In the near future, we are considering to use a common xen/domain heap for
Arm 32-bit V8-R. In this setup, both PADDR_BITS and BITS_PER_LONG will be
32. Therefore, the condition PADDR_BITS >= BITS_PER_LONG will become true.
Looking at the commit that introduce the BUILD_BUG_ON (88e3ed61642b
"x86/NUMA: make init_node_heap() respect Xen heap limit") and the current
use, it seems this check was mainly to ensure that the shift of xenheap_bits
is not going to be undefined (>> N for a N-bit type is undefined).
So far, all the shifts are using "xenheap_bits - PAGE_SHIFT". Therefore, the
existing code should work for 32-bit architecture as long as the result of
the substraction is below 32.
Therefore relax the BUILD_BUG_ON() to check that (PADDR_BITS - PAGE_SHIFT)
is not equal of above BITS_PER_LONG.
Note that we would need further precaution if we ended up to use
'xenheap_bits' alone in shifts.
Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Signed-off-by: Julien Grall <julien@xen.org>
---
Currently this change will not have any impact on the existing architectures.
The following table illustrates PADDR_BITS vs BITS_PER_LONG of different archs
------------------------------------------------
| Arch | PADDR_BITS | BITS_PER_LONG |
------------------------------------------------
| Arm_64 | 48 | 64 |
| Arm_32 | 40 | 32 |
| RISCV_64 | 56 | 64 |
| x86 | 52 | 64 |
-------------------------------------------------
However, this will change when we introduce a platform (For eg Cortex-R52) which
supports 32 bit physical address and BITS_PER_LONG. This platform does not follow
the same code path as Arm_32.
Thus, I have introduced this change as I don't see it causing a regression on
any of the supported platforms.
Changes from v1:-
1. Changed the check from "(PADDR_BITS > BITS_PER_LONG)" to "((PADDR_BITS - PAGE_SHIFT) >= BITS_PER_LONG)"
2. Updated the commit message to explain the reason for this.
v2 :-
1. Updated the commit message.
v3 :-
1. Updated the commit message.
2. Added Julien's SOB.
xen/common/page_alloc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/xen/common/page_alloc.c b/xen/common/page_alloc.c
index 62afb07bc6..c5b8c7444f 100644
--- a/xen/common/page_alloc.c
+++ b/xen/common/page_alloc.c
@@ -2245,7 +2245,7 @@ void __init xenheap_max_mfn(unsigned long mfn)
{
ASSERT(!first_node_initialised);
ASSERT(!xenheap_bits);
- BUILD_BUG_ON(PADDR_BITS >= BITS_PER_LONG);
+ BUILD_BUG_ON((PADDR_BITS - PAGE_SHIFT) >= BITS_PER_LONG);
xenheap_bits = min(flsl(mfn + 1) - 1 + PAGE_SHIFT, PADDR_BITS);
printk(XENLOG_INFO "Xen heap: %u bits\n", xenheap_bits);
}
--
2.17.1
On 05.12.2022 10:48, Ayan Kumar Halder wrote: > In the near future, we are considering to use a common xen/domain heap for > Arm 32-bit V8-R. In this setup, both PADDR_BITS and BITS_PER_LONG will be > 32. Therefore, the condition PADDR_BITS >= BITS_PER_LONG will become true. > > Looking at the commit that introduce the BUILD_BUG_ON (88e3ed61642b > "x86/NUMA: make init_node_heap() respect Xen heap limit") and the current > use, it seems this check was mainly to ensure that the shift of xenheap_bits > is not going to be undefined (>> N for a N-bit type is undefined). > > So far, all the shifts are using "xenheap_bits - PAGE_SHIFT". Therefore, the > existing code should work for 32-bit architecture as long as the result of > the substraction is below 32. > > Therefore relax the BUILD_BUG_ON() to check that (PADDR_BITS - PAGE_SHIFT) > is not equal of above BITS_PER_LONG. > > Note that we would need further precaution if we ended up to use > 'xenheap_bits' alone in shifts. > > Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com> > Signed-off-by: Julien Grall <julien@xen.org> Acked-by: Jan Beulich <jbeulich@suse.com>
© 2016 - 2024 Red Hat, Inc.