----- Original Message ----- > On 04/25, Dave Anderson wrote: > > > > As I see it, this facility is simply another LIVE_SYSTEM memory source, > > of which there currently are /dev/mem, /proc/kcore and /dev/crash. > > Yes, if we are talking about 09/10 and 10/10. > > > The > > essential difference between them is the pc->readmem plugin: > > The essential difference is that this source is remote. See another email. > > > The CRASHBUILTIN stuff and related stuff you ran into is only there because > > live system analysis typically does not require a crash command line option, > > so crash has to figure what to do when a user just enters "crash". > > Ah. Yes, the usage of CRASHBUILTIN is ugly, and I tried to document this. > > We need the new (say) RAW_MEM_DUMP flag. We can't use RAMDUMP because it can > be set along with KDUMP. > > And there is no room in pc->flags, so I decided to abuse CRASHBUILTIN for now. I understand. It's exceeding rare that a new dumpfile type, or a new live system accessor in this case, gets introduced. When that happens, we can move deprecated leftovers from pc->flags to pc->flags2. So for example, all of these can be moved to pc->flags2 in one fell swoop: REM_NETDUMP REM_MCLXCD REM_LKCD REM_S390D Checking everywhere they get used, the (useless) settings in remote.c can be changed, and the very few macros that reference them can be changed to get them from pc->flags2. Dave -- Crash-utility mailing list Crash-utility@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/crash-utility
