[PATCH v3 0/8] ARM: NOMMU: MPU updates

On 14/10/17 15:37, Russell King - ARM Linux wrote:

On Fri, Oct 13, 2017 at 10:40:39AM +0100, Vladimir Murzin wrote:

quoted

On 13/10/17 10:24, Russell King - ARM Linux wrote:

quoted

Sorry, but I have no working MPS3 platform since it killed its uSD card
with all the FPGA bitstreams on.  I'm waiting (since end of September)
for a replacement from ARM Ltd but there hasn't yet been any progress.

Sure, I can replace the uSD card myself, but the problem is the contents
of the card.

Thanks for prompt reply!

I should say that board won't help you with testing this series since it
implements different MPU model (I have patches for that too but they are
on top of this series).

It assists as far as making sure the changes don't cause a
regression for the platform.

quoted

I've been working closely with Alex and Andr?s (since they needed this for
their Cortex-M7 platform) to make sure it works as expected. It was also
tested with R-class FastModels.

I intentionally kept CONFIG_ARM_MPU disabled by default for M-class
(PATCH 6/8), so it won't interfere with existing bootloaders.

It looks like patch 6 will allow me to select ARM_MPU for the MPS3
config (by way of CPU_V7M being enabled).  As I'm unable to test
what effect that has here, and I don't know when I'll be able to do
so, please check if not already done so.

It was done. To ensure that there is a change in PATCH 3/8

@@ -204,13 +205,13 @@ ENTRY(__setup_mpu)
        mrc     p15, 0, r0, c0, c1, 4           @ Read ID_MMFR0
        and     r0, r0, #(MMFR0_PMSA)           @ PMSA field
        teq     r0, #(MMFR0_PMSAv7)             @ PMSA v7
-       bne     __error_p                       @ Fail: ARM_MPU on NOT v7 PMSA
+       bxne    lr

the C side already had a check in mpu_setup()

 	if (!mpu_present())
 		return;

quoted

I'd be glad if you can review the series and give it a chance to sit in
linux-next (I can provide a branch on top of yours for-next).

Please never provide a branch on top of my for-next, it's an "unstable"
commit in that it's regenerated from several sub-branches (fixes, misc,
devel-stable, and whatever other work I want to include in linux-next.)

Ok, got it! It seems the right branch is devel-stable. Unless you have
comments, I'd like to submit patches in your patch system.

Thanks
Vladimir