| Message ID | 1504001833-18097-1-git-send-email-LW@KARO-electronics.de |
|---|---|
| Headers | show |
| Series | mtd: nand: fix regression introduced by splitting off manufacturer dependent code | expand |
On Tue, 29 Aug 2017 12:17:11 +0200 Lothar Waßmann <LW@KARO-electronics.de> wrote: > commit c51d0ac59f24 ("mtd: nand: Move Samsung specific init/detection > logic in nand_samsung.c") introduced a regression for Samsung SLC NAND > chips by skipping the initialization of chip->bits_per_cell that is > done in nand_decode_ext_id() from which the manufacturer dependent > code was extracted. > The regression should also affect Hynix and Macronix chips whose code > was separated out in further commits but which I cannot test. > AMD/Spansion and Toshiba NAND are not affected, since they are calling > nand_decode_ext_id() (which initializes bhip->bits_per_cell) in their > .detect function. > > Fix the regression and add a warning to nand_is_slc() to prevent > further regressions of this kind. Applied both to nand/next and generated a new PR to push this in 4.14. Also added the Fixes and Cc-stable tag to the first patch. Thanks, Boris
commit c51d0ac59f24 ("mtd: nand: Move Samsung specific init/detection logic in nand_samsung.c") introduced a regression for Samsung SLC NAND chips by skipping the initialization of chip->bits_per_cell that is done in nand_decode_ext_id() from which the manufacturer dependent code was extracted. The regression should also affect Hynix and Macronix chips whose code was separated out in further commits but which I cannot test. AMD/Spansion and Toshiba NAND are not affected, since they are calling nand_decode_ext_id() (which initializes bhip->bits_per_cell) in their .detect function. Fix the regression and add a warning to nand_is_slc() to prevent further regressions of this kind.