00:18 <rjo> sb0: yeah. starvation was jsut a guess. in the cases where it works it looked like grinding to a halt after working a little while... also larger DMA regions appear to help.

00:19 <rjo> sb0: afaict the sdram works fine but i am only lookng at it through lm32+lasmi2wishbone.

05:50 <LoneTech> sb0: [15min] you overestimate the performance of my main machine. I should move this job to the workstation.

08:20 <sb0> rjo, it seems it doesn't crash when only the read core is started

08:20 <sb0> (fwiw)

08:22 <sb0> so.. that *might* suggest the sdram contents are being corrupted

08:26 <GitHub51> [misoc] sbourdeauducq pushed 3 new commits to master: http://git.io/PmTZKg

08:26 <GitHub51> misoc/master d3b96a0 Robert Jordens: programmer: make xc3sprog verbose

08:26 <GitHub51> misoc/master 81ed92d Robert Jordens: spiflash: redundant slice

08:26 <GitHub51> misoc/master ed902bf Robert Jordens: crt: add umoddi3

08:27 <sb0> the gdbstub could come handy here

08:31 <GitHub74> [misoc] sbourdeauducq pushed 2 new commits to master: http://git.io/qMZw7g

08:31 <GitHub74> misoc/master b26ac46 Sebastien Bourdeauducq: crt0: remove macadress for or1k as well

08:31 <GitHub74> misoc/master 6deeca0 Robert Jordens: bios/crt0.S: remove unused macaddr, add syscall handler stub

09:32 <sb0_> anyone interested in rewriting LM32 with Migen?

09:35 <GitHub195> [misoc] sbourdeauducq pushed 1 new commit to master: http://git.io/gKTlSg

09:35 <GitHub195> misoc/master ac97815 Sebastien Bourdeauducq: targets/simple: pass kwargs

09:36 <_florent_> hi

09:37 <_florent_> for the dma that crash, I will investigate with miscope

09:38 <_florent_> but I'm still working on rewriting Minimac with Migen for now

09:38 <_florent_> by the way, miscope is now easier to use

09:40 <_florent_> an example:

09:40 <_florent_> http://git.io/0pE7dw

09:41 <sb0_> hmm, memtest crashes on the mixxeo as well now

09:41 <_florent_> and the sofware on the host

09:41 <_florent_> http://git.io/zzxVwg

09:42 <sb0_> so we should probably do a bisect actually

09:44 <_florent_> hmm strange... but I remember being able to have the Mixxeo memtest working and the de0-nano memtest crashing with the same Migen/MiSoC revision...

09:46 <sb0_> it crashes reliably here

09:46 <sb0_> after less than a second

09:46 <sb0_> I'm building with "-Ot with_memtest True -s MiniSoC"

11:13 <sb0_> hm. it works with -s FramebufferSoC

11:13 <sb0_> could be that the bug manifests itself only when wb2lasmi is the only other master besides the memtest cores (even when other masters are inactive)

11:14 <sb0_> let me try adding a dummy master on the simplesoc ...

11:16 <_florent_> when I tested it was on FramebufferSoC or VideoMixerSoC indeed

11:24 <sb0_> hmm, adding a dummy master to simplesoc doesn't seem to help

21:39 <GitHub196> [NetBSD] fallen pushed 1 new commit to master: http://git.io/52G2KQ

21:39 <GitHub196> NetBSD/master ba76682 Yann Sionneau: Rework TLB miss handler...