Hi Luis, On 2/12/24 18:44, Luis de Arquer wrote: > Hi Cristian, > > On Mon, 2024-02-05 at 13:24 +0200, Cristian Ciocaltea wrote: >> + >> +static bool hdptx_phy_clk_pll_calc(unsigned int data_rate, >> + struct ropll_config *cfg) >> +{ >> + const unsigned int fout = data_rate / 2, fref = 24000; >> + unsigned long k = 0, lc, k_sub, lc_sub; >> + unsigned int fvco, sdc; >> + u32 mdiv, sdiv, n = 8; >> + >> + for (sdiv = 16; sdiv >= 1; sdiv--) { >> + if (sdiv % 2 && sdiv != 1) >> + continue; >> + >> + fvco = fout * sdiv; >> + >> + if (fvco < 2000000 || fvco > 4000000) >> + continue; >> + > > What about adding a check to data_rate, maybe like > > if (fout > 0x0FFFFFFF) > return false; > > or similar, before the for loop, to keep the multiplication safe? > > Right now it would be redundant, given that data_rate was, at some > point, encoded in 28 bits within bus_width. But can prevent future pain, > especially after changing to phy_configure_opts_hdmi. Indeed, it makes sense to do the check, thanks for the heads up! I was almost ready to submit v3, so this arrived just in time. :-) Regards, Cristian