Hi, Arnd,
On Thu, Aug 12, 2021 at 8:46 PM Arnd Bergmann <arnd@xxxxxxxx> wrote:
>
> On Thu, Aug 12, 2021 at 1:05 PM Huacai Chen <chenhuacai@xxxxxxxxx> wrote:
> > On Tue, Jul 6, 2021 at 6:16 PM Arnd Bergmann <arnd@xxxxxxxx> wrote:
> > > > +
> > > > +static inline void check_bugs(void)
> > > > +{
> > > > + unsigned int cpu = smp_processor_id();
> > > > +
> > > > + cpu_data[cpu].udelay_val = loops_per_jiffy;
> > > > +}
> > >
> > > This needs a comment to explain what bug you are working around.
> > OK, there is "no bug" at present, just set the CPU0's udelay_val here.
>
> Can you do this elsewhere?
>
> I think the normal place would be somewhere around calibrate_delay(),
> which has various options.
>
> If the CPU has a well-defined hardware timer, you should not need to
> know any loops_per_jiffy value but simply wait for the correct number
> of cycles in delay()
OK, this seems need to be reworked.
>
> > > Can you send a cleanup patch for this? I think we should get rid of the
> > > isa_dma_bridge_buggy variable for architectures that do not have a VIA
> > > VP2 style ISA bridge.
> > We doesn't need ISA, but isa_dma_bridge_buggy is needed in drivers/pci/pci.c
>
> Ah right, of course. We should clean that up one day so architectures
> don't have to
> define this, I don't think anything but x86 can actually set it to a value other
> than zero.
>
> > > > +#ifdef CONFIG_64BIT
> > > > +#define TASK_SIZE32 0x80000000UL
> > >
> > > Why is the task size for compat tasks limited to 2GB? Shouldn't these
> > > be able to use the full 32-bit address space?
> > Because we use 2:2 to split user/kernel space.
>
> Usually in compat mode, the user process can access all 4GB though, regardless
> of what native 32-bit tasks can do.
>
> Is there a limitation on loongarch that prevents you from doing this?
Emm, we can define COMPAT32 to 4GB now.
>
> > > > +#ifdef CONFIG_VA_BITS_40
> > > > +#define TASK_SIZE64 (0x1UL << ((cpu_vabits > 40)?40:cpu_vabits))
> > > > +#endif
> > > > +#ifdef CONFIG_VA_BITS_48
> > > > +#define TASK_SIZE64 (0x1UL << ((cpu_vabits > 48)?48:cpu_vabits))
> > > > +#endif
> > >
> > > Why would the CPU have fewer VA bits than the page table layout allows?
> > PAGESIZE is configurable, so the range a PGD can cover is various, and
> > VA40/VA48 is not exactly 40bit/48bit, but 40bit/48bit in maximum.
>
> Do you mean the page size is not a compile-time constant in this case?
>
> What are the combinations you can support? Is this a per-task setting
> or detected at boot time?
Page size is a compile-time configuration. Maybe you or maybe me get
lost here.:)
What I want to do here:
1, I want to provide VA40 and VA48 for user-space.
2, CPU's VA bits may be 32, 36, 40, 48 or other values (so we need
(0x1UL << ((cpu_vabits > 40)?40:cpu_vabits)) or something similar).
3, The range a PGD can cover depends on PAGE SIZE (so it cannot
exactly equals to 40/48).
>
>
> > > > +/*
> > > > + * Subprogram calling convention
> > > > + */
> > > > +#define _LOONGARCH_SIM_ABILP32 1
> > > > +#define _LOONGARCH_SIM_ABILPX32 2
> > > > +#define _LOONGARCH_SIM_ABILP64 3
> > >
> > > What is the difference between ILP32 and ILPX32 here?
> > >
> > > What is the ILP64 support for, do you support both the regular LP64 and ILP64
> > > with 64-bit 'int'?
> > ABILP is ABI-LP, not AB-ILP. :). LP32 is native 32bit ABI, LP64 is
> > native 64bit ABI, and LPX32 something like MIPS N32/X86 X32 (not exist
> > in the near future).
>
> I would suggest you don't plan to ever add LPX32 in this case, it
> has never really caught on for any of the architectures that support
> it other than mips, and even there I don't think it is used much
> any more.
>
> It's certainly easier to have fewer ABI options.
LPX32 is just a placeholder now, and maybe a placeholder for ever. :)
Huacai
>
> Arnd
