On Friday, February 1, 2019 8:36:13 AM CET Rosen Penev wrote: > The flash chip on the board (Spansion S25FL256SAIF00) is rated to support > at least 50MHz according to the datasheet. > > From testing this, that seems correct. > > time dd if=/dev/mtdblock3 of=/dev/null bs=64k from > > 41.78s to 16.61s > > Signed-off-by: Rosen Penev <ros...@gmail.com>
Hm, this is fine. There's just a small caveat. The mt7621-spi [0] can do at most (SYS_CLOCK) / 2 which works out as 50 MHz / 2 = 25 MHz and this is supported by your results as well: Because 41.78s / 16.61s = 2.51. If the spi-transfers were operating at 50 Mhz, dd would have taken just around 8-9 seconds. The reason why the dts patch this is still fine is because the spi-max-frequency's value is supposed to be taken from the slave device (S25FL256SAIF00 datasheet). So maybe, you could just mention in the commit message that speed it sadly capped at 25 Mhz. (Maybe you could also test, if you get better read performance by adding the "m25p,fast-read;" property [1] to the device's dts) Regards, Christian [0] <https://github.com/torvalds/linux/blob/master/drivers/staging/mt7621-spi/spi-mt7621.c#L304> [1] <https://www.kernel.org/doc/Documentation/devicetree/bindings/mtd/jedec%2Cspi-nor.txt> > --- > target/linux/ramips/dts/GB-PC2.dts | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/target/linux/ramips/dts/GB-PC2.dts > b/target/linux/ramips/dts/GB-PC2.dts > index 788d4e6c08..f951083b8f 100644 > --- a/target/linux/ramips/dts/GB-PC2.dts > +++ b/target/linux/ramips/dts/GB-PC2.dts > @@ -81,7 +81,7 @@ > m25p80@0 { > compatible = "jedec,spi-nor"; > reg = <0>; > - spi-max-frequency = <10000000>; > + spi-max-frequency = <50000000>; > > partitions { > compatible = "fixed-partitions"; > _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
