On Tue, 2022-10-11 at 11:21 +0900, Daejun Park wrote: > Hi Bean Huo, > > I think ufs_is_valid_unit_desc_lun() is also used for > wb_buf_alloc_units_show() in ufs-sysfs.c. > So just removing this if-checkup will make different result when > check lun value. > Hi Daejun, Thanks for your review on the patch. Yes, I understood what you mean. But I don't think that's the problem. Without this patch, access on sysfs node "wb_shared_alloc_units" would get "invalid argument", while with this patch sysfs would return 00. According to the UFS specification: "If this value is ‘0’, then the WriteBooster is not supported for this LU. The Logical unit among LU0 ~ LU7 can be configured for WriteBooster Buffer. Otherwise, whole WriteBooster Buffer configuration in this device is invalid." Per my understanding, with this patch, there is still no miss- explanation of this sysfs node. The key purpose of this patch is to remove any nonsense logical during the booting stage. please have a think my comments. thanks. Kind regards, Bean > Thanks, > Daejun > > > From: Bean Huo <beanhuo@xxxxxxxxxx> > > > > LUs with WB potential support are properly checked in ufshcd_wb_pro > > be() > > before calling ufshcd_read_unit_desc_param(), so remove this unnece > > ssary > > if- > > checkup in ufs_is_valid_unit_desc_lun() to match its function defin > > ition.
