Re: [Qemu-devel] [PATCH 06/10] r2d: Flash memory creation is confused about size, mark FIXME

From: Markus Armbruster <armbru@redhat.com>
To: "Philippe Mathieu-Daudé" <philmd@redhat.com>
Cc: "Magnus Damm" <magnus.damm@gmail.com>,
	"Kevin Wolf" <kwolf@redhat.com>,
	"Peter Maydell" <peter.maydell@linaro.org>,
	Qemu-block <qemu-block@nongnu.org>,
	"Alex Bennée" <alex.bennee@linaro.org>,
	"QEMU Developers" <qemu-devel@nongnu.org>,
	"Max Reitz" <mreitz@redhat.com>, qemu-ppc <qemu-ppc@nongnu.org>,
	"Laszlo Ersek" <lersek@redhat.com>
Subject: Re: [Qemu-devel] [PATCH 06/10] r2d: Flash memory creation is confused about size, mark FIXME
Date: Mon, 04 Mar 2019 16:33:52 +0100	[thread overview]
Message-ID: <87k1he1wnz.fsf@dusky.pond.sub.org> (raw)
In-Reply-To: <f2c8f20d-17ad-b72b-28fb-c835d6b97c10@redhat.com> ("Philippe =?utf-8?Q?Mathieu-Daud=C3=A9=22's?= message of "Mon, 4 Mar 2019 12:43:10 +0100")

Philippe Mathieu-Daudé <philmd@redhat.com> writes:

> On 3/4/19 8:25 AM, Markus Armbruster wrote:
>> Magnus Damm <magnus.damm@gmail.com> writes:
>>> On Wed, Feb 20, 2019 at 2:31 AM Markus Armbruster <armbru@redhat.com> wrote:
>>>> Perhaps Magnus, who maintains the machine, can pick a new value for us.
>>>
>>> According to the old board user document in Japanese (under NDA) what
>>> is referred to as FROM (Area0) is connected via a 32-bit bus and CS0
>>> to CN8. The docs mention s29pl127j60tfi130 but since I don't have the
>>> board handy ATM I don't know how the chips are connected.
>>>
>>> Hope this helps,
>> 
>> If you want me to change our emulated flash memory's size, please give
>> me a number.
>
> datasheet "S29PL-J 002-00615 Rev. *E":
> https://www.cypress.com/file/207091/download
>
> The S29PL127J60TFI130 is a 128Mbit NOR pflash addressable in words of 16bit.
>
> 128Mbit = 16 MiB
>
> At least it matches the "RTS7751R2D Handling Manual"!
> https://elinux.org/RTS7751R2D_Handling_Manual#Kernel_start_from_FROM_extension_card_.28Kernel_space_XIP.29
>
> PL127J:
> - 4 Banks
>   -> we don't model banks.
> - sectors of 4Kw and 32Kw
>   -> we don't model different sector size and only use the
>      biggest available
>
> sector_size = 32Kw = 64KiB // sector_len
> (naive) sector_count = 256 // nb_blocs
>
> ManufID:  0001h
> DeviceID: 227Eh 2220h 2200h
>
> I understand "connected via a 32-bit bus and CS0 to CN8" as the full
> device wordsize is addressable, so this device
>
> So in pflash_cfi02_register() format:
>
> - name = "FROM (Area0)"
> - size = 16 * MiB
> - sector_len = 64 * KiB
> - nb_blocs = 256
> - nb_mappings = 1? /* Machine specific... */
> - width = 2
> - id0 = 0x0001
> - id1 = 0x227e
> - id2 = 0x2220
> - id3 = 0x2200
> - unlock_addr0 = 0x555,
> - unlock_addr1 = 0x2aa
> - be = 0 /* Arch specific... */
>
> Which hopefully is very similar to what we currently use :)

'fraid not:

    $ qemu-system-sh4 -M r2d -display none -S -monitor stdio
    QEMU 3.1.50 monitor - type 'help' for more information
    (qemu) info qtree
    [...]
      dev: cfi.pflash02, id ""
        drive = ""
        num-blocks = 512 (0x200)        <--- instead of 256
        sector-length = 16384 (0x4000)  <--- instead of 65536
        width = 4 (0x4)                 <--- instead of 2
        mappings = 1 (0x1)
        big-endian = 0 (0x0)
        id0 = 0 (0x0)                   <--- instead of 0x0001
        id1 = 0 (0x0)                   <--- instead of 0x227e
        id2 = 0 (0x0)                   <--- instead of 0x2220
        id3 = 0 (0x0)                   <--- instead of 0x2200
        unlock-addr0 = 1365 (0x555)
        unlock-addr1 = 682 (0x2aa)
        name = "r2d.flash"              <--- instead of "FROM (Area0)"
        mmio 0000000000000000/0000000000800000

Not shown: size is num-blocks * sector-length = 512 * 16384 = 8MiB
instead of 16MiB.

Which properties would you like me to change, and how?