[PATCH v2 02/19] ARM: at91/at91x40: remove use of at91_sys_read/write

On 14:59 Thu 23 Feb     , Ryan Mallon wrote:

On 23/02/12 14:25, Jean-Christophe PLAGNIOL-VILLARD wrote:

quoted

On 09:22 Thu 23 Feb     , Ryan Mallon wrote:

quoted

On 22/02/12 20:39, Nicolas Ferre wrote:

quoted

From: Jean-Christophe PLAGNIOL-VILLARD <redacted>

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <redacted>
Acked-by: Nicolas Ferre <redacted>
---
 arch/arm/mach-at91/at91x40.c              |    2 +-
 arch/arm/mach-at91/at91x40_time.c         |   28 +++++++++++++++++-----------
 arch/arm/mach-at91/include/mach/at91x40.h |   18 +++++++++---------
 3 files changed, 27 insertions(+), 21 deletions(-)

diff --git a/arch/arm/mach-at91/at91x40.c b/arch/arm/mach-at91/at91x40.c
index 0154b7f..5400a1d 100644
--- a/arch/arm/mach-at91/at91x40.c
+++ b/arch/arm/mach-at91/at91x40.c

@@ -44,7 +44,7 @@ static void at91x40_idle(void)
 	 * Disable the processor clock.  The processor will be automatically
 	 * re-enabled by an interrupt or by a reset.
 	 */
-	at91_sys_write(AT91_PS_CR, AT91_PS_CR_CPU);
+	__raw_writel(AT91_PS_CR_CPU, AT91_PS_CR);

This doesn't seem to be equivalent, at91_sys_write does:

  void __iomem *addr = (void __iomem *)AT91_VA_BASE_SYS;
  __raw_writel(value, addr + reg_offset);

and this patch doesn't redefine AT91_PS_CR. Was it broken before this
patch? What am I missing?

this is right
#define AT91_PS_CR      (AT91_PS + 0)   /* PS Control register */

That doesn't answer my question.

The old, at91_sys_write, version was writing to (using __raw_writel):

  AT91_VA_BASE_SYS + AT91_PS_CR

The new version is writing, also using __raw_writel, to:

  AT91_PS_CR

The value of AT91_PS_CR is not changed in this patch. Assuming that
AT91_VA_BASE_SYS for at91x40 (which at a quick glance it is not), then
the old and the new version of the code are not writing to the same
address.

Was it previously incorrect, or is it incorrect now?

it's as we update

+#define AT91_PS		0xffff4000	/* Power Save */

Best Regards,
J.