iommu: fix min_not_zero() type mismatch warning

Message ID 20191210195803.866126-1-arnd@arndb.de
State New
Headers show
Series
  • iommu: fix min_not_zero() type mismatch warning
Related show

Commit Message

Arnd Bergmann Dec. 10, 2019, 7:57 p.m.
min()/max() require the arguments to be of the same type.

When dma_addr_t is not compatible with __u64, this causes
a warning:

In file included from include/linux/list.h:9,
                 from include/linux/kobject.h:19,
                 from include/linux/of.h:17,
                 from include/linux/irqdomain.h:35,
                 from include/linux/acpi.h:13,
                 from include/linux/acpi_iort.h:10,
                 from drivers/iommu/dma-iommu.c:11:
drivers/iommu/dma-iommu.c: In function 'iommu_dma_alloc_iova':
include/linux/kernel.h:844:29: error: comparison of distinct pointer types lacks a cast [-Werror]
   (!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
                             ^~
include/linux/kernel.h:858:4: note: in expansion of macro '__typecheck'
   (__typecheck(x, y) && __no_side_effects(x, y))
    ^~~~~~~~~~~
include/linux/kernel.h:868:24: note: in expansion of macro '__safe_cmp'
  __builtin_choose_expr(__safe_cmp(x, y), \
                        ^~~~~~~~~~
include/linux/kernel.h:877:19: note: in expansion of macro '__careful_cmp'
 #define min(x, y) __careful_cmp(x, y, <)
                   ^~~~~~~~~~~~~
include/linux/kernel.h:910:39: note: in expansion of macro 'min'
  __x == 0 ? __y : ((__y == 0) ? __x : min(__x, __y)); })
                                       ^~~
drivers/iommu/dma-iommu.c:424:14: note: in expansion of macro 'min_not_zero'
  dma_limit = min_not_zero(dma_limit, dev->bus_dma_limit);
              ^~~~~~~~~~~~

Add an explicit cast to work around it, as there is no min_not_zero_t()
equivalent of min_t().

Fixes: a7ba70f1787f ("dma-mapping: treat dev->bus_dma_mask as a DMA limit")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

---
 drivers/iommu/dma-iommu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

-- 
2.20.0

Comments

Christoph Hellwig Dec. 11, 2019, 7:43 a.m. | #1
I think we agree that the parameter should be a u64 instead, and either
Nicolas or Robin (sorry, fading memory) promised to send me a patch for
that.

Patch

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 0cc702a70a96..6fa32231dc9f 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -421,7 +421,7 @@  static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
 	if (iova_len < (1 << (IOVA_RANGE_CACHE_MAX_SIZE - 1)))
 		iova_len = roundup_pow_of_two(iova_len);
 
-	dma_limit = min_not_zero(dma_limit, dev->bus_dma_limit);
+	dma_limit = min_not_zero((u64)dma_limit, dev->bus_dma_limit);
 
 	if (domain->geometry.force_aperture)
 		dma_limit = min(dma_limit, domain->geometry.aperture_end);