On 18.12.18 13:28, Thomas Huth wrote: > On 2018-12-18 13:01, Janosch Frank wrote: >> On 18.12.18 12:50, Thomas Huth wrote: > [...] >>> Maybe rather assert(sccb->h.response_code == SCLP_RC_NORMAL_COMPLETION) ? >>> It does not make much sense to continue otherwise, does it? >> >> Will the assert abort without any further prints? > > No ... so that does not work. So maybe do a disabled wait with irqs > disabled instead to signal a crash to the host? > > Thomas > Honestly, I'm thinking about dropping the check. It often works without any init because the stage 3 bootloader does an init and we do not check the print functions for errors. It's best effort at this point :) I'm more concerned with interrupts that print things while a printf holds the sclp_busy lock. But that mostly happens if you mess up a sclp call.
Attachment:
signature.asc
Description: OpenPGP digital signature
