Re: [PATCH v6 1/2] io: Add support for MSG_PEEK for socket channel

[Juan Quintela]

Daniel P. Berrangé <berrange@redhat.com> wrote:
> On Thu, Feb 02, 2023 at 02:39:05PM +0100, Juan Quintela wrote:
>> Daniel P. Berrangé <berrange@redhat.com> wrote:
>> > On Thu, Feb 02, 2023 at 01:51:28PM +0100, Juan Quintela wrote:
>> >> Daniel P. Berrangé <berrange@redhat.com> wrote:
>> >> > On Thu, Feb 02, 2023 at 01:22:12PM +0100, Juan Quintela wrote:
>> >> >> "manish.mishra" <manish.mishra@nutanix.com> wrote:
>> >> >> > MSG_PEEK peeks at the channel, The data is treated as unread and
>> >> >> > the next read shall still return this data. This support is
>> >> >> > currently added only for socket class. Extra parameter 'flags'
>> >> >> > is added to io_readv calls to pass extra read flags like MSG_PEEK.
>> >> >> >
>> >> >> > Reviewed-by: Peter Xu <peterx@redhat.com>
>> >> >> > Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
>> >> >> > Suggested-by: Daniel P. Berrange <berrange@redhat.com>
>> >> >> > Signed-off-by: manish.mishra <manish.mishra@nutanix.com>
>> >> >> 
>> >> >> 
>> >> >> This change breaks RDMA migration.
>> >> >> 
>> >> >> FAILED: libcommon.fa.p/migration_rdma.c.o
>> >> >> cc -m64 -mcx16 -Ilibcommon.fa.p -I/usr/include/pixman-1 
>> >> >> -I/usr/include/libpng16 -I/usr/include/spice-server 
>> >> >> -I/usr/include/spice-1 -I/usr/include/cacard -I/usr/include/glib-2.0 
>> >> >> -I/usr/lib64/glib-2.0/include -I/usr/include/sysprof-4 
>> >> >> -I/usr/include/nss3 -I/usr/include/nspr4 -I/usr/include/PCSC 
>> >> >> -I/usr/include/p11-kit-1 -I/usr/include/libusb-1.0 -I/usr/include/SDL2 
>> >> >> -I/usr/include/libmount -I/usr/include/blkid 
>> >> >> -I/usr/include/gio-unix-2.0 -I/usr/include/slirp 
>> >> >> -fdiagnostics-color=auto -Wall -Winvalid-pch -Werror -std=gnu11 -O2 -g 
>> >> >> -isystem /mnt/code/qemu/full/linux-headers -isystem linux-headers 
>> >> >> -iquote . -iquote /mnt/code/qemu/full -iquote 
>> >> >> /mnt/code/qemu/full/include -iquote /mnt/code/qemu/full/tcg/i386 
>> >> >> -pthread -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2 -D_GNU_SOURCE 
>> >> >> -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -fno-strict-aliasing 
>> >> >> -fno-common -fwrapv -Wundef -Wwrite-strings -Wmissing-prototypes 
>> >> >> -Wstrict-prototypes -Wredundant-decls -Wold-style-declaration 
>> >> >> -Wold-style-definition -Wtype-limits -Wformat-security -Wformat-y2k 
>> >> >> -Winit-self -Wignored-qualifiers -Wempty-body -Wnested-externs 
>> >> >> -Wendif-labels -Wexpansion-to-defined -Wimplicit-fallthrough=2 
>> >> >> -Wmissing-format-attribute -Wno-missing-include-dirs 
>> >> >> -Wno-shift-negative-value -Wno-psabi -fstack-protector-strong -fPIE 
>> >> >> -D_REENTRANT -Wno-undef -DSTRUCT_IOVEC_DEFINED -MD -MQ 
>> >> >> libcommon.fa.p/migration_rdma.c.o -MF 
>> >> >> libcommon.fa.p/migration_rdma.c.o.d -o 
>> >> >> libcommon.fa.p/migration_rdma.c.o -c 
>> >> >> ../../../../mnt/code/qemu/full/migration/rdma.c
>> >> >> ../../../../mnt/code/qemu/full/migration/rdma.c: In function 
>> >> >> ‘qio_channel_rdma_class_init’:
>> >> >> ../../../../mnt/code/qemu/full/migration/rdma.c:4020:25: error: 
>> >> >> assignment to ‘ssize_t (*)(QIOChannel *, const struct iovec *, size_t, 
>> >> >>  int **, size_t *, int,  Error **)’ {aka ‘long int (*)(QIOChannel *, 
>> >> >> const struct iovec *, long unsigned int,  int **, long unsigned int *, 
>> >> >> int,  Error **)’} from incompatible pointer type ‘ssize_t 
>> >> >> (*)(QIOChannel *, const struct iovec *, size_t,  int **, size_t *, 
>> >> >> Error **)’ {aka ‘long int (*)(QIOChannel *, const struct iovec *, long 
>> >> >> unsigned int,  int **, long unsigned int *, Error **)’} 
>> >> >> [-Werror=incompatible-pointer-types]
>> >> >>  4020 |     ioc_klass->io_readv = qio_channel_rdma_readv;
>> >> >>       |                         ^
>> >> >> cc1: all warnings being treated as errors
>> >> >> 
>> >> >> And I don't really know how to fix it, because the problem is that rdma
>> >> >> don't use qio_channel_readv_full() at all.
>> >> >
>> >> > Likely qio_channel_rdma_readv just adds the 'int flags' param added.
>> >> > It doesn't need to actually do anything with the flags as they are
>> >> > checked before
>> >> 
>> >> I can do that.  That would fix the compilation issue.
>> >> 
>> >> But will rdma work?  Because it fakes a qio channel, so what is going to
>> >> implement the MSG_PEEK functionality for it?  It don't end calling
>> >> recv() at all.
>> >
>> > It is no problem - the qio_channel_readv method changes in this patch
>> > add:
>> >
>> > +    if ((flags & QIO_CHANNEL_READ_FLAG_MSG_PEEK) &&
>> > +        !qio_channel_has_feature(ioc, QIO_CHANNEL_FEATURE_READ_MSG_PEEK)) 
>> > {
>> > +        error_setg_errno(errp, EINVAL,
>> > +                         "Channel does not support peek read");
>> > +        return -1;
>> > +    }
>> >
>> >
>> > so it is impossible for qio_channel_rdma_readv to be invoked with
>> > flags having MSG_PEEK set, thus RDMA can ignore the whole concept.
>> 
>> And as we require MSG_PEEK to do migration, we have lost RDMA migration
>> in the process.
>> 
>> The following patch on the series use this functionality to read the
>> beggining of the streams in the channels.
>
> It guards that usage of MSG_PEEK with
>
>    if (... && qio_channel_has_feature(ioc, 
> QIO_CHANNEL_FEATURE_READ_MSG_PEEK)) {

You win.

They are back.

Thanks very much for the explanation.