On 31.01.2013 13:24, Andrey Borzenkov wrote: > Fix autodetection of $prefix when booted from EFI CD-ROM. > > Based on patch from Matthew Garrett, modified to not overwrite > memory returned by device path protocol handler and rebased to > current trunk. > > Additionally fixes potential memory leak - dup_dp was not deallocated > if device path was not found.
Why any special treatement for CD-ROMs? They can have partitions and do
so on CD-ROMs intended for macs. What about following patch:
phcoder@debian.x201.phnet:15:02:18:~/grub2/bzr/mainline$ bzr diff
=== modified file 'grub-core/disk/efi/efidisk.c'
--- grub-core/disk/efi/efidisk.c 2013-04-05 08:59:26 +0000
+++ grub-core/disk/efi/efidisk.c 2013-04-05 13:07:39 +0000
@@ -633,9 +633,6 @@
return d->handle;
case 'c':
- /* FIXME: probably this is not correct. */
- return d->handle;
-
case 'h':
/* If this is the whole disk, just return its own data. */
if (! disk->partition)
@@ -657,7 +654,9 @@
if ((GRUB_EFI_DEVICE_PATH_TYPE (c->last_device_path)
== GRUB_EFI_MEDIA_DEVICE_PATH_TYPE)
&& (GRUB_EFI_DEVICE_PATH_SUBTYPE (c->last_device_path)
- == GRUB_EFI_HARD_DRIVE_DEVICE_PATH_SUBTYPE)
+ == GRUB_EFI_HARD_DRIVE_DEVICE_PATH_SUBTYPE
+ || GRUB_EFI_DEVICE_PATH_SUBTYPE (c->last_device_path)
+ == GRUB_EFI_CDROM_DEVICE_PATH_SUBTYPE)
&& (grub_partition_get_start (disk->partition)
== (hd.partition_start << (disk->log_sector_size
- GRUB_DISK_SECTOR_BITS)))
@@ -770,7 +769,9 @@
if (GRUB_EFI_DEVICE_PATH_TYPE (ldp) == GRUB_EFI_MEDIA_DEVICE_PATH_TYPE
&& (GRUB_EFI_DEVICE_PATH_SUBTYPE (ldp)
- == GRUB_EFI_HARD_DRIVE_DEVICE_PATH_SUBTYPE))
+ == GRUB_EFI_HARD_DRIVE_DEVICE_PATH_SUBTYPE
+ || GRUB_EFI_DEVICE_PATH_SUBTYPE (ldp)
+ == GRUB_EFI_CDROM_DEVICE_PATH_SUBTYPE))
{
struct grub_efidisk_get_device_name_ctx ctx;
char *dev_name;
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel
