Re: [PATCH 2/7] usb: xhci: use '%pad' specifier for DMA address printing

[Michal Pecio <michal.pecio@gmail.com>]

On Wed,  3 Sep 2025 19:01:22 +0200, Niklas Neronin wrote:
> Switch all printing of DMA addresses to '%pad' specifier. This specifier
> ensures that the address is printed correctly, regardless of whether the
> kernel is running in a 32-bit or 64-bit environment.

Old %llx with (long long) cast also prints it corretly.

I had the same idea and even implemented it in some private debugging
patches, but I found %pad just annoying in practice.


%pad isn't guaranteed to be at least 64 bit long, so some DMAs from
64 bit hardware will always need to be printed with %llx or similar.

Secondly, padding is not optional with %pad. Maybe not a big deal, but
on 64 bit systems with comparatively little RAM it adds clutter.

Thirdly, %pad can't be passed by value. Hence pollution like:

> @@ -2654,7 +2654,7 @@ static int handle_tx_event(struct xhci_hcd *xhci,
>  	unsigned int slot_id;
>  	int ep_index;
>  	struct xhci_td *td = NULL;
> -	dma_addr_t ep_trb_dma;
> +	dma_addr_t ep_trb_dma, deq, td_start, td_end;
>  	struct xhci_segment *ep_seg;
>  	union xhci_trb *ep_trb;
>  	int status = -EINPROGRESS;

This function has plenty of variables already, not sure if it needs
three more. We could work around it by introducing {} scopes around
printing, or functions like print_scary_error_message(), but it ends
up being more hassle than type casting at some point.

Maybe a small helper if the verbose casts really bother people?
static inline unsigned long long dma2llx(dma_addr_t dma) {return dma;}

BTW, isn't unsigned unnecessary?