Hi Shimoda-san, On Mon, Feb 1, 2016 at 12:06 PM, Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx> wrote: >> From: geert.uytterhoeven@xxxxxxxxx [mailto:geert.uytterhoeven@xxxxxxxxx] On Behalf Of Geert Uytterhoeven >> > --- a/drivers/clk/shmobile/r8a7795-cpg-mssr.c >> > +++ b/drivers/clk/shmobile/r8a7795-cpg-mssr.c >> > @@ -124,6 +124,8 @@ static const struct mssr_mod_clk r8a7795_mod_clks[] __initconst = { >> > DEF_MOD("pcie0", 319, R8A7795_CLK_S3D1), >> > DEF_MOD("usb3-if1", 327, R8A7795_CLK_S3D1), >> > DEF_MOD("usb3-if0", 328, R8A7795_CLK_S3D1), >> > + DEF_MOD("usb-dmac0", 330, R8A7795_CLK_S3D2), >> > + DEF_MOD("usb-dmac1", 331, R8A7795_CLK_S3D2), >> >> The datasheet is not very clear about whether the module clock is based of s3d1 >> (266 MHz) or s3d2 (133 MHz), but given the comment "... it is recommended to >> allow several tens of clocks if operating at 260 MHz ..." in section 75.2.2 >> "DMA0/1 Software Reset Register" of the datasheet, I'm inclined to believe s3d1 >> would be correct. > > According to the HW team (in case of gen2 though): > > In gen2 (In gen3 I assumed) > ZS (S3D1) : For using DRAM access > HP (S3D2) : For using register access > 48MHz (50MHz) : For using USB transfer > > So, I also think that s3d1 would be correct. > What do you think? Thanks, that makes two of us :-) Hence please use s3d1. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds