On Mon, Oct 28, 2019 at 12:37 PM Carlo Pisani <carlojpisani@xxxxxxxxx> wrote: > > > > > These resources are supplied to the PCI core, probably from DT. A > > > > complete dmesg log would show more. > > The rb532 does not have device tree support, so the answer is no. > I am using a common vanilla kernel without any extra patch. What's your .config file? The knowledge about the host bridge windows must be compiled in somewhere. > 00:02.0 Ethernet controller: VIA Technologies, Inc. VT6105/VT6106S > [Rhine-III] (rev 86) > Subsystem: AST Research Inc Device 086c > Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- > ParErr- Stepping+ SERR- FastB2B- DisINTx- > 00:05.0 Serial controller: Oxford Semiconductor Ltd OX16PCI954 (Quad > 16950 UART) function 0 (Uart) (rev 01) (prog-if 06 [16950]) > Subsystem: Oxford Semiconductor Ltd Device 0000 > Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- > ParErr- Stepping- SERR- FastB2B- DisINTx- > > If those UARTs are capable of DMA, it's conceivable they could corrupt > something. > > yes, it looks like something corrupts the memory. The UARTs do not have DMA enabled (BusMaster-) in the lspci output, so they shouldn't be able to corrupt memory. The NICs *do* have DMA enabled. Does the problem still happen if you turn off the NIC drivers (via-rhine and ath9k, it looks like)?
