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

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Wed, Sep 10, 2025 at 07:56:30AM +0200, Michal Pecio wrote:
> On Tue, 9 Sep 2025 22:44:16 +0200, Michal Pecio wrote:
> > On Tue, 9 Sep 2025 14:29:20 +0300, Andy Shevchenko wrote:
> > > Not really. It prints unnecessary long values on 32-bit machines
> > > making an impression that something works somewhere in 64-bit
> > > address space.  
> > 
> > The %016llx format you are alluding to is used in two error messages
> > actually seen by users, that's an issue. My crazy personal preference
> > would be %08llx, but I concede it's unprofessional, so %pad it seems.
> 
> Actually, I take this back.
> 
> I think that leading zeros are evil and I agree this message is bad.
> But I don't understand why 64 bit users should put up with this:
> 
> [  140.106751] xhci_hcd 0000:07:00.0: Event dma 0x00000000ffeec7f0 for ep 2 status 13 not part of TD at 00000000ffeec800 - 00000000ffeec800
> [  140.476573] xhci_hcd 0000:07:00.0: Event dma 0x00000000ffeec1a0 for ep 2 status 13 not part of TD at 00000000ffeec1b0 - 00000000ffeec1b0
> [  140.502855] xhci_hcd 0000:07:00.0: Event dma 0x00000000ffeecd60 for ep 2 status 13 not part of TD at 00000000ffeecd70 - 00000000ffeecd70
> [  141.225300] xhci_hcd 0000:07:00.0: Event dma 0x00000000ffeeb970 for ep 2 status 13 not part of TD at 00000000ffeeb980 - 00000000ffeeb980
> 
> when we can have this:
> 
> [  419.967755] xhci_hcd 0000:07:00.0: Event dma 0xffc34760 for ep 2 status 13 not part of TD at 0xffc34770 - 0xffc34770
> [  420.100611] xhci_hcd 0000:07:00.0: Event dma 0xffc34bc0 for ep 2 status 13 not part of TD at 0xffc34bd0 - 0xffc34bd0
> [  420.360917] xhci_hcd 0000:07:00.0: Event dma 0xffc34e70 for ep 2 status 13 not part of TD at 0xffc34e80 - 0xffc34e80
> [  421.719530] xhci_hcd 0000:07:00.0: Event dma 0xffc35770 for ep 2 status 13 not part of TD at 0xffc35780 - 0xffc35780
> 
> with a simple change (anything wrong with u64 cast here?):
> 
> -       xhci_err(xhci, "Event dma %pad for ep %d status %d not part of TD at %016llx - %016llx\n",
> -                &ep_trb_dma, ep_index, trb_comp_code,
> +       xhci_err(xhci, "Event dma %#08llx for ep %d status %d not part of TD at %#08llx - %#08llx\n",

How is 0 will be printed with %#08x?

> +                (u64) ep_trb_dma, ep_index, trb_comp_code,
> 
> These zeros only add noise, and in many cases make difference between
> line wrapping or not because this is longer than 99% of kernel messages
> and some people want their terminal window not to take the whole screen.

I disagree on this. The 64-bit platforms are 64-bit. If the address in use is
_capable_ of 64-bit, it should be printed as 64-bit. Otherwise make it u32 in
the code and then I will agree with you.

> The main thing we care about here are the last 3-4 digits and we could
> have made it little more than (ep_trb_dma & 0xffff) long ago, but then
> Niklas asked "what about correlation with tracing/debugfs/dyndbg?", so
> it was left the way it is.

-- 
With Best Regards,
Andy Shevchenko