Holger Waechtler wrote:
Hi Johannes,
did you modified the CRC check so that you will only check packets with
section syntax indicator=1? Doing so in general would be a bad idea i
think, I would rather add a new DMX_CHECK_CRC_SSI suggest this semantic:
DMX_CHECK_CRC not set - always pass packets
DMX_CHECK_CRC bit set - check always
DMX_CHECK_CRC_SSI bit set - check if section syntax indicator is '1'
This would allow CRC checking also for TOT sections and for proprietary
section formats like firmware upgrades etc. What do you think?
I like it like it is now, where the driver behaves the same with
hw_sections=1 and hw_sections=0. As it seems to me that hw_sections=1
works better with MPE (can someone confirm this?), we should make this
the default again. Why did hw_sections=0 get to be the default now?
I fear hw_sections=1 means no CRC at all -- or did somebody implemented
firmware-based crc32 checksums in the meantime?