[PATCH 3/4] ARM: mm: kill unused TLB_CAN_READ_FROM_L1_CACHE and use ALT_SMP instead

[Will Deacon]

On Wed, May 15, 2013 at 03:46:20PM +0100, Gregory CLEMENT wrote:
> On 05/15/2013 04:06 PM, Will Deacon wrote:
> > You could also try deleting both of the ALT_* lines and just putting a
> > W(nop) in there directly.
> 
> If I just delete the both of the ALT_* lines it no more hangs.
> If I put a  W(nop) instead it hangs.

Wow. This doesn't sound good for your CPU and you might want to check with
the Marvell guys...

Extra things you could try:

	- Try adding a 16-bit nop instead (remove the W, build thumb-2 and
	  double check in your diasassembly)

	- Try adding the W(nop) to other places in the kernel and see if you
	  can tickle the lock-up elsewhere.

	- Can you reproduce on the Armada XP? (since I have access to one of
	  those)

> > Early in boot (head.S:__fixup_smp), we detect whether the CPU has the MP
> > extensions then, if we're actually UP but the kernel has CONFIG_SMP=y, we
> > walk over the .alt.smp.init section looking at each entry in there. The
> > ALT_UP macro spits out an (address, instruction) pair, so in
> > __do_fixup_smp_on_up, we store the instruction to the address for each pair,
> > replacing the SMP instruction which sat there in the compiled image.
> > 
> > It could be that the CPUID checks are failing on your Marvell part. Can you
> > tell me what you have in your mpidr (mrc p15, 0, Rd, c0, c0, 5) and also
> > your midr (mrc p15, 0, Rd, c0, c0) please?
> 
> mpidr= 0x0
> midr= 0x561F5811

Should be fine, but it doesn't look like the patching is the issue here.

Will