drtio: reading still needs work buffer after all #154

mwojcik · 2021-11-29T12:33:11+08:00

mwojcik commented

2021-11-29 12:33:11 +08:00

Satellite was hanging after receiving first 40-byte packet that had the routing table data. Investigation showed it was because of axi burst reads (not yet implemented, of course) - they didn't happen on master because the packets received were small enough to get by. Temporary solution is of course to bring back the work buffer, and also change that one to use 16-bit reads/writes instead of 32-bit not to trigger any bursts.

Also noticed in zc706 satellite that the siphaser had "ultrascale" set as False, copied straight from kc705 - but that's not true here. Shouldn't change much.

Satellite was hanging after receiving first 40-byte packet that had the routing table data. Investigation showed it was because of axi burst reads (not yet implemented, of course) - they didn't happen on master because the packets received were small enough to get by. Temporary solution is of course to bring back the work buffer, and also change that one to use 16-bit reads/writes instead of 32-bit not to trigger any bursts. Also noticed in zc706 satellite that the siphaser had "ultrascale" set as False, copied straight from kc705 - but that's not true here. Shouldn't change much.

mwojcik added 1 commit 2021-11-29 12:33:12 +08:00

e8a18a7977 drtio: reading still needs work buffer after all

sb10q merged commit ada3f2e704 into master

2021-11-29 12:46:09 +08:00

Sign in to join this conversation.

No reviewers