From: Danke Xie d.xie@sta.samsung.com
The current TWL 6030 IRQ handler assumes little endianness. This change makes it endian-neutral.
Signed-off-by: Danke Xie d.xie@sta.samsung.com Signed-off-by: Taras Kondratiuk taras.kondratiuk@linaro.org --- v2: fixed sparse warning v1: https://patchwork.kernel.org/patch/2974331/
drivers/mfd/twl6030-irq.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/mfd/twl6030-irq.c b/drivers/mfd/twl6030-irq.c index 517eda8..18a607e 100644 --- a/drivers/mfd/twl6030-irq.c +++ b/drivers/mfd/twl6030-irq.c @@ -176,8 +176,9 @@ static irqreturn_t twl6030_irq_thread(int irq, void *data) int i, ret; union { u8 bytes[4]; - u32 int_sts; + __le32 int_sts; } sts; + u32 int_sts; /* sts.int_sts converted to CPU endianness */ struct twl6030_irq *pdata = data;
/* read INT_STS_A, B and C in one shot using a burst read */ @@ -196,8 +197,9 @@ static irqreturn_t twl6030_irq_thread(int irq, void *data) if (sts.bytes[2] & 0x10) sts.bytes[2] |= 0x08;
- for (i = 0; sts.int_sts; sts.int_sts >>= 1, i++) - if (sts.int_sts & 0x1) { + int_sts = le32_to_cpu(sts.int_sts); + for (i = 0; int_sts; int_sts >>= 1, i++) + if (int_sts & 0x1) { int module_irq = irq_find_mapping(pdata->irq_domain, pdata->irq_mapping_tbl[i]);
On Mon, 2013-12-23 at 19:11 +0200, Taras Kondratiuk wrote:
From: Danke Xie d.xie@sta.samsung.com
The current TWL 6030 IRQ handler assumes little endianness. This change makes it endian-neutral.
[]
diff --git a/drivers/mfd/twl6030-irq.c b/drivers/mfd/twl6030-irq.c
[]
@@ -196,8 +197,9 @@ static irqreturn_t twl6030_irq_thread(int irq, void *data) if (sts.bytes[2] & 0x10) sts.bytes[2] |= 0x08;
- for (i = 0; sts.int_sts; sts.int_sts >>= 1, i++)
if (sts.int_sts & 0x1) {
- int_sts = le32_to_cpu(sts.int_sts);
- for (i = 0; int_sts; int_sts >>= 1, i++)
if (int_sts & 0x1) { int module_irq = irq_find_mapping(pdata->irq_domain, pdata->irq_mapping_tbl[i]);
instead of the
for (...) { if (foo & 1) {
maybe using ffs would be better/faster
while ((bit = ffs(int_sts))) {
etc...
On 23 December 2013 19:20, Joe Perches joe@perches.com wrote:
On Mon, 2013-12-23 at 19:11 +0200, Taras Kondratiuk wrote:
From: Danke Xie d.xie@sta.samsung.com
The current TWL 6030 IRQ handler assumes little endianness. This change makes it endian-neutral.
[]
diff --git a/drivers/mfd/twl6030-irq.c b/drivers/mfd/twl6030-irq.c
[]
@@ -196,8 +197,9 @@ static irqreturn_t twl6030_irq_thread(int irq, void *data) if (sts.bytes[2] & 0x10) sts.bytes[2] |= 0x08;
for (i = 0; sts.int_sts; sts.int_sts >>= 1, i++)
if (sts.int_sts & 0x1) {
int_sts = le32_to_cpu(sts.int_sts);
for (i = 0; int_sts; int_sts >>= 1, i++)
if (int_sts & 0x1) { int module_irq = irq_find_mapping(pdata->irq_domain, pdata->irq_mapping_tbl[i]);
instead of the
for (...) { if (foo & 1) {
maybe using ffs would be better/faster
while ((bit = ffs(int_sts))) {
etc...
Makes sense, but the aim of this patch is to fix endianness issue. This micro-optimisation can be done as a separate patch.
On Mon, 23 Dec 2013, Taras Kondratiuk wrote:
From: Danke Xie d.xie@sta.samsung.com
The current TWL 6030 IRQ handler assumes little endianness. This change makes it endian-neutral.
Signed-off-by: Danke Xie d.xie@sta.samsung.com Signed-off-by: Taras Kondratiuk taras.kondratiuk@linaro.org
v2: fixed sparse warning v1: https://patchwork.kernel.org/patch/2974331/
drivers/mfd/twl6030-irq.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
Applied, thanks.
linaro-kernel@lists.linaro.org