Skip to contentSkip to conversation

Fix default values for address and size cells

PATCH New matthias.bgg at kernel.org 5 participants 23 messages Sep 5, 2019 – Oct 22, 2019 Device tree

lore
Share & tools
lore
#0 lore raw

v2 cover Fix default values for address and size cells

From: Matthias Brugger <[email protected]>

The libftd implementation of U-Boot is outdated with the upstream project. Especially the default number of size-cells was wrong. This series fixes this by backporting the corresponding patches from the upstream project.

Changes since v1:
- fix spelling of U-Boot (patch 1 + 2 /4)
- add Reviewed-by tag (patch 1/4)
- fix the case of size-cells == 0 (patch 3/4)
- fix default address cells for livetree (patch 4/4)

v1 can be found here:
https://patchwork.ozlabs.org/patch/1137331/
https://patchwork.ozlabs.org/patch/1137330/

Matthias Brugger (4):
libfdt: fdt_address_cells() and fdt_size_cells()
libfdt: return correct value if #size-cells property is not present
libfdt: Allow #size-cells of 0
dm: Fix default address cells return value

 include/dm/of.h                    |  2 +-
 scripts/dtc/libfdt/fdt_addresses.c | 45 ++++++++++++++++--------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 3 files changed, 26 insertions(+), 23 deletions(-)
Signature · 1 line
2.22.0
#1 lore raw

PATCH v2 1/4 libfdt: fdt_address_cells() and fdt_size_cells()

From: Matthias Brugger <[email protected]>

Add internal fdt_cells() to avoid copy and paste. Fix typo in fdt_size_cells() documentation comment.

This is based in upstream commit:
c12b2b0 ("libfdt: fdt_address_cells() and fdt_size_cells()") but misses the test cases, as we don't implement them in U-Boot.

---

 scripts/dtc/libfdt/fdt_addresses.c | 35 +++++++++++-------------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 23 deletions(-)
scripts/dtc/libfdt/fdt_addresses.c +13 -22
index eff4dbcc72..49537b578d 100644
@@ -1,6 +1,7 @@
 /*
  * libfdt - Flat Device Tree manipulation
  * Copyright (C) 2014 David Gibson <[email protected]>
+ * Copyright (C) 2018 embedded brains GmbH
  *
  * libfdt is dual licensed: you can use it either under the terms of
  * the GPL, or the BSD license, at your option.
@@ -55,42 +56,32 @@
 
 #include "libfdt_internal.h"
 
-int fdt_address_cells(const void *fdt, int nodeoffset)
+static int fdt_cells(const void *fdt, int nodeoffset, const char *name)
 {
-	const fdt32_t *ac;
+	const fdt32_t *c;
 	int val;
 	int len;
 
-	ac = fdt_getprop(fdt, nodeoffset, "#address-cells", &len);
-	if (!ac)
+	c = fdt_getprop(fdt, nodeoffset, name, &len);
+	if (!c)
 		return 2;
 
-	if (len != sizeof(*ac))
+	if (len != sizeof(*c))
 		return -FDT_ERR_BADNCELLS;
 
-	val = fdt32_to_cpu(*ac);
+	val = fdt32_to_cpu(*c);
 	if ((val <= 0) || (val > FDT_MAX_NCELLS))
 		return -FDT_ERR_BADNCELLS;
 
 	return val;
 }
 
-int fdt_size_cells(const void *fdt, int nodeoffset)
+int fdt_address_cells(const void *fdt, int nodeoffset)
 {
-	const fdt32_t *sc;
-	int val;
-	int len;
-
-	sc = fdt_getprop(fdt, nodeoffset, "#size-cells", &len);
-	if (!sc)
-		return 2;
-
-	if (len != sizeof(*sc))
-		return -FDT_ERR_BADNCELLS;
-
-	val = fdt32_to_cpu(*sc);
-	if ((val < 0) || (val > FDT_MAX_NCELLS))
-		return -FDT_ERR_BADNCELLS;
+	return fdt_cells(fdt, nodeoffset, "#address-cells");
+}
 
-	return val;
+int fdt_size_cells(const void *fdt, int nodeoffset)
+{
+	return fdt_cells(fdt, nodeoffset, "#size-cells");
 }
scripts/dtc/libfdt/libfdt.h +1 -1
index cf86ddba88..66f01fec53 100644
@@ -1109,7 +1109,7 @@ int fdt_address_cells(const void *fdt, int nodeoffset);
  *
  * returns:
  *	0 <= n < FDT_MAX_NCELLS, on success
- *      2, if the node has no #address-cells property
+ *      2, if the node has no #size-cells property
  *      -FDT_ERR_BADNCELLS, if the node has a badly formatted or invalid
  *		#size-cells property
  *	-FDT_ERR_BADMAGIC,
Signature · 1 line
2.22.0
#2 lore raw

PATCH v2 1/4 libfdt: fdt_address_cells() and fdt_size_cells()

From: Matthias Brugger <[email protected]>

Add internal fdt_cells() to avoid copy and paste. Fix typo in fdt_size_cells() documentation comment.

This is based in upstream commit:
c12b2b0 ("libfdt: fdt_address_cells() and fdt_size_cells()") but misses the test cases, as we don't implement them in U-Boot.

---

 scripts/dtc/libfdt/fdt_addresses.c | 35 +++++++++++-------------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 23 deletions(-)

Applied to u-boot-dm/next, thanks!

#3 lore raw

PATCH v2 1/4 libfdt: fdt_address_cells() and fdt_size_cells()

Hi Simon,

Quoted 11 lines · sjg at google.com

From: Matthias Brugger <[email protected]>

Add internal fdt_cells() to avoid copy and paste. Fix typo in fdt_size_cells() documentation comment.

This is based in upstream commit:
c12b2b0 ("libfdt: fdt_address_cells() and fdt_size_cells()") but misses the test cases, as we don't implement them in U-Boot.

Quoted 7 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 35 +++++++++++-------------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 23 deletions(-)

Applied to u-boot-dm/next, thanks!

I just looked through the upstream U-Boot commit log and don't see them queued for 2019.10-rc5
Do I understand correctly that you are planning to merge them after 2019.10 is released. Correct?

I'm asking because without these patches the grub boot path for RPi4 with RAM > 1GB is 'broken', in the sense that the kernel will only see RAM size of 1GB.

I can keep them as backports in openSUSE, that's no problem, just wanted to know what's your plan. And have a quick answer if someone asks on the mailinglist :)

Regards,
Matthias

#4 lore raw

PATCH v2 1/4 libfdt: fdt_address_cells() and fdt_size_cells()

Hi Matthias,

Quoted 28 lines · Matthias Brugger

Hi Simon,

Quoted 11 lines · sjg at google.com

From: Matthias Brugger <[email protected]>

Add internal fdt_cells() to avoid copy and paste. Fix typo in fdt_size_cells() documentation comment.

This is based in upstream commit:
c12b2b0 ("libfdt: fdt_address_cells() and fdt_size_cells()") but misses the test cases, as we don't implement them in U-Boot.

Quoted 7 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 35 +++++++++++-------------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 23 deletions(-)

Applied to u-boot-dm/next, thanks!

I just looked through the upstream U-Boot commit log and don't see them queued for 2019.10-rc5
Do I understand correctly that you are planning to merge them after 2019.10 is released. Correct?

Quoted 6 lines

I'm asking because without these patches the grub boot path for RPi4 with RAM > 1GB is 'broken', in the sense that the kernel will only see RAM size of 1GB.

I can keep them as backports in openSUSE, that's no problem, just wanted to know what's your plan. And have a quick answer if someone asks on the mailinglist :)

Yes that's right. Should get these applied next week.

Regards,
Simon

#5 lore raw

PATCH v2 1/4 libfdt: fdt_address_cells() and fdt_size_cells()

Quoted 31 lines · Simon Glass

Hi Matthias,

Quoted 28 lines · Matthias Brugger

Hi Simon,

Quoted 11 lines · sjg at google.com

From: Matthias Brugger <[email protected]>

Add internal fdt_cells() to avoid copy and paste. Fix typo in fdt_size_cells() documentation comment.

This is based in upstream commit:
c12b2b0 ("libfdt: fdt_address_cells() and fdt_size_cells()") but misses the test cases, as we don't implement them in U-Boot.

Quoted 7 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 35 +++++++++++-------------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 23 deletions(-)

Applied to u-boot-dm/next, thanks!

I just looked through the upstream U-Boot commit log and don't see them queued for 2019.10-rc5
Do I understand correctly that you are planning to merge them after 2019.10 is released. Correct?

Quoted 9 lines
Quoted 6 lines

I'm asking because without these patches the grub boot path for RPi4 with RAM > 1GB is 'broken', in the sense that the kernel will only see RAM size of 1GB.

I can keep them as backports in openSUSE, that's no problem, just wanted to know what's your plan. And have a quick answer if someone asks on the mailinglist :)

Yes that's right. Should get these applied next week.

Ok, thanks for the info :)

#6 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

In reply to matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

---

 scripts/dtc/libfdt/fdt_addresses.c | 16 +++++++++++++---
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 4 deletions(-)
scripts/dtc/libfdt/fdt_addresses.c +13 -3
index 49537b578d..f13a87dfa0 100644
@@ -64,7 +64,7 @@ static int fdt_cells(const void *fdt, int nodeoffset, const char *name)
 
 	c = fdt_getprop(fdt, nodeoffset, name, &len);
 	if (!c)
-		return 2;
+		return len;
 
 	if (len != sizeof(*c))
 		return -FDT_ERR_BADNCELLS;
@@ -78,10 +78,20 @@ static int fdt_cells(const void *fdt, int nodeoffset, const char *name)
 
 int fdt_address_cells(const void *fdt, int nodeoffset)
 {
-	return fdt_cells(fdt, nodeoffset, "#address-cells");
+	int val;
+
+	val = fdt_cells(fdt, nodeoffset, "#address-cells");
+	if (val == -FDT_ERR_NOTFOUND)
+		return 2;
+	return val;
 }
 
 int fdt_size_cells(const void *fdt, int nodeoffset)
 {
-	return fdt_cells(fdt, nodeoffset, "#size-cells");
+	int val;
+
+	val = fdt_cells(fdt, nodeoffset, "#size-cells");
+	if (val == -FDT_ERR_NOTFOUND)
+		return 1;
+	return val;
 }
scripts/dtc/libfdt/libfdt.h +1 -1
index 66f01fec53..5c778b115b 100644
@@ -1109,7 +1109,7 @@ int fdt_address_cells(const void *fdt, int nodeoffset);
  *
  * returns:
  *	0 <= n < FDT_MAX_NCELLS, on success
- *      2, if the node has no #size-cells property
+ *      1, if the node has no #size-cells property
  *      -FDT_ERR_BADNCELLS, if the node has a badly formatted or invalid
  *		#size-cells property
  *	-FDT_ERR_BADMAGIC,
Signature · 1 line
2.22.0
#7 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Hi,

Quoted 11 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

Quoted 5 lines

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

This is v2 but I don't see a change log?

Quoted 5 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 16 +++++++++++++---
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 4 deletions(-)

Regards,
Simon

#8 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Hi Simon,

Quoted 14 lines · Simon Glass

Hi,

Quoted 11 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

Quoted 8 lines
Quoted 5 lines

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

This is v2 but I don't see a change log?

I put the changelog into the cover letter:
https://patchwork.ozlabs.org/cover/1158304/

From your email I understand that you prefer a patch by patch changelog, correct?

Regards,
Matthias

Quoted 9 lines
Quoted 5 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 16 +++++++++++++---
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 2 files changed, 14 insertions(+), 4 deletions(-)

Regards,
Simon

#9 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Hi Matthias,

Quoted 24 lines · Matthias Brugger

Hi Simon,

Quoted 20 lines · Simon Glass

Hi,

Quoted 16 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

Quoted 7 lines
Quoted 2 lines

This is v2 but I don't see a change log?

I put the changelog into the cover letter:
https://patchwork.ozlabs.org/cover/1158304/

From your email I understand that you prefer a patch by patch changelog, correct?

Both is best. If you use patman it does this for you.

Regards,
Simon

#10 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Quoted 15 lines · Simon Glass

Hi Matthias,

Quoted 11 lines · Matthias Brugger

Hi Simon,

Quoted 7 lines · Simon Glass

Hi,

Quoted 4 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for

There seems to be something missing here/...

Quoted 16 lines
Quoted 16 lines
Quoted 15 lines
Quoted 7 lines

was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

Quoted 5 lines

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

This is v2 but I don't see a change log?

Quoted 6 lines
Quoted 4 lines

I put the changelog into the cover letter:
https://patchwork.ozlabs.org/cover/1158304/

From your email I understand that you prefer a patch by patch changelog, correct?

Both is best. If you use patman it does this for you.

Regards,
Simon

#11 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Quoted 11 lines · Simon Glass
Quoted 10 lines · Simon Glass

Hi Matthias,

Quoted 6 lines · Matthias Brugger

Hi Simon,

Quoted 2 lines · Simon Glass

Hi,

Quoted 8 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:
Quoted 4 lines

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for

There seems to be something missing here/...

Argh, you are right, I accidentally deleted one line from the commit message. The correct paragraph should read like this:

<snip>
According to the device tree specification, the default value for #size-cells is 1, but fdt_size_cells() was returning 2 if this property was not present.
</snip>

BTW this series is a fix and I'd like to see it in v2019.10-rc5. I'm not sure who is to take it, you or Tom.

Regards,
Matthias

Quoted 28 lines
Quoted 16 lines
Quoted 16 lines
Quoted 15 lines
Quoted 7 lines

was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

Quoted 5 lines

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

This is v2 but I don't see a change log?

Quoted 6 lines
Quoted 4 lines

I put the changelog into the cover letter:
https://patchwork.ozlabs.org/cover/1158304/

From your email I understand that you prefer a patch by patch changelog, correct?

Both is best. If you use patman it does this for you.

Regards,
Simon

#12 lore raw

PATCH v2 2/4 libfdt: return correct value if #size-cells property is not present

Quoted 11 lines · Simon Glass
Quoted 10 lines · Simon Glass

Hi Matthias,

Quoted 6 lines · Matthias Brugger

Hi Simon,

Quoted 2 lines · Simon Glass

Hi,

Quoted 8 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:
Quoted 4 lines

From: Matthias Brugger <[email protected]>

According to the device tree specification, the default value for

There seems to be something missing here/...

Argh, you are right, I accidentally deleted one line from the commit message. The correct paragraph should read like this:

<snip>
According to the device tree specification, the default value for #size-cells is 1, but fdt_size_cells() was returning 2 if this property was not present.
</snip>

BTW this series is a fix and I'd like to see it in v2019.10-rc5. I'm not sure who is to take it, you or Tom.

Regards,
Matthias

Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 7 lines
Quoted 7 lines

was not present.

This patch also makes fdt_address_cells() and fdt_size_cells() conform to the behaviour documented in libfdt.h. The defaults are only returned if fdt_getprop() returns -FDT_ERR_NOTFOUND, otherwise the actual error is returned.

Quoted 8 lines
Quoted 5 lines

This is based on upstream commit:
aa7254d ("libfdt: return correct value if #size-cells property is not present") but misses the test case part, as we don't implement them in U-Boot.

This is v2 but I don't see a change log?

I put the changelog into the cover letter:
https://patchwork.ozlabs.org/cover/1158304/

Quoted 9 lines
Quoted 3 lines
Quoted 1 line

From your email I understand that you prefer a patch by patch changelog, correct?

Both is best. If you use patman it does this for you.

Regards,
Simon

Applied to u-boot-dm/next, thanks!

#13 lore raw

PATCH v2 3/4 libfdt: Allow #size-cells of 0

In reply to matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

The commit "libfdt: fdt_address_cells() and fdt_size_cells()" introduced a bug as it consolidated code between the helpers for getting be 0, and is frequently found so in practice for /cpus. IEEE1275 only requires implementations to handle 1..4 for #address-cells, although one could make a case for #address-cells == #size-cells == 0 being used to represent a bridge with a single port.

While we're there, it's not totally obvious that the existing implicit cast of a u32 to int will give the correct results according to strict C, although it does work in practice. Straighten that up to cast only after we've made our range checks.

This is based on upstream commit:
b8d6eca ("libfdt: Allow #size-cells of 0")
but misses the test cases,as we don't implement them in U-Boot.

---

 scripts/dtc/libfdt/fdt_addresses.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
scripts/dtc/libfdt/fdt_addresses.c +5 -3
index f13a87dfa0..788c143113 100644
@@ -59,7 +59,7 @@
 static int fdt_cells(const void *fdt, int nodeoffset, const char *name)
 {
 	const fdt32_t *c;
-	int val;
+	uint32_t val;
 	int len;
 
 	c = fdt_getprop(fdt, nodeoffset, name, &len);
@@ -70,10 +70,10 @@ static int fdt_cells(const void *fdt, int nodeoffset, const char *name)
 		return -FDT_ERR_BADNCELLS;
 
 	val = fdt32_to_cpu(*c);
-	if ((val <= 0) || (val > FDT_MAX_NCELLS))
+	if (val > FDT_MAX_NCELLS)
 		return -FDT_ERR_BADNCELLS;
 
-	return val;
+	return (int)val;
 }
 
 int fdt_address_cells(const void *fdt, int nodeoffset)
@@ -81,6 +81,8 @@ int fdt_address_cells(const void *fdt, int nodeoffset)
 	int val;
 
 	val = fdt_cells(fdt, nodeoffset, "#address-cells");
+	if (val == 0)
+		return -FDT_ERR_BADNCELLS;
 	if (val == -FDT_ERR_NOTFOUND)
 		return 2;
 	return val;
Signature · 1 line
2.22.0
#14 lore raw

PATCH v2 3/4 libfdt: Allow #size-cells of 0

Quoted 19 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

The commit "libfdt: fdt_address_cells() and fdt_size_cells()" introduced a bug as it consolidated code between the helpers for getting be 0, and is frequently found so in practice for /cpus. IEEE1275 only requires implementations to handle 1..4 for #address-cells, although one could make a case for #address-cells == #size-cells == 0 being used to represent a bridge with a single port.

While we're there, it's not totally obvious that the existing implicit cast of a u32 to int will give the correct results according to strict C, although it does work in practice. Straighten that up to cast only after we've made our range checks.

This is based on upstream commit:
b8d6eca ("libfdt: Allow #size-cells of 0")
but misses the test cases,as we don't implement them in U-Boot.

Quoted 5 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

This is v2 but I don't see a change log?

- Simon

#15 lore raw

PATCH v2 3/4 libfdt: Allow #size-cells of 0

Quoted 21 lines · Simon Glass
Quoted 19 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

The commit "libfdt: fdt_address_cells() and fdt_size_cells()" introduced a bug as it consolidated code between the helpers for getting be 0, and is frequently found so in practice for /cpus. IEEE1275 only requires implementations to handle 1..4 for #address-cells, although one could make a case for #address-cells == #size-cells == 0 being used to represent a bridge with a single port.

While we're there, it's not totally obvious that the existing implicit cast of a u32 to int will give the correct results according to strict C, although it does work in practice. Straighten that up to cast only after we've made our range checks.

This is based on upstream commit:
b8d6eca ("libfdt: Allow #size-cells of 0")
but misses the test cases,as we don't implement them in U-Boot.

Quoted 7 lines
Quoted 5 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

This is v2 but I don't see a change log?

(please use patman if you can next time)

#16 lore raw

PATCH v2 3/4 libfdt: Allow #size-cells of 0

Quoted 21 lines · Simon Glass
Quoted 19 lines · On Thu, 5 Sep 2019 at 02:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

The commit "libfdt: fdt_address_cells() and fdt_size_cells()" introduced a bug as it consolidated code between the helpers for getting be 0, and is frequently found so in practice for /cpus. IEEE1275 only requires implementations to handle 1..4 for #address-cells, although one could make a case for #address-cells == #size-cells == 0 being used to represent a bridge with a single port.

While we're there, it's not totally obvious that the existing implicit cast of a u32 to int will give the correct results according to strict C, although it does work in practice. Straighten that up to cast only after we've made our range checks.

This is based on upstream commit:
b8d6eca ("libfdt: Allow #size-cells of 0")
but misses the test cases,as we don't implement them in U-Boot.

Quoted 7 lines
Quoted 5 lines

---

 scripts/dtc/libfdt/fdt_addresses.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

This is v2 but I don't see a change log?

(please use patman if you can next time)

Applied to u-boot-dm/next, thanks!

#17 lore raw

PATCH v2 4/4 dm: Fix default address cells return value

In reply to matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

Default address cells value on the livetree access function returns the wrong value. Fix this so that the value returned corresponds to the device tree specification.

---

 include/dm/of.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/dm/of.h +1 -1
index 461e25aa19..6bef73b441 100644
@@ -111,7 +111,7 @@ static inline const char *of_node_full_name(const struct device_node *np)
 
 /* Default #address and #size cells */
 #if !defined(OF_ROOT_NODE_ADDR_CELLS_DEFAULT)
-#define OF_ROOT_NODE_ADDR_CELLS_DEFAULT 1
+#define OF_ROOT_NODE_ADDR_CELLS_DEFAULT 2
 #define OF_ROOT_NODE_SIZE_CELLS_DEFAULT 1
 #endif
 
Signature · 1 line
2.22.0
#18 lore raw

PATCH v2 4/4 dm: Fix default address cells return value

Quoted 7 lines · On Thu, 5 Sep 2019 at 01:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

Default address cells value on the livetree access function returns the wrong value. Fix this so that the value returned corresponds to the device tree specification.

Quoted 5 lines

---

 include/dm/of.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
#19 lore raw

PATCH v2 4/4 dm: Fix default address cells return value

Quoted 7 lines · On Thu, 5 Sep 2019 at 01:49, <[email protected]> wrote:

From: Matthias Brugger <[email protected]>

Default address cells value on the livetree access function returns the wrong value. Fix this so that the value returned corresponds to the device tree specification.

Quoted 5 lines

---

 include/dm/of.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Applied to u-boot-dm/next, thanks!

#20 lore raw

v2 cover Fix default values for address and size cells

In reply to matthias.bgg at kernel.org

Hi Simon,

Quoted 24 lines · matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

The libftd implementation of U-Boot is outdated with the upstream project. Especially the default number of size-cells was wrong. This series fixes this by backporting the corresponding patches from the upstream project.

Changes since v1:
- fix spelling of U-Boot (patch 1 + 2 /4)
- add Reviewed-by tag (patch 1/4)
- fix the case of size-cells == 0 (patch 3/4)
- fix default address cells for livetree (patch 4/4)

v1 can be found here:
https://patchwork.ozlabs.org/patch/1137331/
https://patchwork.ozlabs.org/patch/1137330/

Matthias Brugger (4):
libfdt: fdt_address_cells() and fdt_size_cells()
libfdt: return correct value if #size-cells property is not present
libfdt: Allow #size-cells of 0
dm: Fix default address cells return value

Quoted 5 lines
 include/dm/of.h                    |  2 +-
 scripts/dtc/libfdt/fdt_addresses.c | 45 ++++++++++++++++--------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 3 files changed, 26 insertions(+), 23 deletions(-)

Any comments on this? I tried to find your custodian tree, but wasn't able to do so and check if you took it already.

Regards,
Matthias

#21 lore raw

v2 cover Fix default values for address and size cells

Hi Matthias,

Quoted 28 lines · Matthias Brugger

Hi Simon,

Quoted 24 lines · matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

The libftd implementation of U-Boot is outdated with the upstream project. Especially the default number of size-cells was wrong. This series fixes this by backporting the corresponding patches from the upstream project.

Changes since v1:
- fix spelling of U-Boot (patch 1 + 2 /4)
- add Reviewed-by tag (patch 1/4)
- fix the case of size-cells == 0 (patch 3/4)
- fix default address cells for livetree (patch 4/4)

v1 can be found here:
https://patchwork.ozlabs.org/patch/1137331/
https://patchwork.ozlabs.org/patch/1137330/

Matthias Brugger (4):
libfdt: fdt_address_cells() and fdt_size_cells()
libfdt: return correct value if #size-cells property is not present
libfdt: Allow #size-cells of 0
dm: Fix default address cells return value

Quoted 8 lines
Quoted 5 lines
 include/dm/of.h                    |  2 +-
 scripts/dtc/libfdt/fdt_addresses.c | 45 ++++++++++++++++--------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 3 files changed, 26 insertions(+), 23 deletions(-)

Any comments on this? I tried to find your custodian tree, but wasn't able to do so and check if you took it already.

This was applied. Did you not get the emails on each patch?

Regards,
Simon

#22 lore raw

v2 cover Fix default values for address and size cells

Quoted 37 lines · Simon Glass

Hi Matthias,

Quoted 34 lines · Matthias Brugger

Hi Simon,

Quoted 29 lines · matthias.bgg at kernel.org

From: Matthias Brugger <[email protected]>

The libftd implementation of U-Boot is outdated with the upstream project. Especially the default number of size-cells was wrong. This series fixes this by backporting the corresponding patches from the upstream project.

Changes since v1:
- fix spelling of U-Boot (patch 1 + 2 /4)
- add Reviewed-by tag (patch 1/4)
- fix the case of size-cells == 0 (patch 3/4)
- fix default address cells for livetree (patch 4/4)

v1 can be found here:
https://patchwork.ozlabs.org/patch/1137331/
https://patchwork.ozlabs.org/patch/1137330/

Matthias Brugger (4):
libfdt: fdt_address_cells() and fdt_size_cells()
libfdt: return correct value if #size-cells property is not present
libfdt: Allow #size-cells of 0
dm: Fix default address cells return value

 include/dm/of.h                    |  2 +-
 scripts/dtc/libfdt/fdt_addresses.c | 45 ++++++++++++++++--------------
 scripts/dtc/libfdt/libfdt.h        |  2 +-
 3 files changed, 26 insertions(+), 23 deletions(-)
Quoted 5 lines
Quoted 2 lines

Any comments on this? I tried to find your custodian tree, but wasn't able to do so and check if you took it already.

This was applied. Did you not get the emails on each patch?

I got emails saying that they were applied to u-boot-dm/next, but wasn't able to find that branch.

Seeing it now, not sure if I didn't look correctly. Anyway thank for merging these fixes.

Regards,
Matthias