Re: [PATCH 8/9] monitor: move hmp_info_block* to blockdev-hmp-cmds.c

[Kevin Wolf]

Am 27.01.2020 um 14:33 hat Markus Armbruster geschrieben:
> Maxim Levitsky <address@hidden> writes:
> 
> > On Wed, 2019-11-27 at 09:08 +0100, Markus Armbruster wrote:
> >> I think it makes sense to collect *all* block HMP stuff here.
> >> 
> >> Left in monitor/hmp-cmds.c: hmp_eject(), hmp_nbd_server_start(), ...
> >> 
> >> I guess hmp_change() has to stay there, because it's both block and ui.
> >> 
> >> Left in blockdev.c: hmp_drive_add_node().
> >
> > Thank you very much. I added these and bunch more to my patchset.
> >
> >> 
> >> Quick grep for possible files to check:
> >> 
> >> $ git-grep -l 'monitor[a-z_-]*.h' | xargs grep -l 'block[a-z_-]*\.h'
> >> MAINTAINERS
> >> blockdev-hmp-cmds.c
> >> 
> >
> >> blockdev.c
> > hmp_drive_add_node is there and I moved it too.
> >
> >
> >> cpus.c
> > Nothing suspicious
> >
> >> dump/dump.c
> > qmp_dump_guest_memory is only monitor reference there I think
> >
> >> hw/display/qxl.c
> > No way that is related to the block layer
> >
> >> hw/scsi/vhost-scsi.c
> > All right, the monitor_fd_param is an interesting thing.
> > Not related to block though.
> >
> >> hw/usb/dev-storage.c
> > All right, this for no reason includes monitor/monitor.h,
> > added patch to remove this because why not.
> >
> >> include/monitor/monitor.h
> > Nothing suspicious
> >
> >> migration/migration.c
> > Nothing suspicious
> >
> >> monitor/hmp-cmds.c
> > Added hmp_qemu_io
> >
> > Maybe I need to add hmp_delvm too?
> > savevm/delvm do old style snapshots
> > which are stored to the first block device
> 
> One foot in the block subsystem, the other foot in the migration
> subsystem.  I'm not sure where this should go.  Kevin?

Moving it to block is not an obvious improvement, so I think I'd leave
it alone.

Kevin