We should move the test a bit down, after the request_module() attempts.
Otherwise, when using modules in combination with kmod, on the first open()
the check will return ENXIO unconditionally rather than loading the IDE
disk driver, as nr_sectors will be 0 since the ide-disk module was not
initialized yet.
Gadi
On Sun, 16 May 1999 Andries.Brouwer@cwi.nl wrote:
> The SCSI code (sd_open) returns -ENXIO if one tries to open
> a nonexistent partition of an exiting disk. Good.
> But the IDE code just succeeds in the corresponding
> situation. This means that attempts to mount a nonexistent
> partition yield nonsense like
>
> May 16 14:20:03 nata kernel: 03:0c: rw=0, want=1, limit=0
> May 16 14:20:03 nata kernel: attempt to access beyond end of device
> May 16 14:20:03 nata kernel: hda<: bad access: block=0, count=1
> May 16 14:20:03 nata kernel: end_request: I/O error, dev 03:0c (hda), sector 0
>
> in the syslog. The patch below makes ide_open fail in this situation.
> (Maybe the surrounding `if (drive->media == ide_disk)' is too
> cautious - I have not thought about what other IDE devices there
> are, and what meaning the minor could have for them.)
>
> Andries
>
>
> --- ide.c~ Sat May 15 21:13:29 1999
> +++ ide.c Sun May 16 14:37:23 1999
> @@ -1672,7 +1672,12 @@
> int rc;
>
> if ((drive = get_info_ptr(inode->i_rdev)) == NULL)
> - return -ENXIO;
> + return -ENXIO; /* no such drive */
> + if (drive->media == ide_disk) {
> + int minor = (MINOR(inode->i_rdev) & PARTN_MASK);
> + if (drive->part[minor].nr_sects == 0)
> + return -ENXIO; /* no such partition */
> + }
> MOD_INC_USE_COUNT;
> if (drive->driver == NULL)
> ide_init_module(IDE_DRIVER_MODULE);
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/