I am running into an issue where the metadata ROM is being filled with 2070 dwords, and when the FPGA is loaded, I get the following message when trying to do ocpihdl get: # ocpihdl get Exiting for problem: Metadata ROM appears corrupted: 0x1 0x2048 0xbe897958 The bottom byte and the top byte seem to be stable, but the middle 2 bytes change every time I run the program. I noticed this issue was recently merged in: https://gitlab.com/opencpi/opencpi/-/issues/1328 I looked at metadata_rv and I have the generic number of words being set, so I believe it should all be there, but I am running from the v2.1.0-rc.2 tag. Was something merged in more recently than when v2.1.0-rc.2 was tagged? Was this feature tested with a large design? Thanks, Brian

There was a change made to dynamically size the metadata ROM since several designs were exceeding the previous fixed limit, and we didn’t want to expand the resource usage for all designs. So there are two likely explanations. One is that a complete rebuild was not done after updating OpenCPI, and another is that we have some sort of bug in this change. We’ll look into it a bit, but we may need more information on your failure. Cheers, Jim > On Mar 11, 2021, at 8:18 PM, Brian Padalino <bpadalino@gmail.com> wrote: > > I am running into an issue where the metadata ROM is being filled with 2070 dwords, and when the FPGA is loaded, I get the following message when trying to do ocpihdl get: > > # ocpihdl get > Exiting for problem: Metadata ROM appears corrupted: 0x1 0x2048 0xbe897958 > > The bottom byte and the top byte seem to be stable, but the middle 2 bytes change every time I run the program. > > I noticed this issue was recently merged in: > > https://gitlab.com/opencpi/opencpi/-/issues/1328 <https://gitlab.com/opencpi/opencpi/-/issues/1328> > > I looked at metadata_rv and I have the generic number of words being set, so I believe it should all be there, but I am running from the v2.1.0-rc.2 tag. > > Was something merged in more recently than when v2.1.0-rc.2 was tagged? Was this feature tested with a large design? > > Thanks, > Brian > _______________________________________________ > discuss mailing list -- discuss@lists.opencpi.org > To unsubscribe send an email to discuss-leave@lists.opencpi.org

On Thu, Mar 11, 2021 at 9:50 PM Kulp Jim <jek@parera.com> wrote: > There was a change made to dynamically size the metadata ROM since several > designs were exceeding the previous fixed limit, and we didn’t want to > expand the resource usage for all designs. > Understandable. > > So there are two likely explanations. One is that a complete rebuild was > not done after updating OpenCPI, and another is that we have some sort of > bug in this change. > This was from a fresh and clean build of the v2.1.0-rc.2 tag. > > We’ll look into it a bit, but we may need more information on your failure. > Understandable. I'm willing to give you whatever information you need, just let me know. Thanks, Brian