| Message ID | 20170721203814.3363448-1-arnd@arndb.de |
|---|---|
| State | Accepted |
| Delegated to: | Boris Brezillon |
| Headers | show |
On Fri, 21 Jul 2017 22:38:06 +0200 Arnd Bergmann <arnd@arndb.de> wrote: > orion_nand_read_buf uses an inline assembly with the "ldrd" > instruction, which is only available from ARMv5 upwards. This > used to be fine, since all users have an ARMv5 or ARMv7 CPU, > but now we can also build a multiplatform kernel with ARMv4 > support enabled in addition to the "kirkwood" (mvebu) platform. > > This provides an alternative to call the readsl() function that > is supposed to have the same effect and is also optimized for > performance. > > I first posted a version of this patch back in 2014, and there > was some discussion about it then. This fixes the bugs identified > back then and should be a reasonable alternative for the rare > corner case. Looks good to me. Applied to nand/next. Let's see if anyone complains before the merge window. Thanks, Boris > > Link: https://patchwork.kernel.org/patch/4144791/ > Cc: Jingoo Han <jg1.han@samsung.com> > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > drivers/mtd/nand/orion_nand.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/drivers/mtd/nand/orion_nand.c b/drivers/mtd/nand/orion_nand.c > index 209170ed2b76..41cb7acfc044 100644 > --- a/drivers/mtd/nand/orion_nand.c > +++ b/drivers/mtd/nand/orion_nand.c > @@ -54,13 +54,16 @@ static void orion_nand_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) > { > struct nand_chip *chip = mtd_to_nand(mtd); > void __iomem *io_base = chip->IO_ADDR_R; > +#if __LINUX_ARM_ARCH__ >= 5 > uint64_t *buf64; > +#endif > int i = 0; > > while (len && (unsigned long)buf & 7) { > *buf++ = readb(io_base); > len--; > } > +#if __LINUX_ARM_ARCH__ >= 5 > buf64 = (uint64_t *)buf; > while (i < len/8) { > /* > @@ -74,6 +77,10 @@ static void orion_nand_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) > buf64[i++] = x; > } > i *= 8; > +#else > + readsl(io_base, buf, len/4); > + i = len / 4 * 4; > +#endif > while (i < len) > buf[i++] = readb(io_base); > }
diff --git a/drivers/mtd/nand/orion_nand.c b/drivers/mtd/nand/orion_nand.c index 209170ed2b76..41cb7acfc044 100644 --- a/drivers/mtd/nand/orion_nand.c +++ b/drivers/mtd/nand/orion_nand.c @@ -54,13 +54,16 @@ static void orion_nand_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) { struct nand_chip *chip = mtd_to_nand(mtd); void __iomem *io_base = chip->IO_ADDR_R; +#if __LINUX_ARM_ARCH__ >= 5 uint64_t *buf64; +#endif int i = 0; while (len && (unsigned long)buf & 7) { *buf++ = readb(io_base); len--; } +#if __LINUX_ARM_ARCH__ >= 5 buf64 = (uint64_t *)buf; while (i < len/8) { /* @@ -74,6 +77,10 @@ static void orion_nand_read_buf(struct mtd_info *mtd, uint8_t *buf, int len) buf64[i++] = x; } i *= 8; +#else + readsl(io_base, buf, len/4); + i = len / 4 * 4; +#endif while (i < len) buf[i++] = readb(io_base); }
orion_nand_read_buf uses an inline assembly with the "ldrd" instruction, which is only available from ARMv5 upwards. This used to be fine, since all users have an ARMv5 or ARMv7 CPU, but now we can also build a multiplatform kernel with ARMv4 support enabled in addition to the "kirkwood" (mvebu) platform. This provides an alternative to call the readsl() function that is supposed to have the same effect and is also optimized for performance. I first posted a version of this patch back in 2014, and there was some discussion about it then. This fixes the bugs identified back then and should be a reasonable alternative for the rare corner case. Link: https://patchwork.kernel.org/patch/4144791/ Cc: Jingoo Han <jg1.han@samsung.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- drivers/mtd/nand/orion_nand.c | 7 +++++++ 1 file changed, 7 insertions(+)