpci provides. But if I just want e.g. standard class IDs,
I don't want to depend on libpci.


> Or use the hw database that libudev exports, which is already on your
> machine and exports all of the pci ids from libpci directly.

Same argument really.

> > Standard class IDs are even sillier to duplicate.
> 
> Again, why does userspace need this?
> 
> thanks,
> 
> greg k-h

Mostly because it has a userspace driver, or is emulating
a hardware device.

For example, the VFIO driver uses the PCI hardware formats
as it's userspace/kernel interface. This means I need
a ton of PCI constants in userspace. Most of them are
already in pci_regs.h, and so exported. No problem here.

However class and prog interface IDs are not, they
are in pci_ids.h

Would you support splitting pci_ids.h to
include/uapi/linux/pci_ids.h with standard headers
and include/linux/pci_ids.h with device/vendor
specific ones?


-- 
MST