Re: [PATCH v7 3/7] xen/console: use memcpy() in conring_puts()

On 17.07.2026 08:18, [email protected] wrote:
> On Thu, Jul 16, 2026 at 06:08:01PM +0200, Jan Beulich wrote:
>> On 13.07.2026 20:16, [email protected] wrote:
>>> --- a/xen/drivers/char/console.c
>>> +++ b/xen/drivers/char/console.c
>>> @@ -360,12 +360,22 @@ static DECLARE_SOFTIRQ_TASKLET(conring_tasklet, conring_notify, NULL);
>>>  /* NB: Do not send conring VIRQs during panic. */
>>>  static bool conring_no_notify;
>>>  
>>> -static void conring_puts(const char *str, size_t len)
>>> +static void conring_puts(const char *str, unsigned int len)
>>>  {
>>> +    unsigned int src = likely(len < conring_size) ? 0 : len - conring_size;
>>
>> Isn't the attempt to handle strings longer than conring_size pretty much in
>> vain? Things look correct as far as the rest of the logic in the function
>> goes, but is this useful? It certainly warrants at least a brief comment imo.
> 
> Thanks for taking a look!
> 
> I don't think there ever will callers of conring_puts() with strings longer
> than conring_size... and I could not decide whether to keep the extra check
> or not.

Having though about this a little more over night, I actually wonder whether
e.g. len > PAGE_SIZE wouldn't already be indicative of an error. And which
hence may instead want flagging by a BUG_ON(). REST maintainers, thoughts?

Jan