On Mon, 24 Sep 2018 18:36:36 +0200 Christophe Kerello <christophe.kerello@xxxxxx> wrote: > >> +static int stm32_fmc2_resume(struct device *dev) > >> +{ > >> + struct stm32_fmc2 *fmc2 = dev_get_drvdata(dev); > >> + int i, ret; > >> + > >> + pinctrl_pm_select_default_state(dev); > >> + > >> + ret = clk_prepare_enable(fmc2->clk); > >> + if (ret) { > >> + dev_err(dev, "can not enable the clock\n"); > >> + return ret; > >> + } > >> + > >> + stm32_fmc2_init(fmc2); > >> + stm32_fmc2_timings_init(fmc2); > >> + stm32_fmc2_setup(fmc2); > >> + > >> + for (i = 0; i < fmc2->ncs; i++) > >> + nand_reset(&fmc2->chip, i); > > > > This means you have one different NAND chip wired on each CS. > > > > We could have two CS wired to the same NAND chip. Calling nand_reset > > twice would be harmless but a lost of time. Actually, you have to call nand_reset() for each CS, otherwise not all dies are reset. ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/