On Wed, Feb 28, 2024 at 12:20 PM Sumit Garg <sumit.garg@xxxxxxxxxx> wrote:
>
> On Wed, 28 Feb 2024 at 14:11, Jens Wiklander <jens.wiklander@xxxxxxxxxx> wrote:
> >
> > On Wed, Feb 28, 2024 at 6:58 AM Sumit Garg <sumit.garg@xxxxxxxxxx> wrote:
> > >
> > > On Tue, 27 Feb 2024 at 21:20, Jens Wiklander <jens.wiklander@xxxxxxxxxx> wrote:
> > > >
> > > > On Tue, Feb 27, 2024 at 7:06 AM Sumit Garg <sumit.garg@xxxxxxxxxx> wrote:
> > [snip]
> > > > > > --- a/include/linux/tee_drv.h
> > > > > > +++ b/include/linux/tee_drv.h
> > > > > > @@ -275,6 +275,17 @@ void *tee_get_drvdata(struct tee_device *teedev);
> > > > > > struct tee_shm *tee_shm_alloc_priv_buf(struct tee_context *ctx, size_t size);
> > > > > > struct tee_shm *tee_shm_alloc_kernel_buf(struct tee_context *ctx, size_t size);
> > > > > >
> > > > > > +int tee_shm_pool_op_alloc_helper(struct tee_shm_pool *pool, struct tee_shm *shm,
> > > > > > + size_t size, size_t align,
> > > > > > + int (*shm_register)(struct tee_context *ctx,
> > > > > > + struct tee_shm *shm,
> > > > > > + struct page **pages,
> > > > > > + size_t num_pages,
> > > > > > + unsigned long start));
> > > > > > +void tee_shm_pool_op_free_helper(struct tee_shm_pool *pool, struct tee_shm *shm,
> > > > > > + int (*shm_unregister)(struct tee_context *ctx,
> > > > > > + struct tee_shm *shm));
> > > > > > +
> > > > >
> > > > > These rather belong to drivers/tee/tee_private.h as we shouldn't
> > > > > expose them to other kernel client drivers.
> > > >
> > > > This is the right place, this .h file is for TEE drivers too.
> > > >
> > >
> > > But this is shared with other kernel TEE client drivers and we
> > > shouldn't expose internal APIs which aren't meant for them with a side
> > > effect of API abuse too. Any particular reason to not use
> > > drivers/tee/tee_private.h?
> >
> > drivers/tee/tee_private.h is supposed to be used internally by only
> > the files in drivers/tee.
>
> How about "struct tee_device" being in drivers/tee/tee_private.h?
What about it?
>
> > If you look in include/linux/tee_drv.h
> > you'll find a few functions and other definitions that a kernel TEE
> > client driver should not use, for instance, tee_device_register() and
> > tee_device_unregister(). This kernel TEE client interface was
> > introduced with commit 25559c22cef8 ("tee: add kernel internal client
> > interface"). include/linux/tee_drv.h existed before we even had any
> > kernel TEE client interface.
>
> Anyhow, it looks like there is a chance for refactoring here. How
> about splitting this header further in something like
> include/linux/tee_core.h which will contain all the pieces relevant to
> TEE drivers?
Sure.
>
> BTW, this patch series can keep using include/linux/tee_drv.h for the
> time being.
Agreed.
Cheers,
Jens
>
> -Sumit
>
> >
> > Cheers,
> > Jens
