mtd: Fix partition alignment check on multi-erasesize devices
authorBoris Brezillon <boris.brezillon@free-electrons.com>
Mon, 25 Sep 2017 08:19:57 +0000 (10:19 +0200)
committerBoris Brezillon <boris.brezillon@free-electrons.com>
Wed, 27 Sep 2017 11:53:20 +0000 (13:53 +0200)
Commit 1eeef2d7483a ("mtd: handle partitioning on devices with 0
erasesize") introduced a regression on heterogeneous erase region
devices. Alignment of the partition was tested against the master
eraseblock size which can be bigger than the slave one, thus leading
to some partitions being marked as read-only.

Update wr_alignment to match this slave erasesize after this erasesize
has been determined by picking the biggest erasesize of all the regions
embedded in the MTD partition.

Reported-by: Mathias Thore <Mathias.Thore@infinera.com>
Fixes: 1eeef2d7483a ("mtd: handle partitioning on devices with 0 erasesize")
Cc: <stable@vger.kernel.org>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Tested-by: Mathias Thore <Mathias.Thore@infinera.com>
Reviewed-by: Mathias Thore <Mathias.Thore@infinera.com>
drivers/mtd/mtdpart.c

index 5736b0c90b339b6bc3e8e1ae3189341910b958f2..a308e707392d595902b77a03e2078ffe8da97d9e 100644 (file)
@@ -581,6 +581,14 @@ static struct mtd_part *allocate_partition(struct mtd_info *parent,
                slave->mtd.erasesize = parent->erasesize;
        }
 
+       /*
+        * Slave erasesize might differ from the master one if the master
+        * exposes several regions with different erasesize. Adjust
+        * wr_alignment accordingly.
+        */
+       if (!(slave->mtd.flags & MTD_NO_ERASE))
+               wr_alignment = slave->mtd.erasesize;
+
        tmp = slave->offset;
        remainder = do_div(tmp, wr_alignment);
        if ((slave->mtd.flags & MTD_WRITEABLE) && remainder) {