openwrt/target/linux/generic/pending-5.4/412-mtd-partial_eraseblock_unlock.patch
John Thomson bf2870c1d9 kernel: fix mtd partition erase < parent_erasesize writes
This bug applied where mtd partition end address,
or erase start address, was not cleanly divisible by parent mtd erasesize.

This error would cause the bits following the end of the partition
to the next erasesize block boundary to be erased,
and this partition-overflow data to be written to the partition erase
address (missing additional partition offset address)
of the mtd (top) parent device.

Fixes: FS#2428

Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au>
Tested-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
[shorten commit title, add Fixes, fix kernel 4.19 as well]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
2020-09-19 12:41:29 +02:00

41 lines
1.4 KiB
Diff

From: Tim Harvey <tharvey@gateworks.com>
Subject: mtd: allow partial block unlock
This allows sysupgrade for devices such as the Gateworks Avila/Cambria
product families based on the ixp4xx using the redboot bootloader with
combined FIS directory and RedBoot config partitions on larger FLASH
devices with larger eraseblocks.
This second iteration of this patch addresses previous issues:
- whitespace breakage fixed
- unlock in all scenarios
- simplification and fix logic bug
[john@phrozen.org: this should be moved to the ixp4xx folder]
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
---
drivers/mtd/mtdpart.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
--- a/drivers/mtd/mtdpart.c
+++ b/drivers/mtd/mtdpart.c
@@ -293,7 +293,16 @@ static int part_lock(struct mtd_info *mt
static int part_unlock(struct mtd_info *mtd, loff_t ofs, uint64_t len)
{
struct mtd_part *part = mtd_to_part(mtd);
- return part->parent->_unlock(part->parent, ofs + part->offset, len);
+
+ ofs += part->offset;
+
+ if (mtd->flags & MTD_ERASE_PARTIAL) {
+ /* round up len to next erasesize and round down offset to prev block */
+ len = (mtd_div_by_eb(len, part->parent) + 1) * part->parent->erasesize;
+ ofs &= ~(part->parent->erasesize - 1);
+ }
+
+ return part->parent->_unlock(part->parent, ofs, len);
}
static int part_is_locked(struct mtd_info *mtd, loff_t ofs, uint64_t len)