[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-block] [PATCH 0/2] atapi: fix NetBSD boot regression
From: |
Mark Cave-Ayland |
Subject: |
Re: [Qemu-block] [PATCH 0/2] atapi: fix NetBSD boot regression |
Date: |
Mon, 09 Nov 2015 21:34:36 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Icedove/31.8.0 |
On 09/11/15 19:05, John Snow wrote:
> Marc noticed that a recent ATAPI permissions fix broke NetBSD 7.0's
> installer ISO.
>
> The problem is that it's meaningless to check for !(cmd->flags & nondata)
> if the command isn't supported, since all unsupported commands have
> _no_ flags. Effectively, all commands default to "Transfer Data" in our
> model until we classify them otherwise.
>
> This leads to a problem where we reject a zero byte BCL PIO command that
> transfers no data, simply because we have no properties for the command
> at all.
>
> Getting an ATA rejection for this command greatly confuses NetBSD.
>
> Correct behavior is to reject the command at the SCSI layer for being
> unsupported.
>
> ________________________________________________________________________________
>
> For convenience, this branch is available at:
> https://github.com/jnsnow/qemu.git branch atapi-bclimit-netbsd
> https://github.com/jnsnow/qemu/tree/atapi-bclimit-netbsd
>
> This version is tagged atapi-bclimit-netbsd-v1:
> https://github.com/jnsnow/qemu/releases/tag/atapi-bclimit-netbsd-v1
>
> John Snow (2):
> atapi: add byte_count_limit helper
> atapi: Prioritize unknown cmd error over BCL error
>
> hw/ide/atapi.c | 38 +++++++++++++++++++++++---------------
> 1 file changed, 23 insertions(+), 15 deletions(-)
I've tested this against my OpenBIOS ISO images here and it fixes the
problem without introducing any further regressions, so:
Tested-by: Mark Cave-Ayland <address@hidden>
Many thanks,
Mark.