[RFC v3 21/26] bus: hide bus object
David Marchand
david.marchand at redhat.com
Thu Jul 28 21:26:16 CEST 2022
On Thu, Jul 28, 2022 at 6:57 PM Bruce Richardson
<bruce.richardson at intel.com> wrote:
>
> On Thu, Jul 28, 2022 at 05:26:35PM +0200, David Marchand wrote:
> > Make rte_bus opaque for non internal users.
> > This will make extending this object possible without breaking the ABI.
> >
> > Introduce a new driver header and move rte_bus definition and helpers.
> > Update drivers and library to use the internal header.
> >
> > Some applications may have been dereferencing rte_bus objects, mark
> > this object's accessors as stable.
> >
> > Signed-off-by: David Marchand <david.marchand at redhat.com>
> > ---
> > Changes since RFC v2:
> > - updated release notes,
> > - marked accessors as stable,
> >
> > Changes since RFC v1:
> > - update all existing users of the public header to use the internal one,
> >
>
> Acked-by: Bruce Richardson <bruce.richardson at intel.com>
>
> One small comment below...
>
> > ---
> > app/test/test_devargs.c | 2 +-
> <snip>
> > lib/pcapng/rte_pcapng.c | 2 +-
> > 38 files changed, 343 insertions(+), 319 deletions(-)
> > create mode 100644 lib/eal/include/bus_driver.h
> >
> > diff --git a/app/test/test_devargs.c b/app/test/test_devargs.c
> > index ac5bc34c18..0a4c34a1ad 100644
> > --- a/app/test/test_devargs.c
> > +++ b/app/test/test_devargs.c
> > @@ -9,7 +9,7 @@
> > #include <rte_common.h>
> > #include <rte_devargs.h>
> > #include <rte_kvargs.h>
> > -#include <rte_bus.h>
> > +#include <bus_driver.h>
> > #include <rte_class.h>
> >
>
> This seems a strange file to need to access bus internals. Maybe something
> to investigate in future.
Yeah... I am not sure it is worth fixing.
This code dereferences a bus object to check the ring driver
registered (by calling vdev_bus->parse()).
We could make it possible to list the drivers (maybe per bus, or even globally).
Or we may expose something like bus_can_parse() / rte_bus_find_by_device_name().
But I did not find applications needing this so I let the code as is.
--
David Marchand
More information about the dev
mailing list