On Sep 22, 2010, at 1:18 PM, Ping Cheng <pingli...@gmail.com> wrote:
> Overall, the whole patchset is very clean. I like it. Since a few > places affected the protocol 5 devices, I have to test it before I can > let it in. I am tied to another project and Peter is on vacation so > the final review may take sometime. I do have one comment below for > this patch. > > Ping > > On Mon, Sep 6, 2010 at 5:42 PM, <ch...@cnpbagwell.com> wrote: >> From: Chris Bagwell <ch...@cnpbagwell.com> >> >> The check to find event sync windows with to little information >> was to agressive. For example, if only pressure is changing >> then only two events will be sent: ABS_PRESSURE and EV_SYN. >> >> The previous logic would discard in this case. This is also >> in preparation for simplier tools which do not report serial >> numbers and in general will send small total events per >> sync window. >> >> Also, fixed comment since MSC_SERIAL events are not strictly >> required. >> >> Signed-off-by: Chris Bagwell <ch...@cnpbagwell.com> >> --- >> src/wcmUSB.c | 7 ++++--- >> 1 files changed, 4 insertions(+), 3 deletions(-) >> >> diff --git a/src/wcmUSB.c b/src/wcmUSB.c >> index 4b1806f..72adb0b 100644 >> --- a/src/wcmUSB.c >> +++ b/src/wcmUSB.c >> @@ -814,9 +814,10 @@ static void usbParseEvent(InputInfoPtr pInfo, >> wcmUSBData* private = common->private; >> >> DBG(10, common, "\n"); >> + >> /* store events until we receive the MSC_SERIAL containing >> - * the serial number; without it we cannot determine the >> - * correct channel. */ >> + * the serial number or a SYN_REPORT. >> + */ >> >> /* space left? bail if not. */ >> if (private->wcmEventCnt >= >> @@ -871,7 +872,7 @@ static void usbParseEvent(InputInfoPtr pInfo, >> } >> >> /* ignore events without information */ >> - if ((private->wcmEventCnt <= 2) && private->wcmLastToolSerial) >> + if ((private->wcmEventCnt < 2) && private->wcmLastToolSerial) > > Does this change make a difference to your Bamboo tablet, with both > your new kernel driver and the old linuxwacom kernel driver? I have to > consider backward compatibility seriously since it is my job. Yes for both cases. You can easily see in x logs that it's discarding valid data. It's mostly only a concern with low volume events like button presses. So that means newer kernel drivers. > > This code was mainly for protocol 5 devices to deal with a few special > cases, such as both stylus and pad (tablet buttons) are in use; or > moving the tablet while touching the tablet buttons. Somehow we get > invalid data. I'll look at protocol 5 kernel drivers and see if I can think of a reason that invalid data could happen. Worst case, we can make that check specific to protocol 5 devices. > > Ping ------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ Linuxwacom-devel mailing list Linuxwacom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel
