Re: [PATCH v2 1/2] block: Fixes nfs compiling error on msys2/mingw

[Max Reitz <mreitz@redhat.com>]

On 05.11.20 12:46, 罗勇刚(Yonggang Luo) wrote:
> 
> 
> On Thu, Nov 5, 2020 at 7:44 PM Max Reitz <mreitz@redhat.com 
> <mailto:mreitz@redhat.com>> wrote:
>  >
>  > On 05.11.20 12:32, 罗勇刚(Yonggang Luo) wrote:
>  > >
>  > >
>  > > On Thu, Nov 5, 2020 at 7:29 PM Max Reitz <mreitz@redhat.com 
> <mailto:mreitz@redhat.com>
>  > > <mailto:mreitz@redhat.com <mailto:mreitz@redhat.com>>> wrote:
>  > >  >
>  > >  > On 13.10.20 02:15, Yonggang Luo wrote:
>  > >  > > These compiling errors are fixed:
>  > >  > > ../block/nfs.c:27:10: fatal error: poll.h: No such file or 
> directory
>  > >  > >     27 | #include <poll.h>
>  > >  > >        |          ^~~~~~~~
>  > >  > > compilation terminated.
>  > >  > >
>  > >  > > ../block/nfs.c:63:5: error: unknown type name 'blkcnt_t'
>  > >  > >     63 |     blkcnt_t st_blocks;
>  > >  > >        |     ^~~~~~~~
>  > >  > > ../block/nfs.c: In function 'nfs_client_open':
>  > >  > > ../block/nfs.c:550:27: error: 'struct _stat64' has no member named
>  > > 'st_blocks'
>  > >  > >    550 |     client->st_blocks = st.st_blocks;
>  > >  > >        |                           ^
>  > >  > > ../block/nfs.c: In function 'nfs_get_allocated_file_size':
>  > >  > > ../block/nfs.c:751:41: error: 'struct _stat64' has no member named
>  > > 'st_blocks'
>  > >  > >    751 |     return (task.ret < 0 ? task.ret : st.st_blocks * 
> 512);
>  > >  > >        |                                         ^
>  > >  > > ../block/nfs.c: In function 'nfs_reopen_prepare':
>  > >  > > ../block/nfs.c:805:31: error: 'struct _stat64' has no member named
>  > > 'st_blocks'
>  > >  > >    805 |         client->st_blocks = st.st_blocks;
>  > >  > >        |                               ^
>  > >  > > ../block/nfs.c: In function 'nfs_get_allocated_file_size':
>  > >  > > ../block/nfs.c:752:1: error: control reaches end of non-void
>  > > function [-Werror=return-type]
>  > >  > >    752 | }
>  > >  > >        | ^
>  > >  > >
>  > >  > > On msys2/mingw, there is no st_blocks in struct _stat64 yet, we
>  > > disable the usage of it
>  > >  > > on msys2/mingw, and create a typedef long long blkcnt_t; for
>  > > further implementation
>  > >  > >
>  > >  > > Signed-off-by: Yonggang Luo <luoyonggang@gmail.com 
> <mailto:luoyonggang@gmail.com>
>  > > <mailto:luoyonggang@gmail.com <mailto:luoyonggang@gmail.com>>>
>  > >  > > ---
>  > >  > >   block/nfs.c | 15 +++++++++++++++
>  > >  > >   1 file changed, 15 insertions(+)
>  > >  > >
>  > >  > > diff --git a/block/nfs.c b/block/nfs.c
>  > >  > > index f86e660374..cf8795fb49 100644
>  > >  > > --- a/block/nfs.c
>  > >  > > +++ b/block/nfs.c
>  > >  >
>  > >  > [...]
>  > >  >
>  > >  > > @@ -51,6 +53,10 @@
>  > >  > >   #define QEMU_NFS_MAX_PAGECACHE_SIZE (8388608 / NFS_BLKSIZE)
>  > >  > >   #define QEMU_NFS_MAX_DEBUG_LEVEL 2
>  > >  > >
>  > >  > > +#if defined(_WIN32)
>  > >  > > +typedef long long blkcnt_t;
>  > >  >
>  > >  > This makes me uneasy.  Is there a guarantee that this type will 
> never be
>  > >  > defined in a mingw environment (which would then result in a new 
> compile
>  > >  > error)?
>  > > That's be sure, mingw is compat with msvcrt.dll, so blkcnt_t  are 
> always
>  > > not defined.
>  >
>  > And there is no possibility of that being added in the future?
> If that's the case, what's the better option.

As I said, just making NFSClient.st_blocks a uint64_t.

Max