| Date | Fri, 3 May 2013 22:27:07 +0200 | Subject | Re: [PATCH 20/63] dmaengine: ste_dma40: Remove unnecessary call to d40_phy_cfg() | From | Linus Walleij <> |
| |
On Fri, May 3, 2013 at 4:32 PM, Lee Jones <lee.jones@linaro.org> wrote:
> All configuration left in d40_phy_cfg() is runtime configurable and > there is already a call into it from d40_runtime_config(), so let's > rely on that. > > Acked-by: Vinod Koul <vnod.koul@intel.com> > Acked-by: Arnd Bergmann <arnd@arndb.de> > Signed-off-by: Lee Jones <lee.jones@linaro.org> (...)
> @@ -2027,6 +2027,14 @@ static int d40_config_memcpy(struct d40_chan *d40c) > } else if (dma_has_cap(DMA_MEMCPY, cap) && > dma_has_cap(DMA_SLAVE, cap)) { > d40c->dma_cfg = dma40_memcpy_conf_phy; > + > + /* Generate interrrupt at end of transfer or relink. */ > + d40c->dst_def_cfg |= BIT(D40_SREG_CFG_TIM_POS); > + > + /* Generate interrupt on error. */ > + d40c->src_def_cfg |= BIT(D40_SREG_CFG_EIM_POS); > + d40c->dst_def_cfg |= BIT(D40_SREG_CFG_EIM_POS); > +
This hunk looks like it's fixing a bug introduced in patch 19/63.
Do you try to run a memcpy test after patch 19?
Breaking the drive in one patch and fixing it in the next is a no-no because of bisection.
Maybe things work fine if you just move this hunk of the patch over to 19/63?
Apart from this the patch looks fine.
Yours, Linus Walleij
|