Messages in this thread | | | Date | Mon, 4 Feb 2013 10:38:44 +0100 | Subject | Re: [PATCH v2 0/9] Support of Nexio 42" and new default class for hid-multitouch | From | Benjamin Tissoires <> |
| |
Hi Henrik,
On Sun, Feb 3, 2013 at 2:07 PM, Henrik Rydberg <rydberg@euromail.se> wrote: > Hi Benjamin, > >> so, this is the v2 of the support of win7/8 devices. > > Looks like it is getting there, thanks.
Thanks for the review. However, before sending a new patch series, I'd like to have your answers to my comments as I mostly disagree on everything :)
Thanks again, Benjamin
> >> changes since v1: >> - removed the "optimization" patches, as the benefit was minimum >> - introduce a new callback "report" in hid-core that drivers can use to treat the >> report by having it entirely parsed >> - rely on this new hook to support Nexio 42" > > As noted in the patches comments, using raw_event() seems sufficient. > >> side notes: >> - I've tested removing the heavy call to kzalloc in hid_input_field. The results >> are disapointing -> the processing time remains the same. >> - I've also tested not to rely on .event hook in hid-multitouch but only on .report. >> Idem, I thought it would reduce the code of hid-multitouch and will enhance its >> processing time, but the results are a roughly same number of lines for hid-multitouch >> and the same processing time... :( >> - these 2 tests helped in cleaning the patch set from the last time. >> >> And again, finally, I've pass all the 40 regression tests of my db. \o/ > > Nice. :-) > > Thanks, > Henrik
| |