Thread (9 messages) 9 messages, 3 authors, 2013-05-29

Re: [PATCH] HID: hid-sensor-hub: do not process feature reports in raw_event

From: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Date: 2013-03-25 15:41:49
Also in: lkml

Daniel,

I am looking at 3.9.rc1.
The only place I see the raw_event callback is called from 
hid/hid_input_report(). hid_input_report is called with type 
HID_INPUT_REPORT in all cases, except hid_ctrl(), where it can be 
different depending on xx.report->type. But here, the return value is 
not checked.

Do you know the call chain for HID_FETURE_REPORT, where this is creating 
problem?

Thanks,
Srinivas

On 03/21/2013 10:22 AM, daniel.leung@linux.intel.com wrote:
quoted hunk ↗ jump to hunk
From: Daniel Leung <redacted>

In sensor_hub_raw_event(), HID feature reports are ignored but are
still marked as processed. This causes the in-kernel struct not to be
updated. Any non-updated fields in the feature reports are zero, and
they are being sent to the device. This causes confusion in the
sensor hub firmware, and some sensors are not powered up as a result.

This changes the raw_event rountine to only process input reports,
and let the hid core handle the incoming feature reports.

The issue has been observed on Acer Iconia W700.

Signed-off-by: Daniel Leung <redacted>
---
  drivers/hid/hid-sensor-hub.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/hid/hid-sensor-hub.c b/drivers/hid/hid-sensor-hub.c
index ca749810..3f7df68 100644
--- a/drivers/hid/hid-sensor-hub.c
+++ b/drivers/hid/hid-sensor-hub.c
@@ -413,7 +413,7 @@ static int sensor_hub_raw_event(struct hid_device *hdev,
  			 report->id, size, report->type);
  	hid_dbg(hdev, "maxfield:%d\n", report->maxfield);
  	if (report->type != HID_INPUT_REPORT)
-		return 1;
+		return 0;
  
  	ptr = raw_data;
  	ptr++; /*Skip report id*/
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help