From mboxrd@z Thu Jan 1 00:00:00 1970 From: m.szyprowski@samsung.com (Marek Szyprowski) Date: Tue, 12 Oct 2010 16:15:11 +0200 Subject: [PATCH 0/2 v2] S3C64XX: Fix 48M clock enable In-Reply-To: References: <20101012123945.16469.2779.stgit@localhost.localdomain> <005701cb6a0c$be774370$3b65ca50$%szyprowski@samsung.com> Message-ID: <005d01cb6a17$e53307f0$af9917d0$%szyprowski@samsung.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hello, On Tuesday, October 12, 2010 3:49 PM Paulius Zaleckas wrote: > >> To enable 48M clock USB PHY must be initialized since 48M > >> is output of USB PHY PLL. > >> So fix a problem where other devices requiring 48M clock will > >> not work unless USB OTG driver is enabled. > >> is_osc was removed from OTG drivers platform data and implemented > >> as separate function s3c6400_clk_xusbxti_is_osc which should be > >> called from machine_init. Currently there is no in-kernel users > >> of it. > >> > >> This fix is only for S3C64xx family, but I am sure something > >> similar will be needed for S5P family. I have started the work > >> but I am unable to finnish it since I only have datasheet for > >> S5PC100. Maybe Samsung guys could help me with this..? > >> However current patches doesn't break S5P since it is not > >> using s3c-hsotg driver (yet?). > > > > There are patches that enables the driver on S5PC110/S5PV210 series > > on kgene/for-next tree. Please note that at least Samsung Goni and > > Samsung Aquila uses external oscillator to supply usb clock. Please > > refer to: > > > > http://git.kernel.org/?p=linux/kernel/git/kgene/linux- > samsung.git;a=commit;h=e5e044e637228be8bdf033c0772f9e5926269ae0 > > Interesting patch... If Goni uses oscillator then it will fail with current > otg driver since it will configure phy for crystal, because is_otg is not set > in the platform data. Right, but the interesting is that fact that it was working. I will check it tomorrow again. > > http://git.kernel.org/?p=linux/kernel/git/kgene/linux- > samsung.git;a=commit;h=a33bc34103b4fc4eca13cdd61f2579857797bbf7 > > This patch contains some really nasty workaround for S3C64XX_OTHERS > Mine version allows to get rid of it. Right that approach was a hack. I really like the idea behind your patch. > I have some preliminary version of patch for s5p. I will send it to you in > the evening. Ok, I will check it tomorrow. Best regards -- Marek Szyprowski Samsung Poland R&D Center