Re: [Qemu-devel] [PATCH] slirp: kill ugly macros

[Kevin Wolf <kwolf@redhat.com>]

Am 15.02.2012 19:55, schrieb Jan Kiszka:
> On 2012-02-15 19:45, Michael S. Tsirkin wrote:
>> Remove ugly macros for field names,
>> change done by the following script:
>>
>> s#\bifq_prev\b#m_prev#g;
>> s#\bifq_next\b#m_next#g;
>> s#\bifs_prev\b#m_prevpkt#g;
>> s#\bifs_next\b#m_nextpkt#g;
>> s#\bifq_so\b#m_so#g;
>> s#\bm_next\b#m_hdr.mh_next#g;
>> s#\bm_prev\b#m_hdr.mh_prev#g;
>> s#\bm_nextpkt\b#m_hdr.mh_nextpkt#g;
>> s#\bm_prevpkt\b#m_hdr.mh_prevpkt#g;
>> s#\bm_flags\b#m_hdr.mh_flags#g;
>> s#\bm_len\b#m_hdr.mh_len#g;
>> s#\bm_data\b#m_hdr.mh_data#g;
>> s#\bm_size\b#m_hdr.mh_size#g;
>> s#\bm_dat\b#M_dat.m_dat_#g;
>> s#\bm_ext\b#M_dat.m_ext_#g;
> 
> Could you convert M_dat to m_dat as well (do not script, it's also a
> type)? It looks strange.

What are all these m_ and mh_ prefixes for struct fields even about?
When I have an mbuf, I know perfectly well that it is one, and that
m_hdr is a header of it.

So while we're cleaning up here, wouldn't mbuf->hdr->next make more
sense than mbuf->m_hdr->mh_next?

Kevin