Am 04.02.2015 um 18:51 hat Eric Blake geschrieben: > From: Stefan Hajnoczi <stefa...@redhat.com> > > The block.c file has grown to over 6000 lines. It is time to split this > file so there are fewer conflicts and the code is easier to maintain. > > Extract I/O request processing code: > * Read > * Write > * Flush > * Discard > * ioctl > * Tracked requests and queuing > * Throttling and copy-on-read > > The patch simply moves code from block.c into block/io.c. > > No code changes are made except adding the following block_int.h > functions so they can be called across block.c and block/io.c: > bdrv_drain_one(), bdrv_set_dirty(), bdrv_reset_dirty(). > > I/O request processing needs to set up BlockDriver coroutine and AIO > emulation function pointers, so add bdrv_setup_io_funcs(bdrv) interface > that block.c calls. > > Signed-off-by: Stefan Hajnoczi <stefa...@redhat.com> > --- > > This patch produces identical results to Stefan's email, but is > MUCH more readable (hint: git config diff.algorithm patience)
Any thoughts on moving (some of) the following functions in addition? * bdrv_drain_all() * bdrv_make_zero() * bdrv_flush_all() * bdrv_is_allocated/get_block_status family of functions * Anything with respect to BlockLimits * load/save_vmstate() * qemu_blockalign() and friends These are more or less all of the block.c functions that are clearly not operations on the graph, but more or less directly related to some kind of I/O. If we move out the graph management code, these would be left in block.c, which is probably not what we want. Kevin
