[PATCH 2/3 RFC] ath10k: wmi: match wait_for_completion_timeout return type

Return type of wait_for_completion_timeout is unsigned long not int.
An appropriately named unsigned long is added and the assignments fixed up.
Rather than returning 0 (timeout) or a more or less random remaining time
(completion success) this return 0 or 1 which also resolves the type of the
functions being int.

Signed-off-by: Nicholas Mc Guire <redacted>
---

Checking the call-sites of ath10k_wmi_wait_for_unified_ready and 
ath10k_wmi_wait_for_service_ready the positive return value (remaining
time in jiffies) is never passed up the call-chain nor used so it is 
cleaner to treat this like a boolean success/fail only (actually the two
functions should probably be of type bool - but that does not seem to be
common practice in the ath10k code base)

Patch was only compile tested with x86_64_defconfig + CONFIG_ATH_CARDS=m,
CONFIG_ATH10K=m

Patch is against 4.0-rc3 (localversion-next is -next-20150312)
:
 drivers/net/wireless/ath/ath10k/wmi.c |   20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/wmi.c b/drivers/net/wireless/ath/ath10k/wmi.c
index c7ea77e..a1cdcba 100644
--- a/drivers/net/wireless/ath/ath10k/wmi.c
+++ b/drivers/net/wireless/ath/ath10k/wmi.c

@@ -884,20 +884,24 @@ void ath10k_wmi_put_wmi_channel(struct wmi_channel *ch,
 
 int ath10k_wmi_wait_for_service_ready(struct ath10k *ar)
 {
-	int ret;
+	unsigned long time_left;
 
-	ret = wait_for_completion_timeout(&ar->wmi.service_ready,
-					  WMI_SERVICE_READY_TIMEOUT_HZ);
-	return ret;
+	time_left = wait_for_completion_timeout(&ar->wmi.service_ready,
+						WMI_SERVICE_READY_TIMEOUT_HZ);
+	if(!time_left)
+		return 0;
+	return 1;
 }
 
 int ath10k_wmi_wait_for_unified_ready(struct ath10k *ar)
 {
-	int ret;
+	unsigned long time_left;
 
-	ret = wait_for_completion_timeout(&ar->wmi.unified_ready,
-					  WMI_UNIFIED_READY_TIMEOUT_HZ);
-	return ret;
+	time_left = wait_for_completion_timeout(&ar->wmi.unified_ready,
+						WMI_UNIFIED_READY_TIMEOUT_HZ);
+	if(!time_left)
+		return 0;
+	return 1;
 }
 
 struct sk_buff *ath10k_wmi_alloc_skb(struct ath10k *ar, u32 len)

-- 
1.7.10.4

On 12 March 2015 at 16:49, Nicholas Mc Guire [off-list ref] wrote:

Return type of wait_for_completion_timeout is unsigned long not int.
An appropriately named unsigned long is added and the assignments fixed up.
Rather than returning 0 (timeout) or a more or less random remaining time
(completion success) this return 0 or 1 which also resolves the type of the
functions being int.

Signed-off-by: Nicholas Mc Guire <redacted>
---

Checking the call-sites of ath10k_wmi_wait_for_unified_ready and
ath10k_wmi_wait_for_service_ready the positive return value (remaining
time in jiffies) is never passed up the call-chain nor used so it is
cleaner to treat this like a boolean success/fail only (actually the two
functions should probably be of type bool - but that does not seem to be
common practice in the ath10k code base)

It'd make sense to have these functions return 0 or -ETIMEDOUT. In
that case both call sites would need to be adjusted to treat "< 0" or
"!x" as an error (instead of the current "<= 0") condition and not set
-ETIMEDOUT themselves.


Michał

On Fri, 13 Mar 2015, Michal Kazior wrote:

On 12 March 2015 at 16:49, Nicholas Mc Guire [off-list ref] wrote:

quoted

Return type of wait_for_completion_timeout is unsigned long not int.
An appropriately named unsigned long is added and the assignments fixed up.
Rather than returning 0 (timeout) or a more or less random remaining time
(completion success) this return 0 or 1 which also resolves the type of the
functions being int.

Signed-off-by: Nicholas Mc Guire <redacted>
---

Checking the call-sites of ath10k_wmi_wait_for_unified_ready and
ath10k_wmi_wait_for_service_ready the positive return value (remaining
time in jiffies) is never passed up the call-chain nor used so it is
cleaner to treat this like a boolean success/fail only (actually the two
functions should probably be of type bool - but that does not seem to be
common practice in the ath10k code base)

It'd make sense to have these functions return 0 or -ETIMEDOUT. In
that case both call sites would need to be adjusted to treat "< 0" or
"!x" as an error (instead of the current "<= 0") condition and not set
-ETIMEDOUT themselves.

looking at the call sites in ath10k_core_start more or less 
all other initialization calls will treate 0 as success and
!=0 as failure so this is the cleaner solution. as its all
now

status = call()
if(status)
 	error

patch just posted.

thx!
hofrat