[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCHv2 1/4] fw_cfg: don't map the fw_cfg IO ports in
From: |
Mark Cave-Ayland |
Subject: |
Re: [Qemu-devel] [PATCHv2 1/4] fw_cfg: don't map the fw_cfg IO ports in fw_cfg_io_realize() |
Date: |
Tue, 13 Jun 2017 19:27:11 +0100 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.8.0 |
On 12/06/17 23:27, Laszlo Ersek wrote:
> It turns out that we introduced the "iobase" and "dma_iobase" properties
> *solely* so that we could pass arguments to fw_cfg_io_realize(). But
> fw_cfg_io_realize() only needed those parameters for the *wrong* purpose
> (namely calling sysbus_add_io()). By eliminating the sysbus_add_io()
> calls from fw_cfg_io_realize(), the related properties become
> unnecessary too.
>
> (NB, setting the "dma_enabled" property in fw_cfg_init_io_dma(), and
> setting "data_width" and "dma_enabled" in fw_cfg_init_mem_wide(), remain
> necessary, because those aren't related to region mapping.)
>
> What do you think?
Yeah, I like the sound of this approach much better since it keeps the
functionality similar between both the memory and ioport instances. I've
got something basic going but I don't think I'm going to have time to
finish it tonight.
ATB,
Mark.
- [Qemu-devel] [PATCHv2 0/4] fw_cfg: qdev-related tidy-ups, Mark Cave-Ayland, 2017/06/12
- [Qemu-devel] [PATCHv2 4/4] fw_cfg: move QOM type defines into fw_cfg.h, Mark Cave-Ayland, 2017/06/12
- [Qemu-devel] [PATCHv2 1/4] fw_cfg: don't map the fw_cfg IO ports in fw_cfg_io_realize(), Mark Cave-Ayland, 2017/06/12
- [Qemu-devel] [PATCHv2 2/4] fw_cfg: move setting of FW_CFG_VERSION_DMA bit to fw_cfg_init1(), Mark Cave-Ayland, 2017/06/12
- [Qemu-devel] [PATCHv2 3/4] fw_cfg: move qdev_init_nofail() from fw_cfg_init1() to callers, Mark Cave-Ayland, 2017/06/12
- Re: [Qemu-devel] [PATCHv2 0/4] fw_cfg: qdev-related tidy-ups, Paolo Bonzini, 2017/06/14