[PATCH] arm64: ignore memory outside of the linear range

On 17 August 2015 at 13:04, Catalin Marinas [off-list ref] wrote:

On Mon, Aug 17, 2015 at 12:44:50PM +0200, Ard Biesheuvel wrote:

quoted

On 17 August 2015 at 12:40, Catalin Marinas [off-list ref] wrote:

quoted

On Sat, Aug 15, 2015 at 02:13:44PM +0200, Ard Biesheuvel wrote:

quoted

We need to ensure that we don't try to map system RAM ranges whose
offset relative to the start of the kernel image exceeds the size of
the linear range. This may happen even on systems that don't have
huge amounts of RAM if it is laid out very sparsely.

Signed-off-by: Ard Biesheuvel <redacted>
---

This is the minimal fix for addressing the issue we discussed. I dropped
the other changes for now, let's revisit those when (if) my patches for
decoupling the kernel mapping from the linear mapping are back under
discussion.

I will leave it up to the maintainers whether this constitutes a bugfix or
not, but since this has never worked from the beginning afaict, I don't
think it belongs in stable per se.

Even though it is not a regression, I think it is a bug fix and it's
worth cc'ing stable (though we could push it after 4.3-rc1).

quoted

diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c
index ad87ce826cce..c65e57d4c3e7 100644
--- a/arch/arm64/mm/init.c
+++ b/arch/arm64/mm/init.c

@@ -158,6 +158,19 @@ early_param("mem", early_mem);

 void __init arm64_memblock_init(void)
 {
+     /*
+      * Remove the memory that we will not be able to cover
+      * with the linear mapping.
+      */
+     const s64 linear_region_size = -(s64)PAGE_OFFSET;
+
+     if (memstart_addr + linear_region_size < memblock_end_of_DRAM()) {
+             pr_warn("Ignoring memory outside of linear range (0x%012llx - 0x%012llx)\n",
+                     memstart_addr + linear_region_size,
+                     (u64)memblock_end_of_DRAM() - 1);
+             memblock_remove(memstart_addr + linear_region_size, ULLONG_MAX);
+     }

For memory blocks below PHYS_OFFSET, we ignore them in
early_init_dt_add_memory_arch() directly (and also print a warning).
Could we not do something similar for higher blocks? This function has a
check for MAX_PHYS_ADDR but thats -1ULL, so I guess the check is pretty
much useless. We should allow the arch code define this macro and only
leave the current value as default if not overridden.

That is actually more or less what Mark Rutland proposed here

http://article.gmane.org/gmane.linux.ports.arm.kernel/430239

Ah, I missed this. I prefer Mark's variant more (once it's fixed).

quoted

I proposed to override early_init_dt_add_memory_arch() completely
instead, but Rob had some comments, after which the discussion kind of
fizzled out.

I don't think we should override early_init_dt_add_memory_arch(), unless
we diverge too much from it. Currently, it looks to me like the end of
PA range check is pretty useless in this function.

OK. But note that, when we decouple the linear mapping from the
mapping of the kernel image, we will need to override this function
anyway.