On Wed, Aug 04, 2021 at 01:01:36PM +0300, Lauri Kasanen wrote: > > diff --git a/drivers/block/n64cart.c b/drivers/block/n64cart.c > > index 7b4dd10af9ec..c84be0028f63 100644 > > --- a/drivers/block/n64cart.c > > +++ b/drivers/block/n64cart.c > > @@ -74,7 +74,7 @@ static bool n64cart_do_bvec(struct device *dev, struct bio_vec *bv, u32 pos) > > > > n64cart_wait_dma(); > > > > - n64cart_write_reg(PI_DRAM_REG, dma_addr + bv->bv_offset); > > + n64cart_write_reg(PI_DRAM_REG, dma_addr); > > n64cart_write_reg(PI_CART_REG, (bstart | CART_DOMAIN) & CART_MAX); > > n64cart_write_reg(PI_WRITE_REG, bv->bv_len - 1); > > Hm, then how did it work? Does it always happen to be zero? It isn't always zero, but for your use cases it might. That is if you have a page aligned buffer (e.g. the page cache) it tends to be zero. > Have you tested this? I don't have the equipment currently. No, I don't. I'm just auditing uses of bv_offset and this one is clearly bogus.
