linux-kernel - [RFC PATCH] Xen-SWIOTLB fixes (v1) for 3.7

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1343335407-5465-1-git-send-email-konrad.wilk@oracle.com>
Date:	Thu, 26 Jul 2012 16:43:25 -0400
From:	Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>
To:	linux-kernel@...r.kernel.org
Cc:	xen-devel@...ts.xensource.com
Subject: [RFC PATCH] Xen-SWIOTLB fixes (v1) for 3.7

This is an RFC patch for 3.7.

I thought I had addressed this in the past but I can't seem to find
the patches. There is one bug - if one boots a PV 64-bit guests
with more than 4GB, the SWIOTLB gets turned on - and 64MB of precious
low-memory gets used. If you launch more than 10 of them on a 32GB
machines you are going to run in-to trouble as the lowmem gets
exhausted.

On the other hand, the user might want to have 10 guests with 4GB
and each with a PCI device!

So to fix this, we are going to figure out whether the user had
provided the e820_hole=1 parameter in the guest config. The effect
of that parameter is that a massaged host's E820 is used in the guest
 - and we check if it has E820_ACPI or E820_NVS. If so, the user
really wanted to pass in PCI devices to the guest.

Since we now have a routine to check for the e820_hole we can optimize and
see if the user forgot the "iommu=soft" and automatically turn that
on.

The patches are RFC b/c it looks like something has gone bit-rotten the
last time I used this (when Fedora Core 16 was released) b/c I can't
get a guest to boot with e820_hole :-(. But they [the patches] look
sound to me and they do fix the bug of allocating SWIOTLB for normal
PV guests..

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/