[PATCH v2 08/15] habanalabs: add comments in uapi/misc/habanalabs.h

[PATCH v2 08/15] habanalabs: add comments in uapi/misc/habanalabs.h

[Oded Gabbay]

Add comment about minimum and maximum size of command buffer.
Add some text about the expected input of CS IOCTL.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
---
Changes in v2:
  - Add changelog in the commit message
 
 include/uapi/misc/habanalabs.h | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/include/uapi/misc/habanalabs.h b/include/uapi/misc/habanalabs.h
index 23d6ad3459cb..7fd6f633534c 100644
--- a/include/uapi/misc/habanalabs.h
+++ b/include/uapi/misc/habanalabs.h
@@ -112,7 +112,9 @@ struct hl_cb_in {
 	__u64 cb_handle;
 	/* HL_CB_OP_* */
 	__u32 op;
-	/* Size of CB. Minimum requested size must be PAGE_SIZE */
+	/* Size of CB. Maximum size is 2MB. The minimum size that will be
+	 * allocated, regardless of this parameter's value, is PAGE_SIZE
+	 */
 	__u32 cb_size;
 	/* Context ID - Currently not in use */
 	__u32 ctx_id;
@@ -364,6 +366,12 @@ union hl_mem_args {
  * internal. The driver will get completion notifications from the device only
  * on JOBS which are enqueued in the external queues.
  *
+ * For jobs on external queues, the user needs to create command buffers
+ * through the CB ioctl and give the CB's handle to the CS ioctl. For jobs on
+ * internal queues, the user needs to prepare a "command buffer" with packets
+ * on either the SRAM or DRAM, and give the device address of that buffer to
+ * the CS ioctl.
+ *
  * This IOCTL is asynchronous in regard to the actual execution of the CS. This
  * means it returns immediately after ALL the JOBS were enqueued on their
  * relevant queues. Therefore, the user mustn't assume the CS has been completed
-- 
2.17.1

[PATCH 11/15] habanalabs: print pointer using %p

[Oded Gabbay]

Don't cast pointer to u64 to print it. Instead, print the pointer using
%p.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
---
Changes in v2:
  - Add changelog in the commit message
 
 drivers/misc/habanalabs/goya/goya.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
index 13923f4127af..39824214ce61 100644
--- a/drivers/misc/habanalabs/goya/goya.c
+++ b/drivers/misc/habanalabs/goya/goya.c
@@ -4276,9 +4276,8 @@ static int goya_parse_cb_no_ext_quque(struct hl_device *hdev,
 			return 0;
 
 		dev_err(hdev->dev,
-			"Internal CB address 0x%llx + 0x%x is not in SRAM nor in DRAM\n",
-			(u64) (uintptr_t) parser->user_cb,
-			parser->user_cb_size);
+			"Internal CB address %p + 0x%x is not in SRAM nor in DRAM\n",
+			parser->user_cb, parser->user_cb_size);
 
 		return -EFAULT;
 	}
-- 
2.17.1

Re: [PATCH 11/15] habanalabs: print pointer using %p

[Greg KH]

On Thu, Feb 28, 2019 at 11:55:45AM +0200, Oded Gabbay wrote:
> Don't cast pointer to u64 to print it. Instead, print the pointer using
> %p.
> 
> Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
> ---
> Changes in v2:
>   - Add changelog in the commit message

This isn't marked "v2" :(

Anyway, I'll try to piece this all together, next time it's easier to
just resend the whole series...

greg k-h

Re: [PATCH 11/15] habanalabs: print pointer using %p

[Oded Gabbay]

On Thu, Feb 28, 2019 at 2:02 PM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Thu, Feb 28, 2019 at 11:55:45AM +0200, Oded Gabbay wrote:
> > Don't cast pointer to u64 to print it. Instead, print the pointer using
> > %p.
> >
> > Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
> > ---
> > Changes in v2:
> >   - Add changelog in the commit message
>
> This isn't marked "v2" :(
Yes, sorry about that, will do better next time.

>
> Anyway, I'll try to piece this all together, next time it's easier to
> just resend the whole series...
Got it,
Thanks,
Oded
>
> greg k-h

Re: [PATCH 11/15] habanalabs: print pointer using %p

[Joe Perches]

On Thu, 2019-02-28 at 11:55 +0200, Oded Gabbay wrote:
> Don't cast pointer to u64 to print it. Instead, print the pointer using
> %p.

You might want to use %px here if you _really_
want the actual address and not the hashed output
%p normally produces.

> diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
[]
> @@ -4276,9 +4276,8 @@ static int goya_parse_cb_no_ext_quque(struct hl_device *hdev,
>  			return 0;
>  
>  		dev_err(hdev->dev,
> -			"Internal CB address 0x%llx + 0x%x is not in SRAM nor in DRAM\n",
> -			(u64) (uintptr_t) parser->user_cb,
> -			parser->user_cb_size);
> +			"Internal CB address %p + 0x%x is not in SRAM nor in DRAM\n",
> +			parser->user_cb, parser->user_cb_size);

Re: [PATCH 11/15] habanalabs: print pointer using %p

[Oded Gabbay]

On Sat, Mar 2, 2019 at 3:00 AM Joe Perches <joe@perches.com> wrote:
>
> On Thu, 2019-02-28 at 11:55 +0200, Oded Gabbay wrote:
> > Don't cast pointer to u64 to print it. Instead, print the pointer using
> > %p.
>
> You might want to use %px here if you _really_
> want the actual address and not the hashed output
> %p normally produces.
Yes, you are correct. Thanks for pointing it out, I'll fix that.
Oded


>
> > diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
> []
> > @@ -4276,9 +4276,8 @@ static int goya_parse_cb_no_ext_quque(struct hl_device *hdev,
> >                       return 0;
> >
> >               dev_err(hdev->dev,
> > -                     "Internal CB address 0x%llx + 0x%x is not in SRAM nor in DRAM\n",
> > -                     (u64) (uintptr_t) parser->user_cb,
> > -                     parser->user_cb_size);
> > +                     "Internal CB address %p + 0x%x is not in SRAM nor in DRAM\n",
> > +                     parser->user_cb, parser->user_cb_size);
>
>