[OE-core][PATCH] opensbi: bump to 1.4

[OE-core][PATCH] opensbi: bump to 1.4

[thomas.perrot]

From: Thomas Perrot <thomas.perrot@bootlin.com>

This release has:
- Synopsys DesignWare APB GPIO driver
- Zicntr and Zihpm support
- Console print improvements
- Smepmp support
- Simple FDT based syscon regmap driver
- Syscon based reboot and poweroff driver
- Non-contiguous hpm counters
- Smcntrpmf support
- Full sparse hartid support
- IPI improvements
- RFENCE improvements
- Zkr support
- Andes custom PMU support
Overall, this release mainly adds more ISA extensions, drivers and other improvements.

https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4

Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
---
 ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88 +++++++++++++++++++
 .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
 2 files changed, 92 insertions(+), 4 deletions(-)
 create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
 rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb} (90%)

diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
new file mode 100644
index 000000000000..d14e0b73a9bc
--- /dev/null
+++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
@@ -0,0 +1,88 @@
+From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00 2001
+From: Thomas Perrot <thomas.perrot@bootlin.com>
+Date: Tue, 16 Jan 2024 15:16:58 +0100
+Subject: [PATCH] Revert "firmware: Remove handling of R_RISCV_{32,64}"
+
+This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
+
+Upstream-Status: Inappropriate [oe specific]
+Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
+---
+ firmware/fw_base.S      | 22 ++++++++++++++++++++++
+ firmware/fw_base.ldS    |  6 ++++++
+ include/sbi/riscv_elf.h |  8 ++++++++
+ 3 files changed, 36 insertions(+)
+
+diff --git a/firmware/fw_base.S b/firmware/fw_base.S
+index f7763f4d9e56..60efad39b231 100644
+--- a/firmware/fw_base.S
++++ b/firmware/fw_base.S
+@@ -88,8 +88,30 @@ _try_lottery:
+ 	add	t5, t5, t2
+ 	add	t3, t3, t2
+ 	REG_S	t5, 0(t3)		/* store runtime address to the GOT entry */
++	j	5f
+
+ 3:
++	lla	t4, __dyn_sym_start
++
++4:
++	srli	t6, t5, SYM_INDEX	/* t6 <--- sym table index */
++	andi	t5, t5, 0xFF		/* t5 <--- relocation type */
++	li	t3, RELOC_TYPE
++	bne	t5, t3, 5f
++
++	/* address R_RISCV_64 or R_RISCV_32 cases*/
++	REG_L	t3, 0(t0)
++	li	t5, SYM_SIZE
++	mul	t6, t6, t5
++	add	s5, t4, t6
++	REG_L	t6, (REGBYTES * 2)(t0)	/* t0 <-- addend */
++	REG_L	t5, REGBYTES(s5)
++	add	t5, t5, t6
++	add	t5, t5, t2		/* t5 <-- location to fix up in RAM */
++	add	t3, t3, t2		/* t3 <-- location to fix up in RAM */
++	REG_S	t5, 0(t3)		/* store runtime address to the variable */
++
++5:
+ 	addi	t0, t0, (REGBYTES * 3)
+ 	blt	t0, t1, 2b
+ 	j	_relocate_done
+diff --git a/firmware/fw_base.ldS b/firmware/fw_base.ldS
+index c15ccdbf6612..88b8dfd9cb56 100644
+--- a/firmware/fw_base.ldS
++++ b/firmware/fw_base.ldS
+@@ -40,6 +40,12 @@
+
+ 	. = ALIGN(0x1000); /* Ensure next section is page aligned */
+
++	.dynsym : {
++		PROVIDE(__dyn_sym_start = .);
++		*(.dynsym)
++		PROVIDE(__dyn_sym_end = .);
++	}
++
+ 	.rela.dyn : {
+ 		PROVIDE(__rel_dyn_start = .);
+ 		*(.rela*)
+diff --git a/include/sbi/riscv_elf.h b/include/sbi/riscv_elf.h
+index ed361e346155..3b62c38b4a2d 100644
+--- a/include/sbi/riscv_elf.h
++++ b/include/sbi/riscv_elf.h
+@@ -1,6 +1,14 @@
+ #ifndef __RISCV_ELF_H__
+ #define __RISCV_ELF_H__
+
++#include <sbi/riscv_asm.h>
++
++#define R_RISCV_32		1
++#define R_RISCV_64		2
+ #define R_RISCV_RELATIVE	3
+
++#define RELOC_TYPE		__REG_SEL(R_RISCV_64, R_RISCV_32)
++#define SYM_INDEX		__REG_SEL(0x20,	0x8)
++#define SYM_SIZE		__REG_SEL(0x18,0x10)
++
+ #endif
+--
+2.43.0
diff --git a/meta/recipes-bsp/opensbi/opensbi_1.3.bb b/meta/recipes-bsp/opensbi/opensbi_1.4.bb
similarity index 90%
rename from meta/recipes-bsp/opensbi/opensbi_1.3.bb
rename to meta/recipes-bsp/opensbi/opensbi_1.4.bb
index f01cae34d1ff..6cb1a2680f2a 100644
--- a/meta/recipes-bsp/opensbi/opensbi_1.3.bb
+++ b/meta/recipes-bsp/opensbi/opensbi_1.4.bb
@@ -8,8 +8,10 @@ require opensbi-payloads.inc
 
 inherit autotools-brokensep deploy
 
-SRCREV = "057eb10b6d523540012e6947d5c9f63e95244e94"
-SRC_URI = "git://github.com/riscv/opensbi.git;branch=release-1.3.x;protocol=https"
+SRCREV = "a2b255b88918715173942f2c5e1f97ac9e90c877"
+SRC_URI = "git://github.com/riscv/opensbi.git;branch=master;protocol=https \
+	   file://0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch \
+	  "
 
 S = "${WORKDIR}/git"
 
@@ -43,5 +45,3 @@ FILES:${PN} += "/share/opensbi/*/${RISCV_SBI_PLAT}/firmware/fw_dynamic.*"
 
 COMPATIBLE_HOST = "(riscv64|riscv32).*"
 INHIBIT_PACKAGE_STRIP = "1"
-
-SECURITY_CFLAGS = ""
-- 
2.43.0

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Richard Purdie]

On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
lists.openembedded.org wrote:
> From: Thomas Perrot <thomas.perrot@bootlin.com>
> 
> This release has:
> - Synopsys DesignWare APB GPIO driver
> - Zicntr and Zihpm support
> - Console print improvements
> - Smepmp support
> - Simple FDT based syscon regmap driver
> - Syscon based reboot and poweroff driver
> - Non-contiguous hpm counters
> - Smcntrpmf support
> - Full sparse hartid support
> - IPI improvements
> - RFENCE improvements
> - Zkr support
> - Andes custom PMU support
> Overall, this release mainly adds more ISA extensions, drivers and other improvements.
> 
> https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> 
> Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>

This all seems reasonable but why is there a revert below?

> ---
>  ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88 +++++++++++++++++++
>  .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
>  2 files changed, 92 insertions(+), 4 deletions(-)
>  create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
>  rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb} (90%)
> 
> diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
> new file mode 100644
> index 000000000000..d14e0b73a9bc
> --- /dev/null
> +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-R_RISCV_-32-64.patch
> @@ -0,0 +1,88 @@
> +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00 2001
> +From: Thomas Perrot <thomas.perrot@bootlin.com>
> +Date: Tue, 16 Jan 2024 15:16:58 +0100
> +Subject: [PATCH] Revert "firmware: Remove handling of R_RISCV_{32,64}"
> +
> +This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> +
> +Upstream-Status: Inappropriate [oe specific]
> +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>

If we're going to revert an upstream change, we need to say *why*. We
need to revert it for some OE specific reason but what is that reason?

We cannot take patches marked as Inappropriate without much more
information.

Cheers,

Richard

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Thomas Perrot]

[-- Attachment #1: Type: text/plain, Size: 3557 bytes --]

Hello Richard,

On Fri, 2024-01-19 at 17:22 +0000, Richard Purdie wrote:
> On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
> lists.openembedded.org wrote:
> > From: Thomas Perrot <thomas.perrot@bootlin.com>
> > 
> > This release has:
> > - Synopsys DesignWare APB GPIO driver
> > - Zicntr and Zihpm support
> > - Console print improvements
> > - Smepmp support
> > - Simple FDT based syscon regmap driver
> > - Syscon based reboot and poweroff driver
> > - Non-contiguous hpm counters
> > - Smcntrpmf support
> > - Full sparse hartid support
> > - IPI improvements
> > - RFENCE improvements
> > - Zkr support
> > - Andes custom PMU support
> > Overall, this release mainly adds more ISA extensions, drivers and
> > other improvements.
> > 
> > https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> > 
> > Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> 
> This all seems reasonable but why is there a revert below?

I reverted it otherwise the following link issue occurs:

"build/tmp-glibc/work/riscv64-oe-linux/opensbi/1.4/recipe-sysroot-
native/usr/bin/riscv64-oe-linux/../../libexec/riscv64-oe-
linux/gcc/riscv64-oe-linux/13.2.0/ld.bfd:/src/build/tmp-
glibc/work/riscv64-oe-
linux/opensbi/1.4/git/build/platform/generic/firmware/fw_dynamic.elf.ld
:54: undefined section `.dynsym' referenced in expression"

> 
> > ---
> >  ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88
> > +++++++++++++++++++
> >  .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
> >  2 files changed, 92 insertions(+), 4 deletions(-)
> >  create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > firmware-Remove-handling-of-R_RISCV_-32-64.patch
> >  rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb}
> > (90%)
> > 
> > diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-
> > Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-
> > bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-
> > R_RISCV_-32-64.patch
> > new file mode 100644
> > index 000000000000..d14e0b73a9bc
> > --- /dev/null
> > +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-
> > handling-of-R_RISCV_-32-64.patch
> > @@ -0,0 +1,88 @@
> > +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00
> > 2001
> > +From: Thomas Perrot <thomas.perrot@bootlin.com>
> > +Date: Tue, 16 Jan 2024 15:16:58 +0100
> > +Subject: [PATCH] Revert "firmware: Remove handling of
> > R_RISCV_{32,64}"
> > +
> > +This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> > +
> > +Upstream-Status: Inappropriate [oe specific]
> > +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> 
> If we're going to revert an upstream change, we need to say *why*. We
> need to revert it for some OE specific reason but what is that
> reason?
> 

I agree, maybe you have any advice to fix it in another way?

Kind regards,
Thomas Perrot

> We cannot take patches marked as Inappropriate without much more
> information.
> 
> Cheers,
> 
> Richard
> 
> 
> -=-=-=-=-=-=-=-=-=-=-=-
> Links: You receive all messages sent to this group.
> View/Reply Online (#194059):
> https://lists.openembedded.org/g/openembedded-core/message/194059
> Mute This Topic: https://lists.openembedded.org/mt/103782707/5443093
> Group Owner: openembedded-core+owner@lists.openembedded.org
> Unsubscribe:
> https://lists.openembedded.org/g/openembedded-core/unsub [
> thomas.perrot@bootlin.com]
> -=-=-=-=-=-=-=-=-=-=-=-
> 

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 659 bytes --]

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Richard Purdie]

On Fri, 2024-01-19 at 18:48 +0100, Thomas Perrot wrote:
> Hello Richard,
> 
> On Fri, 2024-01-19 at 17:22 +0000, Richard Purdie wrote:
> > On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
> > lists.openembedded.org wrote:
> > > From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > 
> > > This release has:
> > > - Synopsys DesignWare APB GPIO driver
> > > - Zicntr and Zihpm support
> > > - Console print improvements
> > > - Smepmp support
> > > - Simple FDT based syscon regmap driver
> > > - Syscon based reboot and poweroff driver
> > > - Non-contiguous hpm counters
> > > - Smcntrpmf support
> > > - Full sparse hartid support
> > > - IPI improvements
> > > - RFENCE improvements
> > > - Zkr support
> > > - Andes custom PMU support
> > > Overall, this release mainly adds more ISA extensions, drivers and
> > > other improvements.
> > > 
> > > https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> > > 
> > > Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > 
> > This all seems reasonable but why is there a revert below?
> 
> I reverted it otherwise the following link issue occurs:
> 
> "build/tmp-glibc/work/riscv64-oe-linux/opensbi/1.4/recipe-sysroot-
> native/usr/bin/riscv64-oe-linux/../../libexec/riscv64-oe-
> linux/gcc/riscv64-oe-linux/13.2.0/ld.bfd:/src/build/tmp-
> glibc/work/riscv64-oe-
> linux/opensbi/1.4/git/build/platform/generic/firmware/fw_dynamic.elf.ld
> :54: undefined section `.dynsym' referenced in expression"

The commit should at least mention the issue.

I'm also not sure that is "Inappropriate", more like "Pending" as in
needs further investigation.

> > > ---
> > >  ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88
> > > +++++++++++++++++++
> > >  .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
> > >  2 files changed, 92 insertions(+), 4 deletions(-)
> > >  create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > > firmware-Remove-handling-of-R_RISCV_-32-64.patch
> > >  rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb}
> > > (90%)
> > > 
> > > diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-
> > > Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-
> > > bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-
> > > R_RISCV_-32-64.patch
> > > new file mode 100644
> > > index 000000000000..d14e0b73a9bc
> > > --- /dev/null
> > > +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-
> > > handling-of-R_RISCV_-32-64.patch
> > > @@ -0,0 +1,88 @@
> > > +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00
> > > 2001
> > > +From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > +Date: Tue, 16 Jan 2024 15:16:58 +0100
> > > +Subject: [PATCH] Revert "firmware: Remove handling of
> > > R_RISCV_{32,64}"
> > > +
> > > +This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> > > +
> > > +Upstream-Status: Inappropriate [oe specific]
> > > +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > 
> > If we're going to revert an upstream change, we need to say *why*. We
> > need to revert it for some OE specific reason but what is that
> > reason?
> > 
> 
> I agree, maybe you have any advice to fix it in another way?

I have no idea, I've copied Khem who might or might know who to talk
to.

Cheers,

Richard

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Alexandre Belloni]

On 19/01/2024 19:24:44+0000, Richard Purdie wrote:
> On Fri, 2024-01-19 at 18:48 +0100, Thomas Perrot wrote:
> > Hello Richard,
> > 
> > On Fri, 2024-01-19 at 17:22 +0000, Richard Purdie wrote:
> > > On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
> > > lists.openembedded.org wrote:
> > > > From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > 
> > > > This release has:
> > > > - Synopsys DesignWare APB GPIO driver
> > > > - Zicntr and Zihpm support
> > > > - Console print improvements
> > > > - Smepmp support
> > > > - Simple FDT based syscon regmap driver
> > > > - Syscon based reboot and poweroff driver
> > > > - Non-contiguous hpm counters
> > > > - Smcntrpmf support
> > > > - Full sparse hartid support
> > > > - IPI improvements
> > > > - RFENCE improvements
> > > > - Zkr support
> > > > - Andes custom PMU support
> > > > Overall, this release mainly adds more ISA extensions, drivers and
> > > > other improvements.
> > > > 
> > > > https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> > > > 
> > > > Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > 
> > > This all seems reasonable but why is there a revert below?
> > 
> > I reverted it otherwise the following link issue occurs:
> > 
> > "build/tmp-glibc/work/riscv64-oe-linux/opensbi/1.4/recipe-sysroot-
> > native/usr/bin/riscv64-oe-linux/../../libexec/riscv64-oe-
> > linux/gcc/riscv64-oe-linux/13.2.0/ld.bfd:/src/build/tmp-
> > glibc/work/riscv64-oe-
> > linux/opensbi/1.4/git/build/platform/generic/firmware/fw_dynamic.elf.ld
> > :54: undefined section `.dynsym' referenced in expression"
> 
> The commit should at least mention the issue.
> 
> I'm also not sure that is "Inappropriate", more like "Pending" as in
> needs further investigation.

The original commit states that relocations don't need to be handled
because everything is linked statically which is not true in our case,
else we wouldn't be trying to link dynamically.

> 
> > > > ---
> > > > �...re-Remove-handling-of-R_RISCV_-32-64.patch | 88
> > > > +++++++++++++++++++
> > > > �.../{opensbi_1.3.bb => opensbi_1.4.bb}������� |� 8 +-
> > > > �2 files changed, 92 insertions(+), 4 deletions(-)
> > > > �create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > > > firmware-Remove-handling-of-R_RISCV_-32-64.patch
> > > > �rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb}
> > > > (90%)
> > > > 
> > > > diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-
> > > > Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-
> > > > bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-
> > > > R_RISCV_-32-64.patch
> > > > new file mode 100644
> > > > index 000000000000..d14e0b73a9bc
> > > > --- /dev/null
> > > > +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-
> > > > handling-of-R_RISCV_-32-64.patch
> > > > @@ -0,0 +1,88 @@
> > > > +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00
> > > > 2001
> > > > +From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > +Date: Tue, 16 Jan 2024 15:16:58 +0100
> > > > +Subject: [PATCH] Revert "firmware: Remove handling of
> > > > R_RISCV_{32,64}"
> > > > +
> > > > +This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> > > > +
> > > > +Upstream-Status: Inappropriate [oe specific]
> > > > +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > 
> > > If we're going to revert an upstream change, we need to say *why*. We
> > > need to revert it for some OE specific reason but what is that
> > > reason?
> > > 
> > 
> > I agree, maybe you have any advice to fix it in another way?
> 
> I have no idea, I've copied Khem who might or might know who to talk
> to.
> 
> Cheers,
> 
> Richard

> 
> -=-=-=-=-=-=-=-=-=-=-=-
> Links: You receive all messages sent to this group.
> View/Reply Online (#194065): https://lists.openembedded.org/g/openembedded-core/message/194065
> Mute This Topic: https://lists.openembedded.org/mt/103782707/3617179
> Group Owner: openembedded-core+owner@lists.openembedded.org
> Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [alexandre.belloni@bootlin.com]
> -=-=-=-=-=-=-=-=-=-=-=-
> 


-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Khem Raj]

On Fri, Jan 19, 2024 at 2:00 PM Alexandre Belloni
<alexandre.belloni@bootlin.com> wrote:
>
>
> On 19/01/2024 19:24:44+0000, Richard Purdie wrote:
> > On Fri, 2024-01-19 at 18:48 +0100, Thomas Perrot wrote:
> > > Hello Richard,
> > >
> > > On Fri, 2024-01-19 at 17:22 +0000, Richard Purdie wrote:
> > > > On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
> > > > lists.openembedded.org wrote:
> > > > > From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > >
> > > > > This release has:
> > > > > - Synopsys DesignWare APB GPIO driver
> > > > > - Zicntr and Zihpm support
> > > > > - Console print improvements
> > > > > - Smepmp support
> > > > > - Simple FDT based syscon regmap driver
> > > > > - Syscon based reboot and poweroff driver
> > > > > - Non-contiguous hpm counters
> > > > > - Smcntrpmf support
> > > > > - Full sparse hartid support
> > > > > - IPI improvements
> > > > > - RFENCE improvements
> > > > > - Zkr support
> > > > > - Andes custom PMU support
> > > > > Overall, this release mainly adds more ISA extensions, drivers and
> > > > > other improvements.
> > > > >
> > > > > https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> > > > >
> > > > > Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > >
> > > > This all seems reasonable but why is there a revert below?
> > >
> > > I reverted it otherwise the following link issue occurs:
> > >
> > > "build/tmp-glibc/work/riscv64-oe-linux/opensbi/1.4/recipe-sysroot-
> > > native/usr/bin/riscv64-oe-linux/../../libexec/riscv64-oe-
> > > linux/gcc/riscv64-oe-linux/13.2.0/ld.bfd:/src/build/tmp-
> > > glibc/work/riscv64-oe-
> > > linux/opensbi/1.4/git/build/platform/generic/firmware/fw_dynamic.elf.ld
> > > :54: undefined section `.dynsym' referenced in expression"
> >
> > The commit should at least mention the issue.
> >
> > I'm also not sure that is "Inappropriate", more like "Pending" as in
> > needs further investigation.
>
> The original commit states that relocations don't need to be handled
> because everything is linked statically which is not true in our case,
> else we wouldn't be trying to link dynamically.
>
> >
> > > > > ---
> > > > >  ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88
> > > > > +++++++++++++++++++
> > > > >  .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
> > > > >  2 files changed, 92 insertions(+), 4 deletions(-)
> > > > >  create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > > > > firmware-Remove-handling-of-R_RISCV_-32-64.patch
> > > > >  rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb => opensbi_1.4.bb}
> > > > > (90%)
> > > > >
> > > > > diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-
> > > > > Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-
> > > > > bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-of-
> > > > > R_RISCV_-32-64.patch
> > > > > new file mode 100644
> > > > > index 000000000000..d14e0b73a9bc
> > > > > --- /dev/null
> > > > > +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-firmware-Remove-
> > > > > handling-of-R_RISCV_-32-64.patch
> > > > > @@ -0,0 +1,88 @@
> > > > > +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17 00:00:00
> > > > > 2001
> > > > > +From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > > +Date: Tue, 16 Jan 2024 15:16:58 +0100
> > > > > +Subject: [PATCH] Revert "firmware: Remove handling of
> > > > > R_RISCV_{32,64}"
> > > > > +
> > > > > +This reverts commit 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> > > > > +
> > > > > +Upstream-Status: Inappropriate [oe specific]
> > > > > +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > >
> > > > If we're going to revert an upstream change, we need to say *why*. We
> > > > need to revert it for some OE specific reason but what is that
> > > > reason?
> > > >
> > >
> > > I agree, maybe you have any advice to fix it in another way?
> >
> > I have no idea, I've copied Khem who might or might know who to talk
> > to.

We have so far build a non-PIC version but I would suggest we start
doing PIC version
set FW_PIC=y in EXTRA_OEMAKE and try it out.

> >
> > Cheers,
> >
> > Richard
>
> >
> > -=-=-=-=-=-=-=-=-=-=-=-
> > Links: You receive all messages sent to this group.
> > View/Reply Online (#194065): https://lists.openembedded.org/g/openembedded-core/message/194065
> > Mute This Topic: https://lists.openembedded.org/mt/103782707/3617179
> > Group Owner: openembedded-core+owner@lists.openembedded.org
> > Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [alexandre.belloni@bootlin.com]
> > -=-=-=-=-=-=-=-=-=-=-=-
> >
>
>
> --
> Alexandre Belloni, co-owner and COO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com

Re: [OE-core][PATCH] opensbi: bump to 1.4

[Thomas Perrot]

[-- Attachment #1: Type: text/plain, Size: 5526 bytes --]

Hello Khem,

On Fri, 2024-01-19 at 15:56 -0800, Khem Raj wrote:
> On Fri, Jan 19, 2024 at 2:00 PM Alexandre Belloni
> <alexandre.belloni@bootlin.com> wrote:
> > 
> > 
> > On 19/01/2024 19:24:44+0000, Richard Purdie wrote:
> > > On Fri, 2024-01-19 at 18:48 +0100, Thomas Perrot wrote:
> > > > Hello Richard,
> > > > 
> > > > On Fri, 2024-01-19 at 17:22 +0000, Richard Purdie wrote:
> > > > > On Wed, 2024-01-17 at 10:47 +0100, Thomas Perrot via
> > > > > lists.openembedded.org wrote:
> > > > > > From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > > > 
> > > > > > This release has:
> > > > > > - Synopsys DesignWare APB GPIO driver
> > > > > > - Zicntr and Zihpm support
> > > > > > - Console print improvements
> > > > > > - Smepmp support
> > > > > > - Simple FDT based syscon regmap driver
> > > > > > - Syscon based reboot and poweroff driver
> > > > > > - Non-contiguous hpm counters
> > > > > > - Smcntrpmf support
> > > > > > - Full sparse hartid support
> > > > > > - IPI improvements
> > > > > > - RFENCE improvements
> > > > > > - Zkr support
> > > > > > - Andes custom PMU support
> > > > > > Overall, this release mainly adds more ISA extensions,
> > > > > > drivers and
> > > > > > other improvements.
> > > > > > 
> > > > > > https://github.com/riscv-software-src/opensbi/compare/v1.3.1...v1.4
> > > > > > 
> > > > > > Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > > 
> > > > > This all seems reasonable but why is there a revert below?
> > > > 
> > > > I reverted it otherwise the following link issue occurs:
> > > > 
> > > > "build/tmp-glibc/work/riscv64-oe-linux/opensbi/1.4/recipe-
> > > > sysroot-
> > > > native/usr/bin/riscv64-oe-linux/../../libexec/riscv64-oe-
> > > > linux/gcc/riscv64-oe-linux/13.2.0/ld.bfd:/src/build/tmp-
> > > > glibc/work/riscv64-oe-
> > > > linux/opensbi/1.4/git/build/platform/generic/firmware/fw_dynami
> > > > c.elf.ld
> > > > :54: undefined section `.dynsym' referenced in expression"
> > > 
> > > The commit should at least mention the issue.
> > > 
> > > I'm also not sure that is "Inappropriate", more like "Pending" as
> > > in
> > > needs further investigation.
> > 
> > The original commit states that relocations don't need to be
> > handled
> > because everything is linked statically which is not true in our
> > case,
> > else we wouldn't be trying to link dynamically.
> > 
> > > 
> > > > > > ---
> > > > > >  ...re-Remove-handling-of-R_RISCV_-32-64.patch | 88
> > > > > > +++++++++++++++++++
> > > > > >  .../{opensbi_1.3.bb => opensbi_1.4.bb}        |  8 +-
> > > > > >  2 files changed, 92 insertions(+), 4 deletions(-)
> > > > > >  create mode 100644 meta/recipes-bsp/opensbi/opensbi/0001-
> > > > > > Revert-
> > > > > > firmware-Remove-handling-of-R_RISCV_-32-64.patch
> > > > > >  rename meta/recipes-bsp/opensbi/{opensbi_1.3.bb =>
> > > > > > opensbi_1.4.bb}
> > > > > > (90%)
> > > > > > 
> > > > > > diff --git a/meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > > > > > firmware-
> > > > > > Remove-handling-of-R_RISCV_-32-64.patch b/meta/recipes-
> > > > > > bsp/opensbi/opensbi/0001-Revert-firmware-Remove-handling-
> > > > > > of-
> > > > > > R_RISCV_-32-64.patch
> > > > > > new file mode 100644
> > > > > > index 000000000000..d14e0b73a9bc
> > > > > > --- /dev/null
> > > > > > +++ b/meta/recipes-bsp/opensbi/opensbi/0001-Revert-
> > > > > > firmware-Remove-
> > > > > > handling-of-R_RISCV_-32-64.patch
> > > > > > @@ -0,0 +1,88 @@
> > > > > > +From bfe480929bcc966e1fdf5afdde8d4c22adba7f6f Mon Sep 17
> > > > > > 00:00:00
> > > > > > 2001
> > > > > > +From: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > > > +Date: Tue, 16 Jan 2024 15:16:58 +0100
> > > > > > +Subject: [PATCH] Revert "firmware: Remove handling of
> > > > > > R_RISCV_{32,64}"
> > > > > > +
> > > > > > +This reverts commit
> > > > > > 2a6d72534d44c39e1de0614970a0dad97b1c41ba.
> > > > > > +
> > > > > > +Upstream-Status: Inappropriate [oe specific]
> > > > > > +Signed-off-by: Thomas Perrot <thomas.perrot@bootlin.com>
> > > > > 
> > > > > If we're going to revert an upstream change, we need to say
> > > > > *why*. We
> > > > > need to revert it for some OE specific reason but what is
> > > > > that
> > > > > reason?
> > > > > 
> > > > 
> > > > I agree, maybe you have any advice to fix it in another way?
> > > 
> > > I have no idea, I've copied Khem who might or might know who to
> > > talk
> > > to.
> 
> We have so far build a non-PIC version but I would suggest we start
> doing PIC version
> set FW_PIC=y in EXTRA_OEMAKE and try it out.
> 

Thank you, indeed it's better.

Kind regards,
Thomas

> > > 
> > > Cheers,
> > > 
> > > Richard
> > 
> > > 
> > > -=-=-=-=-=-=-=-=-=-=-=-
> > > Links: You receive all messages sent to this group.
> > > View/Reply Online (#194065):
> > > https://lists.openembedded.org/g/openembedded-core/message/194065
> > > Mute This Topic:
> > > https://lists.openembedded.org/mt/103782707/3617179
> > > Group Owner: openembedded-core+owner@lists.openembedded.org
> > > Unsubscribe:
> > > https://lists.openembedded.org/g/openembedded-core/unsub [
> > > alexandre.belloni@bootlin.com]
> > > -=-=-=-=-=-=-=-=-=-=-=-
> > > 
> > 
> > 
> > --
> > Alexandre Belloni, co-owner and COO, Bootlin
> > Embedded Linux and Kernel engineering
> > https://bootlin.com

-- 
Thomas Perrot, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 659 bytes --]