[PATCH] mtd: physmap: physmap-bt1-rom: Fix __iomem addrspace removal warning

[PATCH] mtd: physmap: physmap-bt1-rom: Fix __iomem addrspace removal warning

[Serge Semin]

sparse is unhappy with us casting the __iomem address space pointer to
a type with no address space attribute specified:

"sparse warnings: (new ones prefixed by >>)"
>> drivers/mtd/maps/physmap-bt1-rom.c:78:18: sparse: sparse: cast removes address space '__iomem' of expression

Indeed we perform the __iomem-less type casting but to an integer
variable. The integer variable isn't dereferenced then, so the casting is
safe and won't cause any problem. But in order to make sparse happy let's
fix the warning just by converting the local "shift" and "chunk" variables
to the "unsigned int" type. Add the same fix to the bt1_rom_map_read()
method for unification.

Fixes: b3e79e7682e0 ("mtd: physmap: Add Baikal-T1 physically mapped ROM support")
Link: https://lore.kernel.org/lkml/202011021254.XC70BaQT-lkp@intel.com/
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Cc: Pavel Parkhomenko <Pavel.Parkhomenko@baikalelectronics.ru>
---
 drivers/mtd/maps/physmap-bt1-rom.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/mtd/maps/physmap-bt1-rom.c b/drivers/mtd/maps/physmap-bt1-rom.c
index 27cfe1c63a2e..b7fd5a3759d0 100644
--- a/drivers/mtd/maps/physmap-bt1-rom.c
+++ b/drivers/mtd/maps/physmap-bt1-rom.c
@@ -31,12 +31,12 @@ static map_word __xipram bt1_rom_map_read(struct map_info *map,
 					  unsigned long ofs)
 {
 	void __iomem *src = map->virt + ofs;
-	unsigned long shift;
+	unsigned int shift;
 	map_word ret;
 	u32 data;
 
 	/* Read data within offset dword. */
-	shift = (unsigned long)src & 0x3;
+	shift = (unsigned int)src & 0x3;
 	data = readl_relaxed(src - shift);
 	if (!shift) {
 		ret.x[0] = data;
@@ -60,7 +60,7 @@ static void __xipram bt1_rom_map_copy_from(struct map_info *map,
 					   ssize_t len)
 {
 	void __iomem *src = map->virt + from;
-	ssize_t shift, chunk;
+	unsigned int shift, chunk;
 	u32 data;
 
 	if (len <= 0 || from >= map->size)
@@ -75,7 +75,7 @@ static void __xipram bt1_rom_map_copy_from(struct map_info *map,
 	 * up into the next three stages: unaligned head, aligned body,
 	 * unaligned tail.
 	 */
-	shift = (ssize_t)src & 0x3;
+	shift = (unsigned int)src & 0x3;
 	if (shift) {
 		chunk = min_t(ssize_t, 4 - shift, len);
 		data = readl_relaxed(src - shift);
-- 
2.29.2

Re: [PATCH] mtd: physmap: physmap-bt1-rom: Fix __iomem addrspace removal warning

[Miquel Raynal]

On Thu, 2020-11-12 at 18:00:37 UTC, Serge Semin wrote:
> sparse is unhappy with us casting the __iomem address space pointer to
> a type with no address space attribute specified:
> 
> "sparse warnings: (new ones prefixed by >>)"
> >> drivers/mtd/maps/physmap-bt1-rom.c:78:18: sparse: sparse: cast removes address space '__iomem' of expression
> 
> Indeed we perform the __iomem-less type casting but to an integer
> variable. The integer variable isn't dereferenced then, so the casting is
> safe and won't cause any problem. But in order to make sparse happy let's
> fix the warning just by converting the local "shift" and "chunk" variables
> to the "unsigned int" type. Add the same fix to the bt1_rom_map_read()
> method for unification.
> 
> Fixes: b3e79e7682e0 ("mtd: physmap: Add Baikal-T1 physically mapped ROM support")
> Link: https://lore.kernel.org/lkml/202011021254.XC70BaQT-lkp@intel.com/
> Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
> Reported-by: kernel test robot <lkp@intel.com>
> Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
> Cc: Pavel Parkhomenko <Pavel.Parkhomenko@baikalelectronics.ru>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git nand/next, thanks.

Miquel

[PATCH v2] mtd: physmap: physmap-bt1-rom: Fix __iomem addrspace removal warning

[Serge Semin]

sparse is unhappy with us casting the __iomem address space pointer to
a type with no address space attribute specified:

"sparse warnings: (new ones prefixed by >>)"
>> drivers/mtd/maps/physmap-bt1-rom.c:78:18: sparse: sparse: cast removes address space '__iomem' of expression

Indeed we perform the __iomem-less type casting but to an integer
variable. The integer variable isn't dereferenced then, so the casting is
safe and won't cause any problem. But in order to make sparse happy and
keep the code coherent let's fix the warning by converting the local
"shift" and "chunk" variables to the "unsigned int" type (since their
value won't ever exceed three) and cast the __iomem-pointers to uintptr_t.
Add the same fix to the bt1_rom_map_read() method for unification.

Fixes: b3e79e7682e0 ("mtd: physmap: Add Baikal-T1 physically mapped ROM support")
Link: https://lore.kernel.org/lkml/202011021254.XC70BaQT-lkp@intel.com/
Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
Cc: Pavel Parkhomenko <Pavel.Parkhomenko@baikalelectronics.ru>

---

Link: https://lore.kernel.org/lkml/20201120113929.0aff2f32@canb.auug.org.au/
Changelog v2:
- Cast the pointers to uintptr_t instead of unsiged int so not to
  cause the "cast from pointer to integer of different size" warning on
  64-bits platforms.
---
 drivers/mtd/maps/physmap-bt1-rom.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/mtd/maps/physmap-bt1-rom.c b/drivers/mtd/maps/physmap-bt1-rom.c
index 27cfe1c63a2e..a35450002284 100644
--- a/drivers/mtd/maps/physmap-bt1-rom.c
+++ b/drivers/mtd/maps/physmap-bt1-rom.c
@@ -31,12 +31,12 @@ static map_word __xipram bt1_rom_map_read(struct map_info *map,
 					  unsigned long ofs)
 {
 	void __iomem *src = map->virt + ofs;
-	unsigned long shift;
+	unsigned int shift;
 	map_word ret;
 	u32 data;
 
 	/* Read data within offset dword. */
-	shift = (unsigned long)src & 0x3;
+	shift = (uintptr_t)src & 0x3;
 	data = readl_relaxed(src - shift);
 	if (!shift) {
 		ret.x[0] = data;
@@ -60,7 +60,7 @@ static void __xipram bt1_rom_map_copy_from(struct map_info *map,
 					   ssize_t len)
 {
 	void __iomem *src = map->virt + from;
-	ssize_t shift, chunk;
+	unsigned int shift, chunk;
 	u32 data;
 
 	if (len <= 0 || from >= map->size)
@@ -75,7 +75,7 @@ static void __xipram bt1_rom_map_copy_from(struct map_info *map,
 	 * up into the next three stages: unaligned head, aligned body,
 	 * unaligned tail.
 	 */
-	shift = (ssize_t)src & 0x3;
+	shift = (uintptr_t)src & 0x3;
 	if (shift) {
 		chunk = min_t(ssize_t, 4 - shift, len);
 		data = readl_relaxed(src - shift);
-- 
2.29.2

Re: [PATCH v2] mtd: physmap: physmap-bt1-rom: Fix __iomem addrspace removal warning

[Miquel Raynal]

On Wed, 2020-11-25 at 07:26:40 UTC, Serge Semin wrote:
> sparse is unhappy with us casting the __iomem address space pointer to
> a type with no address space attribute specified:
> 
> "sparse warnings: (new ones prefixed by >>)"
> >> drivers/mtd/maps/physmap-bt1-rom.c:78:18: sparse: sparse: cast removes address space '__iomem' of expression
> 
> Indeed we perform the __iomem-less type casting but to an integer
> variable. The integer variable isn't dereferenced then, so the casting is
> safe and won't cause any problem. But in order to make sparse happy and
> keep the code coherent let's fix the warning by converting the local
> "shift" and "chunk" variables to the "unsigned int" type (since their
> value won't ever exceed three) and cast the __iomem-pointers to uintptr_t.
> Add the same fix to the bt1_rom_map_read() method for unification.
> 
> Fixes: b3e79e7682e0 ("mtd: physmap: Add Baikal-T1 physically mapped ROM support")
> Link: https://lore.kernel.org/lkml/202011021254.XC70BaQT-lkp@intel.com/
> Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
> Reported-by: kernel test robot <lkp@intel.com>
> Cc: Alexey Malahov <Alexey.Malahov@baikalelectronics.ru>
> Cc: Pavel Parkhomenko <Pavel.Parkhomenko@baikalelectronics.ru>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/next, thanks.

Miquel