Best Regards. Anson huang 黄勇才 Freescale Semiconductor Shanghai 上海浦东新区亮景路192号A座2楼 201203 Tel:021-28937058 >-----Original Message----- >From: Shawn Guo [mailto:shawn.guo@xxxxxxxxxx] >Sent: Thursday, December 19, 2013 3:29 PM >To: Huang Yongcai-B20788 >Cc: kernel@xxxxxxxxxxxxxx; rui.zhang@xxxxxxxxx; eduardo.valentin@xxxxxx; >devicetree@xxxxxxxxxxxxxxx; linux-doc@xxxxxxxxxxxxxxx; linux-arm- >kernel@xxxxxxxxxxxxxxxxxxx; linux-pm@xxxxxxxxxxxxxxx >Subject: Re: [PATCH 2/3] thermal: imx: add necessary clk operation > >On Thu, Dec 19, 2013 at 07:08:16AM +0000, Anson.Huang@xxxxxxxxxxxxx wrote: >> >On Thu, Dec 19, 2013 at 01:17:24PM -0500, Anson Huang wrote: >> >> @@ -427,6 +429,13 @@ static int imx_thermal_probe(struct >> >> platform_device >> >*pdev) >> >> return ret; >> >> } >> >> >> >> + data->thermal_clk = devm_clk_get(&pdev->dev, NULL); >> >> + if (IS_ERR(data->thermal_clk)) { >> >> + ret = IS_ERR(data->thermal_clk); >> >> + dev_err(&pdev->dev, "failed to get thermal clk!\n"); >> >> + return ret; >> >> + } >> >> + >> > >> >So when the new kernel runs on a board with an old DTB installed, >> >thermal driver will be broken. >> > >> Yes, I thought about this case, but the previous implement is >> incorrect, if the PLL3 is not enabled by other drivers, thermal driver >> will not work, so this patch is a bug fix, not enhancement. So we >> still need to consider old dts case? > >The thing is mainline kernel runs on many board with thermal driver being >functional today. That said, PLL3 is already enabled on these platforms when >thermal driver is running. You cannot fix a bug but meanwhile break these >users who use old DTB. Okay, got it, I will replace the return with one warning message in V2. Anson > >Shawn ��.n��������+%������w��{.n����z�{��ܨ}���Ơz�j:+v�����w����ޙ��&�)ߡ�a����z�ޗ���ݢj��w�f