Re: [PATCH v5 06/14] migration/qemu-file: Add qemu_file_get_to_fd()

[Avihai Horon <avihaih@nvidia.com>]

On 09/01/2023 13:20, Cédric Le Goater wrote:
> External email: Use caution opening links or attachments
>
>
> On 12/29/22 12:03, Avihai Horon wrote:
>> Add new function qemu_file_get_to_fd() that allows reading data from
>> QEMUFile and writing it straight into a given fd.
>>
>> This will be used later in VFIO migration code.
>>
>> Signed-off-by: Avihai Horon <avihaih@nvidia.com>
>> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
>> ---
>>   migration/qemu-file.h |  1 +
>>   migration/qemu-file.c | 34 ++++++++++++++++++++++++++++++++++
>>   2 files changed, 35 insertions(+)
>>
>> diff --git a/migration/qemu-file.h b/migration/qemu-file.h
>> index fa13d04d78..9d0155a2a1 100644
>> --- a/migration/qemu-file.h
>> +++ b/migration/qemu-file.h
>> @@ -148,6 +148,7 @@ int qemu_file_shutdown(QEMUFile *f);
>>   QEMUFile *qemu_file_get_return_path(QEMUFile *f);
>>   void qemu_fflush(QEMUFile *f);
>>   void qemu_file_set_blocking(QEMUFile *f, bool block);
>> +int qemu_file_get_to_fd(QEMUFile *f, int fd, size_t size);
>>
>>   void ram_control_before_iterate(QEMUFile *f, uint64_t flags);
>>   void ram_control_after_iterate(QEMUFile *f, uint64_t flags);
>> diff --git a/migration/qemu-file.c b/migration/qemu-file.c
>> index 2d5f74ffc2..79303c9d34 100644
>> --- a/migration/qemu-file.c
>> +++ b/migration/qemu-file.c
>> @@ -940,3 +940,37 @@ QIOChannel *qemu_file_get_ioc(QEMUFile *file)
>>   {
>>       return file->ioc;
>>   }
>> +
>> +/*
>> + * Read size bytes from QEMUFile f and write them to fd.
>> + */
>> +int qemu_file_get_to_fd(QEMUFile *f, int fd, size_t size)
>> +{
>> +    while (size) {
>> +        size_t pending = f->buf_size - f->buf_index;
>> +        ssize_t rc;
>> +
>> +        if (!pending) {
>> +            rc = qemu_fill_buffer(f);
>> +            if (rc < 0) {
>> +                return rc;
>> +            }
>> +            if (rc == 0) {
>> +                return -1;
>
> Given the call stack, -1 would be interpreted  as EPERM. May be EIO 
> instead ?
>
Sure.

I will also change write (down below) to return -errno instead of rc on 
error.

> C.
>
>> +            }
>> +            continue;
>> +        }
>> +
>> +        rc = write(fd, f->buf + f->buf_index, MIN(pending, size));
>> +        if (rc < 0) {
>> +            return rc;
>> +        }
>> +        if (rc == 0) {
>> +            return -1;
>> +        }
>> +        f->buf_index += rc;
>> +        size -= rc;
>> +    }
>> +
>> +    return 0;
>> +}
>