Re: [PATCH net v2 8/8] xsk: fix u64 descriptor address truncation on 32-bit architectures

[Stanislav Fomichev <sdf.kernel@gmail.com>]

On 04/21, Jason Xing wrote:
> On Tue, Apr 21, 2026 at 3:49 AM Stanislav Fomichev <sdf.kernel@gmail.com> wrote:
> >
> > On 04/20, Jason Xing wrote:
> > > From: Jason Xing <kernelxing@tencent.com>
> > >
> > > In copy mode TX, xsk_skb_destructor_set_addr() stores the 64-bit
> > > descriptor address into skb_shinfo(skb)->destructor_arg (void *) via a
> > > uintptr_t cast:
> > >
> > >     skb_shinfo(skb)->destructor_arg = (void *)((uintptr_t)addr | 0x1UL);
> > >
> > > On 32-bit architectures uintptr_t is 32 bits, so the upper 32 bits of
> > > the descriptor address are silently dropped. In XDP_ZEROCOPY unaligned
> > > mode the chunk offset is encoded in bits 48-63 of the descriptor
> > > address (XSK_UNALIGNED_BUF_OFFSET_SHIFT = 48), meaning the offset is
> > > lost entirely. The completion queue then returns a truncated address to
> > > userspace, making buffer recycling impossible.
> > >
> > > Fix this by handling the 32-bit case directly in
> > > xsk_skb_destructor_set_addr(): when !CONFIG_64BIT, allocate an xsk_addrs
> > > struct (the same path already used for multi-descriptor SKBs) to store
> > > the full u64 address.
> >
> > Is it easier to make XSK `depends on 64BIT` to avoid dealing with that? Does
> 
> Of course, it would be super easy. Actually the initial version looks
> like this. One line of coder is simply enough.
> 
> > anybody seriously run af_xdp on 32 bit systems?
> 
> But my worry as you guess is if there exists a 32 bit system? I doubt
> it. That's why I put some effort into adding the compatibility code to
> cover the case. Good news is that it doesn't add any side effects of
> the 64 bit system since they are protected under IS_ENABLED condition.

If someone complains, we can follow up? af_xdp is all u64 everywhere,
including uapi, I doubt someone is using it on 32 bit systems. We don't
test this part on 32 bit systems on nipa either..