Re: [Qemu-devel] [PATCH for-2.3 0/3] Contain drive_get() misuse

[Markus Armbruster]

Paolo Bonzini <address@hidden> writes:

>> I really can't see why we should tie ourselves into knots to avoid an
>> incompatible change here.  I seriously doubt anyone will notice if drop
>> the mistaken automatic backend pickup so that "-drive if=floppy -device
>> i82378 -device pc87312" no longer picks up the floppy.
>
> Ok, it wasn't too hard to convince me.  Let's document the three of them
> together.  I'll do it tomorrow if you haven't beaten me to it.

I started to do it, then stopped to look for misuse of other kinds of
backends.  Here are some:

* serial_hds[] via qemu_char_get_next_serial()

  device                bus     depends on        flawed function
  "cadence_uart"        sysbus  CONFIG_CADENCE    cadence_uart_realize()
  "digic-uart"          sysbus  CONFIG_DIGIC      digic_uart_realize()
  "etraxfs,serial"      sysbus  CONFIG_ETRAXFS    etraxfs_ser_init()
  "lm32-juart"          sysbus  CONFIG_LM32       m32_juart_init()
  "lm32-uart"           sysbus  CONFIG_LM32       lm32_uart_init()
  "milkymist-uart"      sysbus  CONFIG_MILKYMIST  milkymist_uart_realize()
  "pl011"               sysbus  CONFIG_PL011      pl011_realize()
  "stm32f2xx-usart"     sysbus  CONFIG_STM32F2XX_USART
                                                  stm32f2xx_usart_init()
  "xlnx.xps-uartlite"   sysbus  CONFIG_XILINX     xilinx_uartlite_realize()

* serial_hds[] directly

  device                bus     depends on        flawed function
  "allwinner-a10"       none    CONFIG_ALLWINNER_A10
                                                  aw_a10_realize()
  "pc87312"             isa     CONFIG_PC87312    pc87312_realize()

* parallel_hds[] directly

  device                bus     depends on        flawed function
  "pc87312"             isa     CONFIG_PC87312    pc87312_realize()

I'll prepare a patch tomorrow that marks them all FIXME, just like the
drive_get() misuses.

What else needs to be done?

* "pc87312"

   Old news, just more incompatible change to document.

* "allwinner-a10"

   "-nodefaults -serial null -device allwinner-a10" doesn't explode,
   which means I can't exclude the possibility that this might actually
   do something useful for someone.  I'd say we treat it just like
   "pc87312": leave alone now, document incompatible change.

* Of the sysbus devices only "xlnx.xps-uartlite" seems to be available
  with the machines that support -device for sysbus devices (ppce500 and
  pseries-*).  When I try to -device it there, I get "Device
  xlnx.xps-uartlite is not supported by this machine yet."  I'll ask
  Alex to confirm.  I'll prepare a patch that sets
  cannot_instantiate_with_device_add_yet for the unavailable ones to
  keep them unavailable, just like for the drive_get() abusers.

How could this kind of mistake could look like in NIC device models?
"allwinner-a10"'s instance_init aw_a10_init() provides a clue: it messes
with qemu_check_nic_model():

    if (nd_table[0].used) {
        qemu_check_nic_model(&nd_table[0], TYPE_AW_EMAC);
        qdev_set_nic_properties(DEVICE(&s->emac), &nd_table[0]);
    }

No other device does that.  Without -nodefaults, this fails:

    -device allwinner-a10: Unsupported NIC model: xgmac

I haven't checked all uses of nd_table[].