On Thu, 4 Mar 2021, Huang Pei wrote:
> > > @@ -1164,8 +1165,8 @@ build_fast_tlb_refill_handler (u32 **p, struct uasm_label **l,
> > >
> > > if (pgd_reg == -1) {
> > > vmalloc_branch_delay_filled = 1;
> > > - /* 1 0 1 0 1 << 6 xkphys cached */
> > > - uasm_i_ori(p, ptr, ptr, 0x540);
> > > + /* insert bit[63:59] of CAC_BASE into bit[11:6] of ptr */
> > > + uasm_i_ori(p, ptr, ptr, (CAC_BASE >> 53));
> >
> > Instead I'd paper the issue over by casting the constant to `s64'.
> >
> > Or better yet fixed it properly by defining CAC_BASE, etc. as `unsigned
> > long long' long rather than `unsigned long' to stop all this nonsense
> > (e.g. PHYS_TO_XKPHYS already casts the result to `s64'). Thomas, WDYT?
> Sorry, I do not get it , on MIPS32, how can CAC_BASE be unsigned long
> long ?
By using the `ULL' suffix with constants. It won't change code produced,
because they are unsigned anyway and the compiler will truncate them with
no change to the actual value to fit in narrower data types as needed, but
it will silence the warnings.
Maciej
