>> @@ -273,10 +273,12 @@ int kvm_s390_import_bp_data(struct kvm_vcpu *vcpu, >> vcpu->arch.guestdbg.nr_hw_wp = nr_wp; >> vcpu->arch.guestdbg.hw_wp_info = wp_info; >> return 0; >> -error: >> - kfree(bp_data); >> - kfree(wp_info); >> +free_bp_info: >> kfree(bp_info); >> +free_wp_info: >> + kfree(wp_info); >> +free_bp_data: >> + kfree(bp_data); >> return ret; >> } > > I agree with Cornelia,
This is generally fine. > while it seems correct from a technical point of view, Thanks for another bit of acknowledgement. > it will make the code harder to maintain. I agree that there some efforts and challenges involved. > For example if we ever add a new malloc and remove another one Do you see any changes coming from this direction? > over time we would need to reshuffle the labels This can occasionally happen, can't it? > and this did went wrong several times in the past. Would you like to add any corresponding software development experiences to discussions around a topic like "CodingStyle: add some more error handling guidelines"? https://www.spinics.net/lists/linux-doc/msg39307.html http://marc.info/?l=linux-doc&m=147187538413914 Regards, Markus
