[PATCH] Revert "dm: core: Scan "/firmware" node by default"

[PATCH] Revert "dm: core: Scan "/firmware" node by default"

[Thirupathaiah Annapureddy]

Subnodes under "/firmware" node are scanned twice in dm_scan_fdt_node
and dm_extended_scan_fdt. This patch removes the double scanning.

This reverts commit 1712ca21924bff678f19fd3141f435408d23bdbf.

Signed-off-by: Thirupathaiah Annapureddy <thiruan@linux.microsoft.com>
---
 drivers/core/root.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/drivers/core/root.c b/drivers/core/root.c
index e85643819e..daf7c61190 100644
--- a/drivers/core/root.c
+++ b/drivers/core/root.c
@@ -342,14 +342,8 @@ int dm_extended_scan_fdt(const void *blob, bool pre_reloc_only)
 	}
 
 	ret = dm_scan_fdt_ofnode_path("/clocks", pre_reloc_only);
-	if (ret) {
-		debug("scan for /clocks failed: %d\n", ret);
-		return ret;
-	}
-
-	ret = dm_scan_fdt_ofnode_path("/firmware", pre_reloc_only);
 	if (ret)
-		debug("scan for /firmware failed: %d\n", ret);
+		debug("scan for /clocks failed: %d\n", ret);
 
 	return ret;
 }
-- 
2.24.1

[PATCH] Revert "dm: core: Scan "/firmware" node by default"

[Simon Glass]

Hi T,

On Tue, 7 Jan 2020 at 20:24, Thirupathaiah Annapureddy
<thiruan@linux.microsoft.com> wrote:
>
> Subnodes under "/firmware" node are scanned twice in dm_scan_fdt_node
> and dm_extended_scan_fdt. This patch removes the double scanning.
>
> This reverts commit 1712ca21924bff678f19fd3141f435408d23bdbf.
>
> Signed-off-by: Thirupathaiah Annapureddy <thiruan@linux.microsoft.com>
> ---
>  drivers/core/root.c | 8 +-------
>  1 file changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/drivers/core/root.c b/drivers/core/root.c
> index e85643819e..daf7c61190 100644
> --- a/drivers/core/root.c
> +++ b/drivers/core/root.c
> @@ -342,14 +342,8 @@ int dm_extended_scan_fdt(const void *blob, bool pre_reloc_only)
>         }
>
>         ret = dm_scan_fdt_ofnode_path("/clocks", pre_reloc_only);
> -       if (ret) {
> -               debug("scan for /clocks failed: %d\n", ret);
> -               return ret;
> -       }
> -
> -       ret = dm_scan_fdt_ofnode_path("/firmware", pre_reloc_only);
>         if (ret)
> -               debug("scan for /firmware failed: %d\n", ret);
> +               debug("scan for /clocks failed: %d\n", ret);
>
>         return ret;
>  }
> --
> 2.24.1
>

I actually think it would be better to remove this from
dm_scan_fdt_node() since that function looks for f

[PATCH] Revert "dm: core: Scan "/firmware" node by default"

[Thirupathaiah Annapureddy]

On 1/8/2020 9:39 AM, Simon Glass wrote:
> Hi T,
> 
> On Tue, 7 Jan 2020 at 20:24, Thirupathaiah Annapureddy
> <thiruan@linux.microsoft.com> wrote:
>>
>> Subnodes under "/firmware" node are scanned twice in dm_scan_fdt_node
>> and dm_extended_scan_fdt. This patch removes the double scanning.
>>
>> This reverts commit 1712ca21924bff678f19fd3141f435408d23bdbf.
>>
>> Signed-off-by: Thirupathaiah Annapureddy <thiruan@linux.microsoft.com>
>> ---
>>  drivers/core/root.c | 8 +-------
>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>
>> diff --git a/drivers/core/root.c b/drivers/core/root.c
>> index e85643819e..daf7c61190 100644
>> --- a/drivers/core/root.c
>> +++ b/drivers/core/root.c
>> @@ -342,14 +342,8 @@ int dm_extended_scan_fdt(const void *blob, bool pre_reloc_only)
>>         }
>>
>>         ret = dm_scan_fdt_ofnode_path("/clocks", pre_reloc_only);
>> -       if (ret) {
>> -               debug("scan for /clocks failed: %d\n", ret);
>> -               return ret;
>> -       }
>> -
>> -       ret = dm_scan_fdt_ofnode_path("/firmware", pre_reloc_only);
>>         if (ret)
>> -               debug("scan for /firmware failed: %d\n", ret);
>> +               debug("scan for /clocks failed: %d\n", ret);
>>
>>         return ret;
>>  }
>> --
>> 2.24.1
>>
> 
> I actually think it would be better to remove this from
> dm_scan_fdt_node() since that function looks for f
> 

Hi Simon,

Thanks. I will address your comment and send new patch. 
My rationale for removing the scanning from dm_extended_scan_fdt() was
based on the order of the scan. Why do you think removing the scan from 
dm_extended_scan_fdt() is better?

--Thiru