On 30/10/18 9:18, Richard Henderson wrote:
On 10/29/18 11:20 PM, Philippe Mathieu-Daudé wrote:+static inline void pl330_init(uint32_t base, qemu_irq irq, int nreq) +{ + SysBusDevice *busdev; + DeviceState *dev; + + dev = qdev_create(NULL, TYPE_PL330); + qdev_prop_set_uint8(dev, "num_chnls", 8); + qdev_prop_set_uint8(dev, "num_periph_req", nreq); + qdev_prop_set_uint8(dev, "num_events", 16); + qdev_prop_set_uint8(dev, "data_width", 64); + qdev_prop_set_uint8(dev, "wr_cap", 8); + qdev_prop_set_uint8(dev, "wr_q_dep", 16); + qdev_prop_set_uint8(dev, "rd_cap", 8); + qdev_prop_set_uint8(dev, "rd_q_dep", 16); + qdev_prop_set_uint16(dev, "data_buffer_dep", 256); + qdev_init_nofail(dev); + + busdev = SYS_BUS_DEVICE(dev); + sysbus_mmio_map(busdev, 0, base); + sysbus_connect_irq(busdev, 0, irq); +}Why is this inline instead of in hw/dma/pl300.c? There should be nothing performance sensative here...
Yeah I didn't like it much neither and wondered the same :) I was looking a examples in hw/char: [phil@x1w qemu]$ git grep -hA 2 'static inline' include/hw/char/ 52:static inline DeviceState *cadence_uart_create(hwaddr addr, 53- qemu_irq irq, 54- Chardev *chr) -- 52:static inline DeviceState *cmsdk_apb_uart_create(hwaddr addr, 53- qemu_irq txint, 54- qemu_irq rxint, -- 18:static inline DeviceState *pl011_create(hwaddr addr, 19- qemu_irq irq, 20- Chardev *chr) -- 35:static inline DeviceState *pl011_luminary_create(hwaddr addr, 36- qemu_irq irq, 37- Chardev *chr) -- 18:static inline DeviceState *xilinx_uartlite_create(hwaddr addr, 19- qemu_irq irq, 20- Chardev *chr) I'll clean it and add a docstring. Thanks for the review, Phil.
