Having further investigated the problems of armmp with Marvell I would like to thank Uwe for submitting a patch which resolves problem a).
Problem b) arises because code intended only for TI chips gets initialized on Marvell. When disabling in kconfigs "POWER_AVS, POWER_AVS_OMAP, CONFIG_POWER_AVS_OMAP_CLASS3" the red error of sr_init vanishes. My first guess would be that this code should be in a module that can be blacklisted for Marvell? In the meantime I have become aware that there is yet another major problem c): When the system is idle, one CPU is always busy at least 50% with a kworker thread which IMHO does not perform any useful work. When idle, the kernel 4.9.y built exclusively for Marvell has both CPU’s at below 1% load. This additional load is not acceptable and prevents use of the Debian mp-kernel on my Marvell :-( Unfortunately I do not know how to find out what part of code results in this waste... Somehow I start doubting that this kernel has ever been in productive use on Marvell??? LinAdmin On 25.09.2018 10:45, LinAdmin wrote: > Hi experts! > > I recently bought a NAS with the Marvell Armada 385 dual > core CPU and 2x 8TB disks. The performance of the hardware > is excellent but the software from the manufacturer is > ugly with a rather old kernel (3.10). > > I succeeded cross-building a recent 4.9.y kernel using > kernel config and dts from the manufacturers GPL source > with Debian Stretch armhf. All this works perfect. > > Since I prefer not to repeatedly build a kernel I tried > using the Debian multi platform kernel 4.9.0-8-armmp. It > does work, but I was not able to solve the following problems: > > a) In the kernel config I used, the buffer management > CONFIG_MVNETA_BM_ENABLE was set, while in > config-4.9.0-8-armmp it is not set. This seems to lower > network throughput by approx. 20%. > > Imho all Marvell 38x are capable of using this buffer > management and therefore it seems preferable to switch it > on by default? > > b) Dmesg always shows "sr_init: platform driver register > failed for SR". This comes from activated SmartReflex/OMAP > power management which lacks in all Marvell designs. It > seems not to have other consequences but I would prefer > not to see any errors in dmesg. > > Many thanks for your help. > > LinAdmin >