Re: [Qemu-devel] [PATCH] block/vdi: Zero unused parts when allocating a new block (fix #919242)

[Stefan Weil <sw@weilnetz.de>]

Am 21.01.2012 13:54, schrieb Stefan Weil:
> The new block was filled with zero when it was allocated by g_malloc0,
> but when it was reused later and only partially used, data from the
> previously allocated block were still present and written to the new
> block.
>
> This caused the problems reported by bug #919242
> (https://bugs.launchpad.net/qemu/+bug/919242).
>
> Now the unused parts of the new block which are before and after the data
> are always filled with zero, so it is no longer necessary to zero the whole
> block with g_malloc0.
>
> I also updated the copyright comment.
>
> Signed-off-by: Stefan Weil<sw@weilnetz.de>
> ---
>   block/vdi.c |    8 ++++++--
>   1 files changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/block/vdi.c b/block/vdi.c
> index 31cdfab..6a0011f 100644
> --- a/block/vdi.c
> +++ b/block/vdi.c
> @@ -1,7 +1,7 @@
>   /*
>    * Block driver for the Virtual Disk Image (VDI) format
>    *
> - * Copyright (c) 2009 Stefan Weil
> + * Copyright (c) 2009, 2012 Stefan Weil
>    *
>    * This program is free software: you can redistribute it and/or modify
>    * it under the terms of the GNU General Public License as published by
> @@ -756,15 +756,19 @@ static void vdi_aio_write_cb(void *opaque, int ret)
>                    (uint64_t)bmap_entry * s->block_sectors;
>           block = acb->block_buffer;
>           if (block == NULL) {
> -            block = g_malloc0(s->block_size);
> +            block = g_malloc(s->block_size);
>               acb->block_buffer = block;
>               acb->bmap_first = block_index;
>               assert(!acb->header_modified);
>               acb->header_modified = 1;
>           }
>           acb->bmap_last = block_index;
> +        /* Copy data to be written to new block and zero unused parts. */
> +        memset(block, 0, sector_in_block * SECTOR_SIZE);
>           memcpy(block + sector_in_block * SECTOR_SIZE,
>                  acb->buf, n_sectors * SECTOR_SIZE);
> +        memset(block + (sector_in_block + n_sectors) * SECTOR_SIZE, 0,
> +               (s->block_sectors - n_sectors - sector_in_block) * SECTOR_SIZE);
>           acb->hd_iov.iov_base = (void *)block;
>           acb->hd_iov.iov_len = s->block_size;
>           qemu_iovec_init_external(&acb->hd_qiov,&acb->hd_iov, 1);
>    

Hi,

this patch should also be applied to the stable branches of QEMU,
at least to stable-1.0 (after the review, of course).

Regards,
Stefan Weil