qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH 1/3] PPC4xx IIC and MAL


From: Salvatore Lionetti
Subject: Re: [Qemu-devel] [PATCH 1/3] PPC4xx IIC and MAL
Date: Wed, 10 Dec 2008 22:09:47 +0000 (GMT)

Hi,

I organized my little work in three 'patch queue', one for each argument:

- IIC
  MAL (only moved in other place plus 4,5 line added)
- EMAC
- WALNUT MACHINE

There is really need of endless soap for these three change?
Really having only a piece of entire patch about eg IIC will help people to 
comment this work?

Perhaps is wrong my point of view, but after three times that i mistake in 
patch , this is a reciprocal time wasting.

My personal opinion is that an integrator, when possible, for a cpu family is 
needed: not only saying 'this patch is big, this one is uncomprensible', but 
also to validate how qemu architecture is used or modified relatively to device 
specification, and how such device specification is reached, at least with 
coarse-grained; in the case comment about, i thing, is welcome. At least for me.

Have a good day

--- Mer 10/12/08, Hollis Blanchard <address@hidden> ha scritto:

> Da: Hollis Blanchard <address@hidden>
> Oggetto: Re: [Qemu-devel] [PATCH 1/3] PPC4xx IIC and MAL
> A: address@hidden
> Cc: "qemu-devel" <address@hidden>
> Data: Mercoledì 10 dicembre 2008, 19:53
> On Wed, 2008-12-10 at 14:13 +0000, Salvatore Lionetti wrote:
> > Ok
> > 
> > 1) The line code modified is 80% IIC, only few row
> about MAL
> > 2) MAL until today (if ever this patch enter in qemu)
> is only used to
> > rw registers, so no malfunction is present.
> > 3) As you can see many ppc* file are to modify since
> I2C previous
> > support it totally not integrated: have sense modify
> ppc* and never
> > use an IIC device?
> > 4) Is error prone and very tedious to extract some
> code, expecially if
> > there is no modularity advantage: one piece require
> the other,
> > 5) I already have successfully tested the expected
> patch.
> > 6) I can't do that directly so every time have 2
> see if main line is
> > updated or not, merge svn file with current my patch
> status
> > 7) Who apply the patch only apply one click?
> > 
> > In this moment i have no time to make all this retail
> and test.
> 
> I would be very surprised if somebody applies your patch
> as-is, because
> it's messy (there are a lot of unnecessary/unrelated
> changes) and it's
> very difficult to see the functional changes.
> 
> Yes, there is a natural dependency of a patch on the
> previous ones. This
> is called a "patch queue" or "patch
> series", and you can see other
> people submitting them on qemu-devel. (In fact the most
> recent series
> contains 13 patches, with #13 depending on #1-12 being
> committed
> already.)
> 
> There are tools (such as quilt, guilt, and MQ) that help
> you manage
> patch series. You can use these tools to maintain your
> patches and
> update them to match the latest qemu tree.
> 
> If you submit small independent patches, people will be
> able to easily
> see that they're good and commit them instantly. When
> you submit a large
> patch with lots of changes, it takes a lot more time and
> effort to
> review, and it's much less likely to be accepted.
> 
> -- 
> Hollis Blanchard
> IBM Linux Technology Center







reply via email to

[Prev in Thread] Current Thread [Next in Thread]