On Wed, 25 Feb 2026 at 16:38, Daniel P. Berrangé <[email protected]> wrote: > > On Thu, Feb 19, 2026 at 10:23:36AM +0000, Peter Maydell wrote: > > Why do we care about the qemu_log output matching the > > error-report output? The logs are expected to be > > quite frequent, to only be there if you've explicitly > > turned them on, and to be usually directed to a log file. > > Personally I never use the log-to-file facility. Instead typical use > case is relying on the trace points 'log' facility as illustrated in > the commit message example to dump to stderr for simple debugging, > and having consistent formatting between the logs & errors is useful > there.
Do you not find that the tracepoint output is often just too big to be manageable via stderr? (To take a random example, the last thing I did with tracepoint logging produced a file 196MB large for a not very long run of the guest.) -- PMM
