Re: [Qemu-devel] [PATCH for-2.9? v2 resend] ipmi: Fix macro issues

From: Eric Blake <eblake@redhat.com>
To: minyard@acm.org, qemu-devel@nongnu.org
Cc: qemu-trivial@nongnu.org, Ed Maste <emaste@freebsd.org>,
	Corey Minyard <cminyard@mvista.com>
Subject: Re: [Qemu-devel] [PATCH for-2.9? v2 resend] ipmi: Fix macro issues
Date: Fri, 31 Mar 2017 09:01:48 -0500	[thread overview]
Message-ID: <7ba4f927-8d0b-2387-c324-42670c8265f3@redhat.com> (raw)
In-Reply-To: <f0dd4986-add0-b3d5-745e-e56ca0d23a22@acm.org>

[-- Attachment #1: Type: text/plain, Size: 1109 bytes --]

On 03/31/2017 08:32 AM, Corey Minyard wrote:

>> Note, on the other hand, that a call such as
>> IPMI_BT_SET_HBUSY(ib->control_reg, 2) would result in writing 0 to the
>> HBUSY bit. In other words, the IPMI_BT_SET_HBUSY() macro is rather weird
>> in that it sets or clears the HBUSY bit based solely on whether its v
>> parameter is even or odd, rather than the more usual semantics of
>> whether the v parameter is 0 or non-zero.  We could change that if we
>> wanted - by having the macro expand to "(left | (!!(v) << shift))"
>> instead of our current expansion of "(left | (((v) & 1) << shift))" -
>> but I still don't think it would change the semantics of any existing
>> caller.
> 
> Yeah, that would be better.  Should I do another patch?

Up to you; I'm fine with either the existing patch as proposed or with
reviewing a v4 (interesting that you resent v2, even though you had also
posted a v3 at one point, where the only difference seems to have been
adding my R-b).

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]