RE: [PATCH v3] Fix endian issues and remove the board specific codes
From: Hn Chen <hidden>
Date: 2015-05-25 13:23:26
Also in:
lkml
Hi, Dmitry, Just send the v4 and please feel free to prompt me a message if there is anything wrong. Remove the fw_id checking redundant codes in this version since it has no meaning. I just want to make sure that values are all correct originally. Best Regard, Hn.chen -----Original Message----- From: Dmitry Torokhov [mailto:dmitry.torokhov@gmail.com] Sent: Saturday, May 23, 2015 7:37 AM To: Hn Chen Cc: linux-input@vger.kernel.org; linux-kernel@vger.kernel.org Subject: Re: [PATCH v3] Fix endian issues and remove the board specific codes Hi Hn, On Thu, May 21, 2015 at 03:43:08PM +0800, Hn Chen wrote:
Hi, Dmitry, Thanks for your comments. Here are my replies for some comments.quoted
+ fw_id = ((fw_chunk_info.chunk_info.version_number >> 12) & 0xF); + chip_id = (((dev_wdt87xx->sys_param.fw_id) >> 12) & 0xF); + + if (fw_id != chip_id) { + wdt87xx_get_sysparam(client, dev_wdt87xx);quoted
Why would it be different now?This driver is intend to support two controllers, WDT8753 & WDT8756. The chip_id and the the firmware are different for them, so we have to check if The firmware is suitable for this controller.
What I meant is as far as I understand dev_wdt87xx->sys_param.fw_id is what we've read from the controller when we probed the device. What is the point in trying to re-read it with wdt87xx_get_sysparam()? Do we expect it to somehow change?
quoted
+ + return (unsigned short) y;quoted
I think I might have asked this before, by anyway: are we sure that we don't already have this checksum implementation in lib/? And if we don't maybe we should?The MISR is a checksum function which the full name is "Multiple Input Signature Register". I checked the kernel library and not found any algorithm like this. You mean I have to submit the checksum function into kernel/lib ?
No, you do not have to. I just wanted to make sure we are not duplicating something that already exists.
quoted
Can you tell me a bit more about the way device sends touch data? Does it always send all 10 contacts regardless of data in raw_buf[TOUCH_PK_OFFSET_FNGR_NUM]?Yes, it always send all 10 contacts data.
I see. -- Dmitry