10.07.2013 00:45, Anthony Liguori wrote: > Michael Tokarev <m...@tls.msk.ru> writes: > >> 09.07.2013 22:37, Anthony Liguori wrote: >>> Michael Tokarev <m...@tls.msk.ru> writes: >>> >>>> It looks like initially there was -nographic option to turn >>>> off display, now there's another option of the same sort, >>>> -display none. But code in other places of qemu checks for >>>> DT_NOGRAPHIC and does not work well with -display none. >>>> Make DT_NOGRAPHIC an internal version which selects DT_NONE, >>>> and check for that in all other places where previously we >>>> checked for DT_NOGRAPHIC. >> [] >>> Breaks make check: >>> >>> main-loop: WARNING: I/O thread spun for 1000 iterations >>> ** >>> ERROR:/home/aliguori/git/qemu/tests/fw_cfg-test.c:63:test_fw_cfg_nographic: >>> assertion failed (qfw_cfg_get_u16(fw_cfg, FW_CFG_NOGRAPHIC) == 0): (1 == 0) >>> GTester: last random seed: R02S25031265f05e4d41efcf758c9ef6043b >> >> Sure, because the test is bogus. > > No, it's a guest ABI. You cannot change the guest ABI. > > -display none != -nographic. > > nographic gives you -display none plus a stdio serial port (with muxing > magic). > > -display none should not imply stdio serial port. The vc goes to a > dummy display. That's a major semantic difference.
So we're back to the original bugreport by Todd Fries -- apparently all my attempts to describe the proposed ways to change it failed. Note that -nographics does NOT imply a serial port either, because it can be redirected elsewhere. /mjt
