Messages in this thread | | | Date | Sat, 7 Aug 2021 12:32:43 +0200 | From | Miquel Raynal <> | Subject | Re: [PATCH 1/2] mtd: mtdconcat: Judge callback function existence getting from master for each partition |
| |
Hi Zhihao,
Zhihao Cheng <chengzhihao1@huawei.com> wrote on Sat, 7 Aug 2021 10:15:46 +0800:
> 在 2021/8/7 3:28, Miquel Raynal 写道: > Hi Miquel, > > Hi Zhihao, > > > > Zhihao Cheng <chengzhihao1@huawei.com> wrote on Sat, 31 Jul 2021 > > 10:32:42 +0800: > > @@ -721,14 +724,15 @@ struct mtd_info *mtd_concat_create(struct mtd_info *subdev[], /* subdevices to c > > subdev[i]->flags & MTD_WRITEABLE; > > } > > > + subdev_master = mtd_get_master(subdev[i]); > > concat->mtd.size += subdev[i]->size; > > concat->mtd.ecc_stats.badblocks += > > subdev[i]->ecc_stats.badblocks; > > if (concat->mtd.writesize != subdev[i]->writesize || > > concat->mtd.subpage_sft != subdev[i]->subpage_sft || > > concat->mtd.oobsize != subdev[i]->oobsize || > > - !concat->mtd._read_oob != !subdev[i]->_read_oob || > > - !concat->mtd._write_oob != !subdev[i]->_write_oob) { > > + !concat->mtd._read_oob != !subdev_master->_read_oob || > > + !concat->mtd._write_oob != !subdev_master->_write_oob) { > > Do you really need this change? > > I think both "!concat->mtd._read_oob != !subdev[i]->_read_oob" and "!concat->mtd._write_oob != !subdev[i]->_write_oob" need to be modified otherwise concatenating goes failure. > > I thought there exists two problems: > > 1. Wrong callback fetching in mtd partition device > > 2. Warning for existence of _read and _read_oob at the same time > > so I solved them in two steps to make history commit logs a bit clear. > > Though these two patches can be combined to one.
No please keep the split.
What I mean here is that I don't think your fix is valid. Maybe we should propagate these callbacks as well instead of trying to hack into this condition. I don't see why you should check against subdev[i] for half of the callbacks and check for subdev_master for the last two.
Thanks, Miquèl
| |