On 07/30/2020 09:20 PM, Mathieu Poirier wrote:
On Wed, Jul 22, 2020 at 06:20:34PM +0100, Suzuki K Poulose wrote:
Convert all register accesses from etm4x driver to use a wrapper to allow switching the access at runtime with little overhead.
co-developed by sed tool ;-), mostly equivalent to :
s/readl(_relaxed)?(drvdata->base + (.*))/etm4x_\1_read32(csdev, \2) s/writel(_relaxed)?((.*), drvdata->base + (.*))/etm4x_\1_write32(csdev, \2, \3)
We don't want to replace them with the csdev_access_* to avoid a function call for every register access for system register access.
Cc: Mathieu Poirier mathieu.poirier@linaro.org Cc: Mike Leach mike.leach@linaro.org Signed-off-by: Suzuki K Poulose suzuki.poulose@arm.com
.../coresight/coresight-etm4x-sysfs.c | 9 +- drivers/hwtracing/coresight/coresight-etm4x.c | 334 +++++++++--------- drivers/hwtracing/coresight/coresight-etm4x.h | 24 ++ 3 files changed, 189 insertions(+), 178 deletions(-)
Agreed to all the comments.
diff --git a/drivers/hwtracing/coresight/coresight-etm4x.h b/drivers/hwtracing/coresight/coresight-etm4x.h index b8283e1d6d88..2b51d03ab6d7 100644 --- a/drivers/hwtracing/coresight/coresight-etm4x.h +++ b/drivers/hwtracing/coresight/coresight-etm4x.h @@ -120,6 +120,30 @@ #define TRCCIDR2 0xFF8 #define TRCCIDR3 0xFFC +#define etm4x_relaxed_read32(csa, offset) \
- readl_relaxed((csa)->base + (offset))
+#define etm4x_read32(csa, offset) \
- readl((csa)->base + (offset))
+#define etm4x_relaxed_write32(csa, val, offset) \
- writel_relaxed((val), (csa)->base + (offset))
+#define etm4x_write32(csa, val, offset) \
- writel((val), (csa)->base + (offset))
+#define etm4x_relaxed_read64(csa, offset) \
- readq_relaxed((csa)->base + (offset))
+#define etm4x_read64(csa, offset) \
- readq((csa)->base + (offset))
+#define etm4x_relaxed_write64(csa, val, offset) \
- writeq_relaxed((val), (csa)->base + (offset))
+#define etm4x_write64(csa, val, offset) \
- writeq((val), (csa)->base + (offset))
Since I haven't gone through the rest of the patches I'll assume you want to enhance the above to pick an access type at some point in the future.
Yes, they are plumbed in with the introduction of system instruction support. We don't want to the overhead of a function call for each register access. Also, we would like to avoid jumping through the large switch..cases for a compile time constant offset. Hence this macro.
Cheers Suzuki