Re: [Xen-devel] [PATCH] Add dm-userspace to the Xen kernel

[Dan Smith <danms@xxxxxxxxxx>]

BB> 1. this struct differs from the old patches published on dm-devel
BB> and lkml?

I added an 'id' field, which contains a unique integer for each
request, which helps me to match up the response to the correct item
in the kernel queue.  Previously, I was matching based on the original
block, which has the potential to be wrong if there are two requests
on the queue for the same block (i.e. one for a read mapping and one
for a write mapping).  I thought the id would be easier.

BB> 2. Why do you use a construct which gives different allignment on
BB> i386 and x86_64? (i386 alligns 64bit ints on 32bit, x86_64 on
BB> 64bit

I assume you mean because there are an odd number of 32-bit fields, is
that correct?  The answer to the question is: "because I haven't given
much thought to x86_64 issues yet" :).  This week, I plan to test on
x86_64, so I can submit another patch with resolutions to any other
x86_64 issues that may be present, if it's likely to be accepted.

-- 
Dan Smith
IBM Linux Technology Center
Open Hypervisor Team
email: danms@xxxxxxxxxx

Attachment: pgpZOrpu5EXUZ.pgp
Description: PGP signature

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel