From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40243) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YDFI5-000560-ON for qemu-devel@nongnu.org; Mon, 19 Jan 2015 11:35:02 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YDFI0-0000aR-Uw for qemu-devel@nongnu.org; Mon, 19 Jan 2015 11:35:01 -0500 Message-ID: <54BD3214.4060209@redhat.com> Date: Mon, 19 Jan 2015 17:34:28 +0100 From: Paolo Bonzini MIME-Version: 1.0 References: <1421667328-11800-1-git-send-email-mark.cave-ayland@ilande.co.uk> <1421667328-11800-3-git-send-email-mark.cave-ayland@ilande.co.uk> <54BCFC57.3010709@redhat.com> <54BD1C2D.9070603@suse.de> <54BD236E.1040200@redhat.com> In-Reply-To: Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH 2/2] m48t59: add mem_base value to m48t59_init_isa() List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Artyom Tarasenko Cc: Mark Cave-Ayland , Alexander Graf , qemu-devel , qemu-ppc@nongnu.org, =?UTF-8?B?SGVydsOpIFBvdXNzaW5lYXU=?= , =?UTF-8?B?QW5kcmVhcyBGw6RyYmVy?= On 19/01/2015 17:17, Artyom Tarasenko wrote: > On Mon, Jan 19, 2015 at 4:31 PM, Paolo Bonzini wr= ote: >> >> >> On 19/01/2015 16:22, Artyom Tarasenko wrote: >>>>>>> >>>>>>> On physical machines it's EBus, which is pretty much like 8-bit I= SA. >>>>>>> So, I think modelling it as ISA is closer to to the reality. >>>>>>> But out of curiosity, would it be possible to have a sysbus devic= e >>>>>>> somewhere in a middle of PCI space? [...] >>>>> >>>>> Why would you want to use a SysBusDevice in the first place? >>> Ask Paolo. :-) For me it's only important to have a MMIO device in th= e >>> proper address range. >> >> The reason I asked is simply because ISA devices never do MMIO (apart >> for the VGA window). >=20 > You mean in the QEMU world? At least physical SCSI and Ethernet > adapters had a MMIO space for the onboard ROM. Uh right, ROMs count as MMIO too. >>>>> I previously discussed with Mark that it should be an EBusDevice, n= ot an >>>>> ISADevice or SysBusDevice. >>> Interesting. I can't find this discussion in the list archive. Do you= suggest to >>> create EBusDevices for all ISA devices (serial, parallel, keyboard, >>> floppy) used in sun4u, or only for m48t59? >>> What would be the advantage of using EBusDevice over ISADevice? >> >> Is there a description of EBus and the sun4u memory map somewhere? >=20 > I could find only sparse pieces. "Uniprocessor System Controller > User's Manual" (805-0170.pdf) has some brief description, it's also > mentioned in the STP2223BGA and STP2200ABGA data sheets. >=20 >> Is there an "EBus bridge" PCI device similar to the PCI-to-ISA bridge? >=20 > As physical devices there are integrated SBus-to-EBus and PCI-to-EBus b= ridges. >=20 > But actually I may have been wrong about NVRAM always sitting on the > EBus: looking at the page 28 of "UltraSPARC=E2=84=A2-IIi User's Manual" > (805-0087.pdf), I see that NVRAM, Serial and other controllers reside > in a "PC compatible SuperIO" chip which sits on a PCI bus. That's an ISA bridge basically. I understand a little more of how this is supposed to work now, but I think it makes little sense to add this patch without the corresponding user. Paolo