qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCHv5] block: add native support for NFS


From: Peter Lieven
Subject: Re: [Qemu-devel] [PATCHv5] block: add native support for NFS
Date: Fri, 10 Jan 2014 19:47:27 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0

Am 10.01.2014 19:24, schrieb Paolo Bonzini:
> Il 10/01/2014 19:07, Peter Lieven ha scritto:
>>
>>
>> Von meinem iPad gesendet
>>
>> Am 10.01.2014 um 19:05 schrieb "Paolo Bonzini" <address@hidden>:
>>
>>> Il 10/01/2014 18:16, ronnie sahlberg ha scritto:
>>>> There is a common exception though, for the case where you read past
>>>> the end of file.
>>>> So short reads should normally not happen. Unless QEMU or the guest
>>>> sends a request to libnfs to read past the end of the file.
>>> Yes, this can happen in QEMU and the various drivers are careful to pad
>>> with zeroes.  It could perhaps be moved to block.c, but for now each
>>> driver handles it separately.
>> ok i will add this as well. however, i thought i had seen code for this in 
>> block.c  already?,
> No, it corresponds to this code in block/raw-posix.c:
>
> static int aio_worker(void *arg)
> {
>     RawPosixAIOData *aiocb = arg;
>     ssize_t ret = 0;
>
>     switch (aiocb->aio_type & QEMU_AIO_TYPE_MASK) {
>     case QEMU_AIO_READ:
>         ret = handle_aiocb_rw(aiocb);
>         if (ret >= 0 && ret < aiocb->aio_nbytes && aiocb->bs->growable) {
>             iov_memset(aiocb->aio_iov, aiocb->aio_niov, ret,
>                       0, aiocb->aio_nbytes - ret);
>
>             ret = aiocb->aio_nbytes;
>         }
>         if (ret == aiocb->aio_nbytes) {
>             ret = 0;
>         } else if (ret >= 0 && ret < aiocb->aio_nbytes) {
>             ret = -EINVAL;
>         }
>         break;

I am a little confused... but it seems what I had in mind just fills up full 
sectors....?!

    if (!(bs->zero_beyond_eof && bs->growable)) {
        ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
    } else {
        /* Read zeros after EOF of growable BDSes */
        int64_t len, total_sectors, max_nb_sectors;

        len = bdrv_getlength(bs);
        if (len < 0) {
            ret = len;
            goto out;
        }

        total_sectors = DIV_ROUND_UP(len, BDRV_SECTOR_SIZE);
        max_nb_sectors = MAX(0, total_sectors - sector_num);
        if (max_nb_sectors > 0) {
            ret = drv->bdrv_co_readv(bs, sector_num,
                                     MIN(nb_sectors, max_nb_sectors), qiov);
        } else {
            ret = 0;
        }

        /* Reading beyond end of file is supposed to produce zeroes */
        if (ret == 0 && total_sectors < sector_num + nb_sectors) {
            uint64_t offset = MAX(0, total_sectors - sector_num);
            uint64_t bytes = (sector_num + nb_sectors - offset) *
                              BDRV_SECTOR_SIZE;
            qemu_iovec_memset(qiov, offset * BDRV_SECTOR_SIZE, 0, bytes);
        }
    }

Peter



>
> Paolo




reply via email to

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