> On Oct 4, 2021, at 11:41 AM, Trond Myklebust <trondmy@xxxxxxxxxxxxxxx> wrote: > > On Mon, 2021-10-04 at 10:10 -0400, Chuck Lever wrote: >> ida_simple_get() returns a signed integer. Negative values are error >> returns, but this suggests the range of valid client IDs is zero to >> 2^31 - 1. >> >> tk_pid is currently an unsigned short, so its range is zero to >> 65535. >> >> For certain special cases, RPC-related tracepoints record a -1 as >> the task ID or the client ID. It's ugly for a trace event to display >> 4 billion in these cases. > > Ugh... I emphatically do not like the idea of an identifier field that > is signed, whatever its range of validity may be. > > If we're going to change anything, then let's rather turn that > identifier into a fixed size hex field in the traces. A fixed-size field might be more legible for eyeballing large complex traces, so I can do that. Thanks for having a look! -- Chuck Lever
