diff options
| author | Arnd Bergmann <arnd@arndb.de> | 2016-01-13 22:38:08 +0100 |
|---|---|---|
| committer | Brian Norris <computersforpeace@gmail.com> | 2016-01-15 10:02:09 -0800 |
| commit | f9bdbd6c46c8ce0bb95f5b708a4a4a4b6b9a5917 (patch) | |
| tree | d8a2468922595fc3437cad0aa41dc33f84170834 | |
| parent | 92752d9974882e2e5384e92668f02a134f9c7463 (diff) | |
| download | linux-f9bdbd6c46c8ce0bb95f5b708a4a4a4b6b9a5917.tar.bz2 | |
mtd: nuc900_nand: read correct SMISR register
The nuc900_nand driver has always passed an incorrect register
address in its nuc900_check_rb() function, which cannot possibly
work, and in some configurations gives us a build warning:
drivers/mtd/nand/nuc900_nand.c: In function 'nuc900_check_rb':
drivers/mtd/nand/nuc900_nand.c:27:23: warning: passing argument 1 of '__raw_readl' makes pointer from integer without a cast [-Wint-conversion]
#define REG_SMISR 0xac
drivers/mtd/nand/nuc900_nand.c:118:20: note: in expansion of macro 'REG_SMISR'
val = __raw_readl(REG_SMISR);
This makes sure we actually read from the register rather than
from (void *)0x000000ac in user space.
I suspect nobody noticed this before because the nuc900_nand_devready()
function never gets called, or nobody uses this driver on an upstream
kernel. Possibly even both.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
| -rw-r--r-- | drivers/mtd/nand/nuc900_nand.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/drivers/mtd/nand/nuc900_nand.c b/drivers/mtd/nand/nuc900_nand.c index 220ddfcf29f5..dbc5b571c2bb 100644 --- a/drivers/mtd/nand/nuc900_nand.c +++ b/drivers/mtd/nand/nuc900_nand.c @@ -113,7 +113,7 @@ static int nuc900_check_rb(struct nuc900_nand *nand) { unsigned int val; spin_lock(&nand->lock); - val = __raw_readl(REG_SMISR); + val = __raw_readl(nand->reg + REG_SMISR); val &= READYBUSY; spin_unlock(&nand->lock); |