* [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast
@ 2024-07-24 9:47 Ahmad Fatoum
2024-07-24 9:47 ` [PATCH master 2/2] usb: gadget: fsl_udc: dma_map buffers instead of using dma_sync Ahmad Fatoum
2024-07-30 8:13 ` [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Sascha Hauer
0 siblings, 2 replies; 3+ messages in thread
From: Ahmad Fatoum @ 2024-07-24 9:47 UTC (permalink / raw)
To: barebox; +Cc: ske, Ahmad Fatoum
All other prints in the file use the DMA address except for this single
print inside debug_dma_map(). Let's use the DMA address there as well to
fix the warning about casting a 32-bit pointer to a 64 bit integer and
to align error messages in case of non-1:1 mapped DMA addresses.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
drivers/dma/debug.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/dma/debug.c b/drivers/dma/debug.c
index 32a417504479..befe837d6cd8 100644
--- a/drivers/dma/debug.c
+++ b/drivers/dma/debug.c
@@ -131,7 +131,7 @@ void debug_dma_map(struct device *dev, void *addr,
if (!IS_ALIGNED(dev_addr, DMA_ALIGNMENT))
dma_dev_warn(dev, "Mapping insufficiently aligned %s buffer 0x%llx+0x%zx: %u bytes required!\n",
- dir2name[direction], (u64)addr, size, DMA_ALIGNMENT);
+ dir2name[direction], (u64)dev_addr, size, DMA_ALIGNMENT);
}
void debug_dma_unmap(struct device *dev, dma_addr_t addr,
--
2.39.2
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH master 2/2] usb: gadget: fsl_udc: dma_map buffers instead of using dma_sync
2024-07-24 9:47 [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Ahmad Fatoum
@ 2024-07-24 9:47 ` Ahmad Fatoum
2024-07-30 8:13 ` [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Sascha Hauer
1 sibling, 0 replies; 3+ messages in thread
From: Ahmad Fatoum @ 2024-07-24 9:47 UTC (permalink / raw)
To: barebox; +Cc: ske, Ahmad Fatoum
CONFIG_DMA_API_DEBUG depends on DMA buffers being mapped with
dma_map_single before ownership is switched with dma_sync_single.
As this is currently not done, CONFIG_DMA_API_DEBUG reports a lot of
false positives when the driver is enabled. Fix it by using
usb_gadget_map_request which internally calls dma_map_single.
This fixes the false positives, aligns us with what the Linux driver
is doing and reduces overhead when unmapping the buffer for an IN
target endpoint:
Previously with DMA_BIDIRECTIONAL, we always invalidated buffers
when taking back ownership for the CPU. This is unnecessary when
unmapping a buffer for an IN endpoint as the direction is DMA_TO_DEVICE
and we won't be reading back the result.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
drivers/usb/gadget/udc/fsl_udc.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/usb/gadget/udc/fsl_udc.c b/drivers/usb/gadget/udc/fsl_udc.c
index 41de44b30d12..e1d0d0b48272 100644
--- a/drivers/usb/gadget/udc/fsl_udc.c
+++ b/drivers/usb/gadget/udc/fsl_udc.c
@@ -197,8 +197,7 @@ static void done(struct fsl_ep *ep, struct fsl_req *req, int status)
dma_free_coherent(curr_td, 0, sizeof(struct ep_td_struct));
}
- dma_sync_single_for_cpu(udc->gadget.dev.parent, (unsigned long)req->req.buf,
- req->req.length, DMA_BIDIRECTIONAL);
+ usb_gadget_unmap_request(&udc->gadget, &req->req, ep_is_in(ep));
if (status && (status != -ESHUTDOWN))
VDBG("complete %s req %p stat %d len %u/%u",
@@ -861,7 +860,7 @@ fsl_ep_queue(struct usb_ep *_ep, struct usb_request *_req)
struct fsl_ep *ep = container_of(_ep, struct fsl_ep, ep);
struct fsl_req *req = container_of(_req, struct fsl_req, req);
struct fsl_udc *udc;
- int is_iso = 0;
+ int ret, is_iso = 0;
/* catch various bogus parameters */
if (!_req || !req->req.complete || !req->req.buf
@@ -885,8 +884,9 @@ fsl_ep_queue(struct usb_ep *_ep, struct usb_request *_req)
req->ep = ep;
- dma_sync_single_for_device(udc->gadget.dev.parent, (unsigned long)req->req.buf,
- req->req.length, DMA_BIDIRECTIONAL);
+ ret = usb_gadget_map_request(&ep->udc->gadget, &req->req, ep_is_in(ep));
+ if (ret)
+ return ret;
req->req.status = -EINPROGRESS;
req->req.actual = 0;
--
2.39.2
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast
2024-07-24 9:47 [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Ahmad Fatoum
2024-07-24 9:47 ` [PATCH master 2/2] usb: gadget: fsl_udc: dma_map buffers instead of using dma_sync Ahmad Fatoum
@ 2024-07-30 8:13 ` Sascha Hauer
1 sibling, 0 replies; 3+ messages in thread
From: Sascha Hauer @ 2024-07-30 8:13 UTC (permalink / raw)
To: barebox, Ahmad Fatoum; +Cc: ske
On Wed, 24 Jul 2024 11:47:31 +0200, Ahmad Fatoum wrote:
> All other prints in the file use the DMA address except for this single
> print inside debug_dma_map(). Let's use the DMA address there as well to
> fix the warning about casting a 32-bit pointer to a 64 bit integer and
> to align error messages in case of non-1:1 mapped DMA addresses.
>
>
Applied, thanks!
[1/2] dma: debug: fix compiler warning about pointer u64 cast
https://git.pengutronix.de/cgit/barebox/commit/?id=4cd2dbea7bda (link may not be stable)
[2/2] usb: gadget: fsl_udc: dma_map buffers instead of using dma_sync
https://git.pengutronix.de/cgit/barebox/commit/?id=c825a35f0f35 (link may not be stable)
Best regards,
--
Sascha Hauer <s.hauer@pengutronix.de>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2024-07-30 8:13 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-07-24 9:47 [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Ahmad Fatoum
2024-07-24 9:47 ` [PATCH master 2/2] usb: gadget: fsl_udc: dma_map buffers instead of using dma_sync Ahmad Fatoum
2024-07-30 8:13 ` [PATCH master 1/2] dma: debug: fix compiler warning about pointer u64 cast Sascha Hauer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox