Re: [PATCH v3 01/10] drivers/char: Add support for USB3 DbC debugger

[Jan Beulich <jbeulich@xxxxxxxx>]

On 04.08.2022 16:41, Marek Marczykowski-Górecki wrote:
> On Thu, Aug 04, 2022 at 04:36:35PM +0200, Jan Beulich wrote:
>> On 04.08.2022 16:28, Marek Marczykowski-Górecki wrote:
>>> On Thu, Aug 04, 2022 at 04:21:01PM +0200, Jan Beulich wrote:
>>>> L"Xen" looks sufficiently readable to me. We use this all over the
>>>> place in the EFI interfacing code.
>>>
>>> Ok, I can try that. But given later adjustments, IIUC it will make the
>>> whole 50+ pages structure land in .data. Is that okay?
>>
>> No. I was actually expecting the piece of data we're talking about here
>> to land in .rodata, and hence be re-usable at the same address for all
>> devices. Hence my reference to string literals.
> 
> "all devices" - this driver supports only a single xhci debug console at
> a time.

Oh, sorry - I've got confused by your multiple consoles patch here.

> Anyway, as explained earlier, it would require reserving the
> whole page for it (there are no other xhci-related structures that can
> live in .rodata), which given your earlier comments about memory usage
> is probably worse.

In your earlier reply you did say you'd see no issue with this sitting
side by side with other string literals. Which is precisely how I
would envision to avoid the need to reserve the entire page. But yes,
if that's not feasible, then the current model of keeping the stuff
in .bss is likely best. A remark in the description towards the purpose
here and the further intentions might help, not the least to avoid me
coming up with the same comment again for a future version of the
series.

Jan