Re: [RFC PATCH-for-5.1] hw/ide: Avoid #DIV/0! FPU exception by setting CD-ROM sector count

[Philippe Mathieu-Daudé]

On 7/17/20 12:16 PM, Philippe Mathieu-Daudé wrote:
> libFuzzer found an undefined behavior (#DIV/0!) in ide_set_sector()
> when using a CD-ROM (reproducer available on the BugLink):
> 
>   UndefinedBehaviorSanitizer:DEADLYSIGNAL
>   ==12163==ERROR: UndefinedBehaviorSanitizer: FPE on unknown address 
> 0x5616279cffdc (pc 0x5616279cffdc bp 0x7ffcdaabae90 sp 0x7ffcdaabae30 T12163)
> 
> Fix by initializing the CD-ROM number of sectors in ide_dev_initfn().
> 
> Reported-by: Alexander Bulekov <alxndr@bu.edu>
> Fixes: b2df431407 ("ide scsi: Mess with geometry only for hard disk devices")
> BugLink: https://bugs.launchpad.net/qemu/+bug/1887309
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> ---
>  hw/ide/qdev.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/hw/ide/qdev.c b/hw/ide/qdev.c
> index 27ff1f7f66..6ce7fc317c 100644
> --- a/hw/ide/qdev.c
> +++ b/hw/ide/qdev.c
> @@ -201,6 +201,15 @@ static void ide_dev_initfn(IDEDevice *dev, IDEDriveKind 
> kind, Error **errp)
>                                errp)) {
>              return;
>          }
> +    } else {
> +        uint64_t nb_sectors;
> +
> +        blk_get_geometry(dev->conf.blk, &nb_sectors);
> +        if (!nb_sectors) {
> +            error_setg(errp, "CD-ROM size can not be zero");

Hmm this doesn't work because for some machines configure_blockdev()
creates blocks of zero size under your feet:

    default_drive(default_cdrom, snapshot,
machine_class->block_default_type, 2,
                  CDROM_OPTS);

> +            return;
> +        }
> +        dev->conf.secs = nb_sectors;
>      }
>      if (!blkconf_apply_backend_options(&dev->conf, kind == IDE_CD,
>                                         kind != IDE_CD, errp)) {
>