Syslinux - Jul 2010 - chain.c: fix v2 for public indexes in iterators (for master and disklib)

>From the earlier patches - that simple fix for indexes not counting non-data
partitions in gpt and mbr cases. Patches don't touch any other parts of the
code at this stage - bigger stuff later for disklib as we agreed - but this is
potentially useful right now.

Michal Soltys (1):
  chain.c: fix public index value in mbr and gpt iterators

mbr and gpt iterators skip empty / non-data partitions properly, but don't
update publicly visible index. With this patch, such partitions are counted as
well.

Signed-off-by: Michal Soltys <soltys at ziu.info>
---
 com32/modules/chain.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/com32/modules/chain.c b/com32/modules/chain.c
index 5078ed3..60e3abc 100644
--- a/com32/modules/chain.c
+++ b/com32/modules/chain.c
@@ -335,8 +335,8 @@ static struct disk_part_iter *next_mbr_part(struct
disk_part_iter *part)
 
     /* Update parameters to reflect this new partition.  Re-use iterator */
     part->lba_data = table[part->private.mbr_index].start_lba;
-    dprintf("Partition %d primary lba %u\n", part->index,
part->lba_data);
-    part->index++;
+    dprintf("Partition %d primary lba %u\n",
part->private.mbr_index, part->lba_data);
+    part->index = part->private.mbr_index + 1;
     part->record = table + part->private.mbr_index;
     return part;
 
@@ -369,7 +369,7 @@ static struct disk_part_iter *next_gpt_part(struct
disk_part_iter *part)
     part->private.gpt.part_guid = &gpt_part->uid;
     part->private.gpt.part_label = gpt_part->name;
     /* Update our index */
-    part->index++;
+    part->index = part->private.gpt.index + 1;
 #ifdef DEBUG
     disk_gpt_part_dump(gpt_part);
 #endif
-- 
1.6.3.1

mbr and gpt iterators skip empty / non-data partitions properly, but don't
update publicly visible index. With this patch, such partitions are counted as
well.

Signed-off-by: Michal Soltys <soltys at ziu.info>
---
 com32/modules/chain.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/com32/modules/chain.c b/com32/modules/chain.c
index 48a83d2..a5f6401 100644
--- a/com32/modules/chain.c
+++ b/com32/modules/chain.c
@@ -624,8 +624,8 @@ static struct disk_part_iter *next_mbr_part(struct
disk_part_iter *part)
 
     /* Update parameters to reflect this new partition.  Re-use iterator */
     part->lba_data = table[part->private.mbr_index].start_lba;
-    dprintf("Partition %d primary lba %u\n", part->index,
part->lba_data);
-    part->index++;
+    dprintf("Partition %d primary lba %u\n",
part->private.mbr_index, part->lba_data);
+    part->index = part->private.mbr_index + 1;
     part->record = table + part->private.mbr_index;
     return part;
 
@@ -869,7 +869,7 @@ static struct disk_part_iter *next_gpt_part(struct
disk_part_iter *part)
     part->private.gpt.part_guid = &gpt_part->uid;
     part->private.gpt.part_label = gpt_part->name;
     /* Update our index */
-    part->index++;
+    part->index = part->private.gpt.index + 1;
     gpt_part_dump(gpt_part);
 
     /* In a GPT scheme, we re-use the iterator */
-- 
1.6.3.1

Thanks again, Michal!  Matching what indices Linux uses is what I think
H. Peter expressed as desired behaviour all along.  - Shao

On 07/30/2010 12:52 AM, Miller, Shao wrote:
> Thanks again, Michal!  Matching what indices Linux uses is what I think
> H. Peter expressed as desired behaviour all along.  - Shao
Indeed.  I wonder if this is the problem Alec Moskvin is seeing.

	-hpa