From mboxrd@z Thu Jan 1 00:00:00 1970 From: hdegoede@redhat.com (Hans de Goede) Date: Wed, 24 Jun 2015 09:16:54 +0200 Subject: [PATCH] touchscreen: sun4i-ts: Really fix A10 temperature reporting In-Reply-To: <20150623195424.GE40168@dtor-ws> References: <1425893870-28715-1-git-send-email-hdegoede@redhat.com> <41a7a857-96fd-48ab-9778-127cfc091ef0@googlegroups.com> <5589B78C.4080600@redhat.com> <20150623195424.GE40168@dtor-ws> Message-ID: <558A5966.9010908@redhat.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi, On 23-06-15 21:54, Dmitry Torokhov wrote: > On Tue, Jun 23, 2015 at 09:46:20PM +0200, Hans de Goede wrote: >> Hi, >> >> On 06/23/2015 05:06 PM, m.silentcreek at gmail.com wrote: >>> Hi Hans, >>> >>> is it possible that this patch (and the parent commit "Input: sun4i-ts - allow controlling filter and sensitivity via DT") has quite an adversial effect on other boards than the ones tested? >>> >>> On my Lemaker Bananapi the temperature is now being reported as somwhere between -15?C to -10?C with the latest Kernel 4.1.0. These negative values are obviously way off. On Kernel 4.0.5 the temperature reading was usually around 40?C. While that might not have been accurate either, it was at least plausible. >> >> The temperature curve for the A20 SoC was not changed, if it did change then >> you're using an old dtb file with a new kernel, you must always update your >> dtb file together with the kernel. > > Umm, we should keep compatibility with old DTS... Right, unfortunately that is not always possible, and with sunxi boards the dtb is always shipped together with the kernel, and the dtb as ABI rule really is intended for the case where the dtb ships with the board/firmware. That does not mean that we can go and break it every other day, but the problem in this case is that we shipped a dts for all boards with a compatible of allwinner,sun4i-a10-ts, and then later on we found out that the temperature curve for the build in temp sensor is different for the A10 compared to the A13/A20, so we adjusted the kernel driver to have 2 temperature curves, 1 for the A10 and one for the later models, and adjusted the compatible in the dts for later models to allwinner,sun5i-a13-ts, note that the actual touchscreen functionality is unchanged and works fine with both compatibles, the only thing which is happening now is that Timo is seeing a different temperature because he is using the new A10 temperature curve on his A20, other then that everything still works 100%, and there really was no other way to deal with this. Let me also reply to one of Timo's later mails here, as that explains the real problem: On 23-06-15 23:06, m.silentcreek at gmail.com wrote:> Hi again, > > Am Dienstag, 23. Juni 2015 21:46:25 UTC+2 schrieb Hans de Goede: >>> On my Lemaker Bananapi the temperature is now being reported as somwhere between -15?C to -10?C with the latest Kernel 4.1.0. These negative values are obviously way off. On Kernel 4.0.5 the temperature reading was usually around 40?C. While that might not have been accurate either, it was at least plausible. >> >> The temperature curve for the A20 SoC was not changed, if it did change then >> you're using an old dtb file with a new kernel, you must always update your >> dtb file together with the kernel. > > I just checked the diff for Linux 4.0 and 4.1 and noticed that sun7i-a20.dtsi in fact has not changed. Is it possible that something that should have been merged got delayed? > > I looked in Maxime's tree "dt-for-4.2" and found this patch that looks like it could be related: https://git.kernel.org/cgit/linux/kernel/git/mripard/linux.git/commit/?h=sunxi/dt-for-4.2&id=8bf1b9b3d90194a174493febc731f7783f2adf1a Ah yes, something went wrong during the 4.1 cycle wrt merging ARM patches which have caused almost all ARM patches intended for 4.1 to get delayed to 4.2, including that one, that is the problem. We did our best to update the driver and the dts in lockstep, and you should have never seen this problem, but the missing of the merge window for all ARM patches during the 4.1 cycle has caused our carefully planned lock-step update to fail, sorry about that. Maxime I think we should send the mentioned dts patch to stable for 4.1 , do you agree ? Regards, Hans