On 10/19/2018 12:02 PM, Boris Brezillon wrote: > Commit 5390a8df769e ("mtd: spi-nor: add support to non-uniform SFDP SPI > NOR flash memories") removed the 'nor->addr_width = 0;' statement when > spi_nor_parse_sfdp() returns an error, thus leaving ->addr_width in an > undefined state which can cause trouble when spi_nor_scan() checks its > value. > > Reported-by: Cyrille Pitchen <cyrille.pitchen@xxxxxxxxxx> > Fixes: 5390a8df769e ("mtd: spi-nor: add support to non-uniform SFDP SPI NOR flash memories") > Signed-off-by: Boris Brezillon <boris.brezillon@xxxxxxxxxxx> Reviewed-by: Tudor Ambarus <tudor.ambarus@xxxxxxxxxxxxx> Commit 5390a8df769e removed the following: @@ -2521,20 +2962,20 @@ static int spi_nor_init_params(struct spi_nor *nor, - /* Override the parameters with data read from SFDP tables. */ - nor->addr_width = 0; - nor->mtd.erasesize = 0; this is good because nor is allocated with _kzalloc by all it's callers. Should we add a comment when declaring spi_nor structure or it's widely known that it should be initialized with zeros? [cut] - if (spi_nor_parse_sfdp(nor, &sfdp_params)) { - nor->addr_width = 0; removal of nor->addr_width = 0; is bad, and must be reintroduced because nor->addr_width is modified inside spi_nor_parse_sfdp() - nor->mtd.erasesize = 0; this removal is good, because we don't overwrite nor->mtd.erasesize in spi_nor_parse_sfdp() I should have sent this changes in a separate patch, sorry. Thanks! ta > --- > drivers/mtd/spi-nor/spi-nor.c | 6 ++++-- > 1 file changed, 4 insertions(+), 2 deletions(-) > > diff --git a/drivers/mtd/spi-nor/spi-nor.c b/drivers/mtd/spi-nor/spi-nor.c > index 9407ca5f9443..3e54e31889c7 100644 > --- a/drivers/mtd/spi-nor/spi-nor.c > +++ b/drivers/mtd/spi-nor/spi-nor.c > @@ -3250,12 +3250,14 @@ static int spi_nor_init_params(struct spi_nor *nor, > memcpy(&sfdp_params, params, sizeof(sfdp_params)); > memcpy(&prev_map, &nor->erase_map, sizeof(prev_map)); > > - if (spi_nor_parse_sfdp(nor, &sfdp_params)) > + if (spi_nor_parse_sfdp(nor, &sfdp_params)) { > + nor->addr_width = 0; > /* restore previous erase map */ > memcpy(&nor->erase_map, &prev_map, > sizeof(nor->erase_map)); > - else > + } else { > memcpy(params, &sfdp_params, sizeof(*params)); > + } > } > > return 0; > ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/