Re: [Crash-utility] [PATCH v2] Fix 'waitq' command for 4.13 and later kernels

> > } else { > > - return; > > + error(FATAL, "cannot determine wait queue structure\n"); > > oh, I should have checked the replacement.. this emits compilation warnings: > > $ make clean ; make warn > ... > cc -c -g -DX86_64 -DLZO -DSNAPPY -DGDB_7_6 kernel.c -Wall -O2 -Wstrict-prototypes -Wmissing-prototypes -fstack-protector -Wformat-security > kernel.c: In function 'cmd_waitq': > kernel.c:9380:6: warning: 'start_index' may be used uninitialized in this function [-Wmaybe-uninitialized] > int start_index; /* where to start in wq array */ > ^~~~~~~~~~~ > kernel.c:9454:22: warning: 'task_offset' may be used uninitialized in this function [-Wmaybe-uninitialized] > readmem(wq_list[i] + task_offset, KVADDR, &task, > ~~~~~~~~~~~^~~~~~~~~~~~~ > kernel.c:9378:8: note: 'task_offset' was declared here > ulong task_offset; /* offset of task in wq element */ > ^~~~~~~~~~~ > > Hmm, in this case, I'd like to put the return back when merging. > Lianbo, is this ok? > > } else { > error(FATAL, "cannot determine wait queue structure\n"); After the log level is set to FATAL, it will exit, and the "return" has no chance to execute . So it could be good to use the log level "WARNING or INFO" for error()?