Re: [PATCH 4/6] x86/mtrr: don't let mtrr_type_lookup() return MTRR_TYPE_INVALID

[Juergen Gross <jgross@suse.com>]

[-- Attachment #1.1.1: Type: text/plain, Size: 1367 bytes --]

On 07.02.23 17:20, Linus Torvalds wrote:
> On Mon, Feb 6, 2023 at 11:29 PM Juergen Gross <jgross@suse.com> wrote:
>>
>> mtrr_type_lookup() should always return a valid memory type. In case
>> there is no information available, it should return the default UC.
> 
> Why isn't the second case (ie MTRR_STATE_MTRR_ENABLED not being set)
> returning 'mtrr_state.def_type'. That's what the hw does, no?

Are you sure? In the SDM I'm reading:

* E (MTRRs enabled) flag, bit 11 — MTRRs are enabled when set; all MTRRs are
   disabled when clear, and the UC memory type is applied to all of physical
   memory.

So UC it should be IMHO.

>> At the same time the mtrr_type_lookup() stub for the !CONFIG_MTRR
>> case should set uniform to 1, as if the memory range would be
>> covered by no MTRR at all.
> 
> .. but you didn't do that for the CONFIG_MTRR, so now they return
> MTRR_TYPE_UNCACHABLE, but don't set uniform=1.

I agree that setting uniform should be done at least for the case of not
enabled MTRRs.

The case of !mtrr_state_set is different, as the data regarding setting
uniform isn't known yet. OTOH there is no caller of mtrr_type_lookup()
interested in the uniform setting so early in boot, so I guess not
setting it is fine (setting it would be okay for the same reason, but
it would be technically wrong IMO).


Juergen

[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 3149 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 495 bytes --]