On Fri, Nov 19, 2021 at 12:48:39AM IST, Yonghong Song wrote:
>
>
> On 11/18/21 10:30 AM, Kumar Kartikeya Dwivedi wrote:
> > On Thu, Nov 18, 2021 at 10:57:15PM IST, Yonghong Song wrote:
> > >
> > >
> > > On 11/15/21 9:42 PM, Kumar Kartikeya Dwivedi wrote:
> > > > In CRIU, we need to be able to determine whether the page pinned by
> > > > io_uring is still present in the same range in the process VMA.
> > > > /proc/<pid>/pagemap gives us the PFN, hence using this helper we can
> > > > establish this mapping easily from the iterator side.
> > > >
> > > > It is a simple wrapper over the in-kernel page_to_pfn helper, and
> > > > ensures the passed in pointer is a struct page PTR_TO_BTF_ID. This is
> > > > obtained from the bvec of io_uring_ubuf for the CRIU usecase.
> > > >
> > > > Signed-off-by: Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx>
> > > > ---
> > > > fs/io_uring.c | 17 +++++++++++++++++
> > > > include/linux/bpf.h | 1 +
> > > > include/uapi/linux/bpf.h | 9 +++++++++
> > > > kernel/trace/bpf_trace.c | 2 ++
> > > > scripts/bpf_doc.py | 2 ++
> > > > tools/include/uapi/linux/bpf.h | 9 +++++++++
> > > > 6 files changed, 40 insertions(+)
> > > >
> > > > diff --git a/fs/io_uring.c b/fs/io_uring.c
> > > > index 46a110989155..9e9df6767e29 100644
> > > > --- a/fs/io_uring.c
> > > > +++ b/fs/io_uring.c
> > > > @@ -11295,6 +11295,23 @@ static struct bpf_iter_reg io_uring_buf_reg_info = {
> > > > .seq_info = &bpf_io_uring_buf_seq_info,
> > > > };
> > > > +BPF_CALL_1(bpf_page_to_pfn, struct page *, page)
> > > > +{
> > > > + /* PTR_TO_BTF_ID can be NULL */
> > > > + if (!page)
> > > > + return U64_MAX;
> > > > + return page_to_pfn(page);
> > > > +}
> > > > +
> > > > +BTF_ID_LIST_SINGLE(btf_page_to_pfn_ids, struct, page)
> > > > +
> > > > +const struct bpf_func_proto bpf_page_to_pfn_proto = {
> > > > + .func = bpf_page_to_pfn,
> > > > + .ret_type = RET_INTEGER,
> > > > + .arg1_type = ARG_PTR_TO_BTF_ID,
> > > > + .arg1_btf_id = &btf_page_to_pfn_ids[0],
> > >
> > > Does this helper need to be gpl_only?
>
> The typically guideline whether the same info can be retrieved from
> userspace. If yes, no gpl is needed. Otherwe, it needs to be gpl.
>
I think it can already be obtained using /proc/<PID>/pagemap, so not needed.
> Also, the helper is implemented in io_uring.c and the helper
> is used by tracing programs. Maybe we can put the helper
> in bpf_trace.c? The helper itself is not tied to io_uring, right?
>
Right, I'll move it outside CONFIG_IO_URING in v2. That's also why the build
failed on this patch.
--
Kartikeya
