Skip to contentSkip to conversation

Update CI to using GCC 11.0 / LLVM-11

RFC In review Tom Rini 7 participants 28 messages May 3, 2021 – Jul 15, 2021

Subsystems (4)
lore
Share & tools
lore

Tom Rini

author
#0 lore raw

cover Update CI to using GCC 11.0 / LLVM-11

Hey all,

This series does most of the work required to move to using GCC-11.0 to build in CI. The first thing is I realized I forgot to update buildman to know to where to fetch gcc-11 from, but I'll take care of that later. Next, this ends up using the last gcc-11 snapshot as the gcc-11.0 prebuilds aren't quite up on kernel.org, but will be soon.

The real issues here and I'd like some feedback on, and need some help with are:
- grub + riscv32 + gcc-11 doesn't build, and this seems known but not addressed yet. I don't know what we want to do here as it's for running some tests.
- I cannot get x86 to build. With this series the failures are:
https://source.denx.de/u-boot/u-boot/-/jobs/263745
But I can't find a good way to address that multiple definition error of _X86EMU_env that doesn't break some other platform. At heart, Chromebooks enable the emulation and we have both chromebook_link and chromebook_link64 as builds, so we have both X86_64 and !X86_64 builds. I don't know the best way to deal with this, but I also wonder if the bios_emulation code doesn't need a rather large cleaning perhaps?

Signature · 1 line
Tom
author
#1 lore raw

RFC 1/7 ARM: mvebu: a38x: Correct mismatched bound warnings

With gcc-11 we see:
drivers/ddr/marvell/a38x/ddr3_debug.c:672:47: error: argument 2 of type 'u32[5]' {aka 'unsigned int[5]'} with mismatched bound [-Werror=array-parameter=]
672 | int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],

      |                                           ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/ddr/marvell/a38x/ddr3_training_ip_engine.h:10,
                 from drivers/ddr/marvell/a38x/ddr3_init.h:17,
                 from drivers/ddr/marvell/a38x/ddr3_debug.c:6:

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h:116:47: note: previously declared as 'u32[]' {aka 'unsigned int[]'}

And similar warnings. Correct these by updating the prototype. Remove the prototype for ddr3_tip_read_pup_value as it is unused.

---
 drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +7 -3
index ab152cb455e9..55832a55405b 100644
@@ -8,6 +8,7 @@
 
 #include "ddr3_training_ip.h"
 #include "ddr3_training_ip_db.h"
+#include "mv_ddr_plat.h"
 
 #define KILLER_PATTERN_LENGTH		32
 #define EXT_ACCESS_BURST_LENGTH		8
@@ -112,9 +113,12 @@ int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type,
 int ddr3_tip_write_mrs_cmd(u32 dev_num, u32 *cs_mask_arr, enum mr_number mr_num, u32 data, u32 mask);
 int ddr3_tip_write_cs_result(u32 dev_num, u32 offset);
 int ddr3_tip_reset_fifo_ptr(u32 dev_num);
-int ddr3_tip_read_pup_value(u32 dev_num, u32 pup_values[], int reg_addr, u32 mask);
-int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr, u32 mask);
-int ddr3_tip_write_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr);
+int ddr3_tip_read_adll_value(u32 dev_num,
+			     u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			     u32 reg_addr, u32 mask);
+int ddr3_tip_write_adll_value(u32 dev_num,
+			      u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			      u32 reg_addr);
 int ddr3_tip_tune_training_params(u32 dev_num, struct tune_train_params *params);
 
 #endif /* _DDR3_TRAINING_IP_FLOW_H_ */
Signature · 1 line
2.17.1
author
#2 lore raw

Re: [RFC 1/7] ARM: mvebu: a38x: Correct mismatched bound warnings

Quoted 7 lines · Tom Rini

With gcc-11 we see:
drivers/ddr/marvell/a38x/ddr3_debug.c:672:47: error: argument 2 of type 'u32[5]' {aka 'unsigned int[5]'} with mismatched bound [-Werror=array-parameter=]
672 | int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],

      |                                           ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/ddr/marvell/a38x/ddr3_training_ip_engine.h:10,
                 from drivers/ddr/marvell/a38x/ddr3_init.h:17,
                 from drivers/ddr/marvell/a38x/ddr3_debug.c:6:
Quoted 6 lines

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h:116:47: note: previously declared as 'u32[]' {aka 'unsigned int[]'}

And similar warnings. Correct these by updating the prototype. Remove the prototype for ddr3_tip_read_pup_value as it is unused.

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
#3 lore raw

Re: [RFC 1/7] ARM: mvebu: a38x: Correct mismatched bound warnings

In reply to Tom Rini
Quoted 35 lines · On Monday 03 May 2021 16:48:53 Tom Rini wrote:

With gcc-11 we see:
drivers/ddr/marvell/a38x/ddr3_debug.c:672:47: error: argument 2 of type 'u32[5]' {aka 'unsigned int[5]'} with mismatched bound [-Werror=array-parameter=]
672 | int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],

      |                                           ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/ddr/marvell/a38x/ddr3_training_ip_engine.h:10,
                 from drivers/ddr/marvell/a38x/ddr3_init.h:17,
                 from drivers/ddr/marvell/a38x/ddr3_debug.c:6:

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h:116:47: note: previously declared as 'u32[]' {aka 'unsigned int[]'}

And similar warnings. Correct these by updating the prototype. Remove the prototype for ddr3_tip_read_pup_value as it is unused.

---
 drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +1 -2
index ab152cb455e9..55832a55405b 100644
@@ -8,6 +8,7 @@
 
 #include "ddr3_training_ip.h"
 #include "ddr3_training_ip_db.h"
+#include "mv_ddr_plat.h"
 
 #define KILLER_PATTERN_LENGTH		32
 #define EXT_ACCESS_BURST_LENGTH		8
@@ -112,9 +113,12 @@ int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type,
 int ddr3_tip_write_mrs_cmd(u32 dev_num, u32 *cs_mask_arr, enum mr_number mr_num, u32 data, u32 mask);
 int ddr3_tip_write_cs_result(u32 dev_num, u32 offset);
 int ddr3_tip_reset_fifo_ptr(u32 dev_num);
-int ddr3_tip_read_pup_value(u32 dev_num, u32 pup_values[], int reg_addr, u32 mask);
-int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr, u32 mask);

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type, · line 124

Quoted 10 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +6 -1
-int ddr3_tip_write_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr);
+int ddr3_tip_read_adll_value(u32 dev_num,
+			     u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			     u32 reg_addr, u32 mask);
+int ddr3_tip_write_adll_value(u32 dev_num,
+			      u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			      u32 reg_addr);
 int ddr3_tip_tune_training_params(u32 dev_num, struct tune_train_params *params);
 
 #endif /* _DDR3_TRAINING_IP_FLOW_H_ */

CC Stefan, Marek, Konstantin

Could you please send this patch to upstream?
https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell

Any change done in drivers/ddr/marvell/a38x cause issue in future code synchronization with upstream repository.

#4 lore raw

Re: [RFC 1/7] ARM: mvebu: a38x: Correct mismatched bound warnings

Hi Pali,
Hi Kosta,

Quoted 49 lines · Pali Rohár
Quoted 35 lines · On Monday 03 May 2021 16:48:53 Tom Rini wrote:

With gcc-11 we see:
drivers/ddr/marvell/a38x/ddr3_debug.c:672:47: error: argument 2 of type 'u32[5]' {aka 'unsigned int[5]'} with mismatched bound [-Werror=array-parameter=]
672 | int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],

       |                                           ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from drivers/ddr/marvell/a38x/ddr3_training_ip_engine.h:10,
                  from drivers/ddr/marvell/a38x/ddr3_init.h:17,
                  from drivers/ddr/marvell/a38x/ddr3_debug.c:6:

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h:116:47: note: previously declared as 'u32[]' {aka 'unsigned int[]'}

And similar warnings. Correct these by updating the prototype. Remove the prototype for ddr3_tip_read_pup_value as it is unused.

---
  drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h | 10 +++++++---
  1 file changed, 7 insertions(+), 3 deletions(-)
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +1 -2
index ab152cb455e9..55832a55405b 100644
@@ -8,6 +8,7 @@
 
  #include "ddr3_training_ip.h"
  #include "ddr3_training_ip_db.h"
+#include "mv_ddr_plat.h"
 
  #define KILLER_PATTERN_LENGTH		32
  #define EXT_ACCESS_BURST_LENGTH		8
@@ -112,9 +113,12 @@ int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type,
  int ddr3_tip_write_mrs_cmd(u32 dev_num, u32 *cs_mask_arr, enum mr_number mr_num, u32 data, u32 mask);
  int ddr3_tip_write_cs_result(u32 dev_num, u32 offset);
  int ddr3_tip_reset_fifo_ptr(u32 dev_num);
-int ddr3_tip_read_pup_value(u32 dev_num, u32 pup_values[], int reg_addr, u32 mask);
-int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr, u32 mask);

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type, · line 124

Quoted 10 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +6 -1
-int ddr3_tip_write_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr);
+int ddr3_tip_read_adll_value(u32 dev_num,
+			     u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			     u32 reg_addr, u32 mask);
+int ddr3_tip_write_adll_value(u32 dev_num,
+			      u32 pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],
+			      u32 reg_addr);
  int ddr3_tip_tune_training_params(u32 dev_num, struct tune_train_params *params);
 
  #endif /* _DDR3_TRAINING_IP_FLOW_H_ */

CC Stefan, Marek, Konstantin

Quoted 5 lines

Could you please send this patch to upstream?
https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell

Any change done in drivers/ddr/marvell/a38x cause issue in future code synchronization with upstream repository.

Sure. It absolutely makes sense to try to keep these repo's at least a bit in sync.

Kosta, could you please apply this patch to your "mv-ddr-marvell" repo as well?

Thanks,
Stefan

#5 lore raw

RE: [EXT] Re: [RFC 1/7] ARM: mvebu: a38x: Correct mismatched bound warnings

Hi Pali and Stefan,

Quoted 90 lines

-----Original Message-----
From: Stefan Roese <[email protected]>
Sent: Thursday, July 15, 2021 11:00
To: Pali Rohár <[email protected]>; Tom Rini <[email protected]>
Cc: [email protected]; Marek Behún <[email protected]>; Kostya Porotchkin <[email protected]>
Subject: [EXT] Re: [RFC 1/7] ARM: mvebu: a38x: Correct mismatched bound warnings

External Email

----------------------------------------------------------------------
Hi Pali,
Hi Kosta,

Quoted 3 lines · Pali Rohár
Quoted 2 lines · On Monday 03 May 2021 16:48:53 Tom Rini wrote:

With gcc-11 we see:
drivers/ddr/marvell/a38x/ddr3_debug.c:672:47: error: argument 2 of type

'u32[5]' {aka 'unsigned int[5]'} with mismatched bound [-Werror=array- parameter=]

Quoted 1 line
Quoted 1 line
   672 | int ddr3_tip_read_adll_value(u32 dev_num, u32

pup_values[MAX_INTERFACE_NUM * MAX_BUS_NUM],

Quoted 1 line
Quoted 1 line
       |

~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Quoted 1 line
Quoted 1 line

In file included from

drivers/ddr/marvell/a38x/ddr3_training_ip_engine.h:10,

Quoted 20 lines
Quoted 20 lines
                  from drivers/ddr/marvell/a38x/ddr3_init.h:17,
                  from drivers/ddr/marvell/a38x/ddr3_debug.c:6:

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h:116:47: note:
previously declared as 'u32[]' {aka 'unsigned int[]'}

And similar warnings. Correct these by updating the prototype. Remove the prototype for ddr3_tip_read_pup_value as it is unused.

---
  drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h | 10 +++++++---
  1 file changed, 7 insertions(+), 3 deletions(-)
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +0 -0
index ab152cb455e9..55832a55405b 100644
@@ -8,6 +8,7 @@
 
Quoted 7 lines

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h int ddr3_tip_configure_odpg(u32 dev_num, enum

Quoted 7 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +1 -0
  #include "ddr3_training_ip.h"
  #include "ddr3_training_ip_db.h"
+#include "mv_ddr_plat.h"
 
  #define KILLER_PATTERN_LENGTH		32
  #define EXT_ACCESS_BURST_LENGTH		8
@@ -112,9 +113,12 @@ int ddr3_tip_configure_odpg(u32 dev_num, enum

hws_access_type access_type,

Quoted 1 line

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h int ddr3_tip_configure_odpg(u32 dev_num, enum · line 113

Quoted 1 line
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +0 -0
  int ddr3_tip_write_mrs_cmd(u32 dev_num, u32 *cs_mask_arr, enum

mr_number mr_num, u32 data, u32 mask);

Quoted 9 lines

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h

Quoted 9 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +0 -0
  int ddr3_tip_write_cs_result(u32 dev_num, u32 offset);
  int ddr3_tip_reset_fifo_ptr(u32 dev_num); -int
 ddr3_tip_read_pup_value(u32 dev_num, u32 pup_values[], int reg_addr,

u32 mask); -int ddr3_tip_read_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr, u32 mask); -int ddr3_tip_write_adll_value(u32 dev_num, u32 pup_values[], u32 reg_addr);

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +2 -0
+int ddr3_tip_read_adll_value(u32 dev_num,
+			     u32 pup_values[MAX_INTERFACE_NUM *

MAX_BUS_NUM],

Quoted 3 lines

drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h int ddr3_tip_configure_odpg(u32 dev_num, enum hws_access_type access_type,

Quoted 3 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +3 -0
+			     u32 reg_addr, u32 mask);
+int ddr3_tip_write_adll_value(u32 dev_num,
+			      u32 pup_values[MAX_INTERFACE_NUM *

MAX_BUS_NUM],

Quoted 11 lines
Quoted 5 lines
drivers/ddr/marvell/a38x/ddr3_training_ip_flow.h +1 -0
+			      u32 reg_addr);
  int ddr3_tip_tune_training_params(u32 dev_num, struct

tune_train_params *params);

  #endif /* _DDR3_TRAINING_IP_FLOW_H_ */

CC Stefan, Marek, Konstantin

Could you please send this patch to upstream?
https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_Marvel
lEmbeddedProcessors_mv-2Dddr-

2Dmarvell&d=DwIDaQ&c=nKjWec2b6R0mOyPaz7xt

Quoted 1 line

fQ&r=-

N9sN4p5NSr0JGQoQ_2UCOgAqajG99W1EbSOww0WU8o&m=uCuiUzuJ- xKl2xMLf1k

Quoted 1 line

3uSHRcVkxkdrRn7NXczjBNrk&s=IXYIGGlj-Y6Z9-VWGf8iWAiGQAbAYt-

3zROqOhn32Sk

Quoted 2 lines

&e=

Quoted 8 lines
Quoted 2 lines

Any change done in drivers/ddr/marvell/a38x cause issue in future code synchronization with upstream repository.

Sure. It absolutely makes sense to try to keep these repo's at least a bit in sync.

Kosta, could you please apply this patch to your "mv-ddr-marvell" repo as well?

[KP] I am porting this to our development stream and once merged, will issue a patch to Github.

Regards
Kosta

Quoted 3 lines

Thanks,
Stefan

author
#6 lore raw

RFC 2/7 freescale: Drop unnecessary cpld_data_t non-typedef

In reply to Tom Rini

In some board cpld.h files the definition of the cpld_data struct not-quite makes a typedef for cpld_data_t. This problem is caught with gcc-11 as a multiple definition error. As there are no users of this non-typedef, fix this by not declaring it one to begin with.

Cc Priyanka Jain <[email protected]>
Cc Shengzhou Liu <[email protected]>
---
 board/freescale/t102xrdb/cpld.h | 2 +-
 board/freescale/t104xrdb/cpld.h | 3 +--
 board/freescale/t208xrdb/cpld.h | 2 +-
 3 files changed, 3 insertions(+), 4 deletions(-)
board/freescale/t102xrdb/cpld.h +1 -1
index c05f536806fb..bd40cc319a89 100644
@@ -21,7 +21,7 @@ struct cpld_data {
 	u8 boot_override;	/* 0x18 - Boot override register */
 	u8 boot_config1;	/* 0x19 - Boot config override register*/
 	u8 boot_config2;	/* 0x1A - Boot config override register*/
-} cpld_data_t;
+};
 
 
 /* Pointer to the CPLD register set */
board/freescale/t104xrdb/cpld.h +1 -2
index a816aef10a42..769883f9461f 100644
@@ -30,8 +30,7 @@ struct cpld_data {
 	u8 boot_override;	/* 0x18 - Boot override register */
 	u8 boot_config1;	/* 0x19 - Boot config override register*/
 	u8 boot_config2;	/* 0x1A - Boot config override register*/
-} cpld_data_t;
-
+};
 
 /* Pointer to the CPLD register set */
 
board/freescale/t208xrdb/cpld.h +1 -1
index bd6c203742e3..a623b1811faf 100644
@@ -21,7 +21,7 @@ struct cpld_data {
 	u8 boot_or;		/* 0x16 - Boot config override register */
 	u8 boot_cfg1;		/* 0x17 - Boot configuration register 1 */
 	u8 boot_cfg2;		/* 0x18 - Boot configuration register 2 */
-} cpld_data_t;
+};
 
 u8 cpld_read(unsigned int reg);
 void cpld_write(unsigned int reg, u8 value);
Signature · 1 line
2.17.1
author
#7 lore raw

Re: [RFC 2/7] freescale: Drop unnecessary cpld_data_t non-typedef

Quoted 8 lines · Tom Rini

In some board cpld.h files the definition of the cpld_data struct not-quite makes a typedef for cpld_data_t. This problem is caught with gcc-11 as a multiple definition error. As there are no users of this non-typedef, fix this by not declaring it one to begin with.

Cc Priyanka Jain <[email protected]>
Cc Shengzhou Liu <[email protected]>

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
author
#8 lore raw

RFC 3/7 eb_cpu5282: Declare diplay_width / display_height as externs

In reply to Tom Rini

The board code here references the display_width / display_height variables set in the video driver, declare these as externs as gcc-11 will notice and lead to a multiple definition error.

---
 board/BuS/eb_cpu5282/eb_cpu5282.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
board/BuS/eb_cpu5282/eb_cpu5282.c +2 -2
index 153ed0e963b0..144a08922b10 100644
@@ -22,8 +22,8 @@
 DECLARE_GLOBAL_DATA_PTR;
 
 #if IS_ENABLED(CONFIG_VIDEO_VCXK)
-unsigned long display_width;
-unsigned long display_height;
+extern unsigned long display_width;
+extern unsigned long display_height;
 #endif
 
 /*---------------------------------------------------------------------------*/
Signature · 1 line
2.17.1
author
#9 lore raw

Re: [RFC 3/7] eb_cpu5282: Declare diplay_width / display_height as externs

Quoted 5 lines · Tom Rini

The board code here references the display_width / display_height variables set in the video driver, declare these as externs as gcc-11 will notice and lead to a multiple definition error.

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
author
#10 lore raw

RFC 4/7 pinctrl: mscc: Fix multiple definition error

In reply to Tom Rini

With gcc-11 we get a multiple errors here as the declarations for mscc_pinctrl_ops and mscc_gpio_ops are missing an extern.

CC Gregory CLEMENT <[email protected]>
Cc Lars Povlsen <[email protected]>
Cc Horatiu Vultur <[email protected]>
---
 drivers/pinctrl/mscc/mscc-common.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/pinctrl/mscc/mscc-common.h +2 -2
index 3c5c1faf840f..49c84a9f410b 100644
@@ -61,6 +61,6 @@ int mscc_pinctrl_probe(struct udevice *dev, int num_func,
 		       const struct mscc_pin_data *mscc_pins, int num_pins,
 		       char * const *function_names,
 		       const unsigned long *mscc_gpios);
-const struct pinctrl_ops mscc_pinctrl_ops;
+extern const struct pinctrl_ops mscc_pinctrl_ops;
 
-const struct dm_gpio_ops mscc_gpio_ops;
+extern const struct dm_gpio_ops mscc_gpio_ops;
Signature · 1 line
2.17.1
#11 lore raw

RFC 4/7 pinctrl: mscc: Fix multiple definition error

Quoted 3 lines · Tom Rini

With gcc-11 we get a multiple errors here as the declarations for mscc_pinctrl_ops and mscc_gpio_ops are missing an extern.

Quoted 25 lines
CC Gregory CLEMENT <[email protected]>
Cc Lars Povlsen <[email protected]>
Cc Horatiu Vultur <[email protected]>
---
 drivers/pinctrl/mscc/mscc-common.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/pinctrl/mscc/mscc-common.h +2 -3
index 3c5c1faf840f..49c84a9f410b 100644
@@ -61,6 +61,6 @@ int mscc_pinctrl_probe(struct udevice *dev, int num_func,
                       const struct mscc_pin_data *mscc_pins, int num_pins,
                       char * const *function_names,
                       const unsigned long *mscc_gpios);
-const struct pinctrl_ops mscc_pinctrl_ops;
+extern const struct pinctrl_ops mscc_pinctrl_ops;
 
-const struct dm_gpio_ops mscc_gpio_ops;
+extern const struct dm_gpio_ops mscc_gpio_ops;
--

2.17.1

Signature · 1 line
/Horatiu
#12 lore raw

RFC 4/7 pinctrl: mscc: Fix multiple definition error

In reply to Tom Rini

Am Montag, den 03.05.2021, 16:48 -0400 schrieb Tom Rini:

Quoted 11 lines

With gcc-11 we get a multiple errors here as the declarations for mscc_pinctrl_ops and mscc_gpio_ops are missing an extern.

CC Gregory CLEMENT <[email protected]>
Cc Lars Povlsen <[email protected]>
Cc Horatiu Vultur <[email protected]>
---
 drivers/pinctrl/mscc/mscc-common.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

-- Daniel

author
#13 lore raw

Re: [RFC 4/7] pinctrl: mscc: Fix multiple definition error

In reply to Tom Rini
Quoted 9 lines · Tom Rini

With gcc-11 we get a multiple errors here as the declarations for mscc_pinctrl_ops and mscc_gpio_ops are missing an extern.

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
author
#14 lore raw

RFC 5/7 Makefile: Disable gcc-10.0 warnings

In reply to Tom Rini

Based on 5c45de21a2223, 44720996e2d79 and 5a76021c2eff7 in Linux

---
 Makefile | 4 ++++
 1 file changed, 4 insertions(+)
Makefile +4 -0
index 404977efa527..aaa9a0495e69 100644
@@ -687,6 +687,10 @@ KBUILD_CFLAGS += $(call cc-option,-fno-delete-null-pointer-checks)
 # disable stringop warnings in gcc 8+
 KBUILD_CFLAGS += $(call cc-disable-warning, stringop-truncation)
 
+KBUILD_CFLAGS += $(call cc-disable-warning, zero-length-bounds)
+KBUILD_CFLAGS += $(call cc-disable-warning, array-bounds)
+KBUILD_CFLAGS += $(call cc-disable-warning, stringop-overflow)
+
 # Enabled with W=2, disabled by default as noisy
 KBUILD_CFLAGS += $(call cc-disable-warning, maybe-uninitialized)
 
Signature · 1 line
2.17.1
author
#15 lore raw

Re: [RFC 5/7] Makefile: Disable gcc-10.0 warnings

Quoted 3 lines · Tom Rini

Based on 5c45de21a2223, 44720996e2d79 and 5a76021c2eff7 in Linux

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
author
#16 lore raw

RFC 6/7 bootstage: Eliminate when not enabled

In reply to Tom Rini

When we do not have bootstage enabled, rather than include an empty dummy function, we just don't reference it. This saves us space in some tight builds. This also shows a few cases where show_boot_progress was incorrectly guarded before.

---
Moving to gcc-11 causes at least one well maintained and used board to now run out of space, and this saves us 400 bytes in SPL.

---
 arch/x86/cpu/cpu.c                    | 2 ++
 board/Seagate/dockstar/dockstar.c     | 2 ++
 board/Seagate/goflexhome/goflexhome.c | 2 ++
 board/bosch/shc/board.c               | 6 +++---
 board/buffalo/lsxl/lsxl.c             | 2 +-
 board/k+p/kp_imx53/kp_imx53.c         | 2 ++
 board/st/stv0991/stv0991.c            | 2 +-
 common/init/board_init.c              | 2 ++
 common/spl/spl.c                      | 2 ++
 include/bootstage.h                   | 4 +++-
 10 files changed, 20 insertions(+), 6 deletions(-)
arch/x86/cpu/cpu.c +2 -0
index 9c4edfcbfdaa..01dece5769c8 100644
@@ -178,10 +178,12 @@ int default_print_cpuinfo(void)
 	return 0;
 }
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 void show_boot_progress(int val)
 {
 	outb(val, POST_PORT);
 }
+#endif
 
 #if !defined(CONFIG_SYS_COREBOOT) && !defined(CONFIG_EFI_STUB)
 /*
board/Seagate/dockstar/dockstar.c +2 -0
index 380e37be5502..fb6919315804 100644
@@ -140,6 +140,7 @@ void reset_phy(void)
 }
 #endif /* CONFIG_RESET_PHY_R */
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 #define GREEN_LED	(1 << 14)
 #define ORANGE_LED	(1 << 15)
 #define BOTH_LEDS	(GREEN_LED | ORANGE_LED)
@@ -169,3 +170,4 @@ void show_boot_progress(int val)
 		break;
 	}
 }
+#endif
board/Seagate/goflexhome/goflexhome.c +2 -0
index 4c19fa7195e8..af8cab7bdc73 100644
@@ -142,6 +142,7 @@ void reset_phy(void)
 }
 #endif /* CONFIG_RESET_PHY_R */
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 #define GREEN_LED	(1 << 14)
 #define ORANGE_LED	(1 << 15)
 #define BOTH_LEDS	(GREEN_LED | ORANGE_LED)
@@ -175,3 +176,4 @@ void show_boot_progress(int val)
 		break;
 	}
 }
+#endif
board/bosch/shc/board.c +3 -3
index e89378188728..01d9ce4f71f9 100644
@@ -188,7 +188,7 @@ static void __maybe_unused leds_set_booting(void)
 /*
  * Function to set the LEDs in the state "Bootloader error"
  */
-static void leds_set_failure(int state)
+static void __maybe_unused leds_set_failure(int state)
 {
 #if defined(CONFIG_B_SAMPLE)
 	/* Turn all blue and green LEDs off */
@@ -479,7 +479,7 @@ int board_eth_init(struct bd_info *bis)
 }
 #endif
 
-#ifdef CONFIG_SHOW_BOOT_PROGRESS
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 static void bosch_check_reset_pin(void)
 {
 	if (readl(GPIO1_BASE + OMAP_GPIO_IRQSTATUS_SET_0) & RESET_MASK) {
@@ -525,9 +525,9 @@ void show_boot_progress(int val)
 		break;
 	}
 }
+#endif
 
 void arch_preboot_os(void)
 {
 	leds_set_finish();
 }
-#endif
board/buffalo/lsxl/lsxl.c +1 -1
index 34be601fdde6..738b6bc25cea 100644
@@ -271,7 +271,7 @@ int misc_init_r(void)
 }
 #endif
 
-#ifdef CONFIG_SHOW_BOOT_PROGRESS
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 void show_boot_progress(int progress)
 {
 	if (progress > 0)
board/k+p/kp_imx53/kp_imx53.c +2 -0
index cc8118b4ad4f..7c3a695cb258 100644
@@ -155,6 +155,7 @@ int board_late_init(void)
 	return ret;
 }
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 #define GPIO_DR 0x0
 #define GPIO_GDIR 0x4
 #define GPIO_ALT1 0x1
@@ -203,3 +204,4 @@ void show_boot_progress(int status)
 		gpio_direction_output(LED_RED, 1);
 	}
 }
+#endif
board/st/stv0991/stv0991.c +1 -1
index 3b0de82db283..57ca9f659c16 100644
@@ -37,7 +37,7 @@ U_BOOT_DRVINFO(stv09911_serials) = {
 };
 #endif
 
-#ifdef CONFIG_SHOW_BOOT_PROGRESS
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 void show_boot_progress(int progress)
 {
 	printf("%i\n", progress);
common/init/board_init.c +2 -0
index 3f183ee11326..0965b96fa3ad 100644
@@ -166,7 +166,9 @@ void board_init_f_init_reserve(ulong base)
 		board_init_f_init_stack_protection();
 }
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 /*
  * Board-specific Platform code can reimplement show_boot_progress () if needed
  */
 __weak void show_boot_progress(int val) {}
+#endif
common/spl/spl.c +2 -0
index a0a608fd7725..eba77cace6df 100644
@@ -58,10 +58,12 @@ binman_sym_declare(ulong, spl, size);
 /* Define board data structure */
 static struct bd_info bdata __attribute__ ((section(".data")));
 
+#if CONFIG_IS_ENABLED(BOOTSTAGE)
 /*
  * Board-specific Platform code can reimplement show_boot_progress () if needed
  */
 __weak void show_boot_progress(int val) {}
+#endif
 
 #if defined(CONFIG_SPL_OS_BOOT) || CONFIG_IS_ENABLED(HANDOFF) || \
 	defined(CONFIG_SPL_ATF)
include/bootstage.h +3 -1
index 00c85fb86aa2..f837a387c8c7 100644
@@ -11,6 +11,8 @@
 #ifndef _BOOTSTAGE_H
 #define _BOOTSTAGE_H
 
+#include <linux/kconfig.h>
+
 /* Flags for each bootstage record */
 enum bootstage_flags {
 	BOOTSTAGEF_ERROR	= 1 << 0,	/* Error record */
@@ -218,7 +220,7 @@ enum bootstage_id {
  */
 ulong timer_get_boot_us(void);
 
-#if defined(USE_HOSTCC)
+#if defined(USE_HOSTCC) || !CONFIG_IS_ENABLED(BOOTSTAGE)
 #define show_boot_progress(val) do {} while (0)
 #else
 /**
Signature · 1 line
2.17.1
author
#17 lore raw

Re: [RFC 6/7] bootstage: Eliminate when not enabled

Quoted 7 lines · Tom Rini

When we do not have bootstage enabled, rather than include an empty dummy function, we just don't reference it. This saves us space in some tight builds. This also shows a few cases where show_boot_progress was incorrectly guarded before.

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
author
#18 lore raw

RFC 7/7 TEST: Move to gcc-11.0 prerelease and LLVM-11

In reply to Tom Rini

- gcc-11.0 proper is coming soon to kernel.org, use this snapshot for now to see what fails.
- grub-2.04 does not build with gcc-11.0 due to -Werror, grub-2.06-rc1 fails on riscv32-linux as seen here:
https://www.mail-archive.com/grub-devel at gnu.org/msg30736.html

Cc Heinrich Schuchardt <[email protected]>
---
 .azure-pipelines.yml    |  3 +-
 .gitlab-ci.yml          |  3 +-
 tools/docker/Dockerfile | 74 +++++++++++++++++------------------------
 3 files changed, 32 insertions(+), 48 deletions(-)
.azure-pipelines.yml +1 -2
index 59e99b8894c5..28e4d53ece71 100644
@@ -2,7 +2,7 @@ variables:
   windows_vm: vs2017-win2016
   ubuntu_vm: ubuntu-18.04
   macos_vm: macOS-10.15
-  ci_runner_image: trini/u-boot-gitlab-ci-runner:bionic-20200807-02Sep2020
+  ci_runner_image: trini/u-boot-gitlab-ci-runner:bionic-20210416-29Apr2021
   # Add '-u 0' options for Azure pipelines, otherwise we get "permission
   # denied" error when it tries to "useradd -m -u 1001 vsts_azpcontainer",
   # since our $(ci_runner_image) user is not root.
@@ -312,7 +312,6 @@ jobs:
           cp ~/grub_x86.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp ~/grub_x64.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp /opt/grub/grubriscv64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv64.efi
-          cp /opt/grub/grubriscv32.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv32.efi
           cp /opt/grub/grubaa64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm64.efi
           cp /opt/grub/grubarm.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm.efi
           virtualenv -p /usr/bin/python3 /tmp/venv
.gitlab-ci.yml +1 -2
index bff487404f30..02dc8449ecf9 100644
@@ -2,7 +2,7 @@
 
 # Grab our configured image.  The source for this is found at:
 # https://source.denx.de/u-boot/gitlab-ci-runner
-image: trini/u-boot-gitlab-ci-runner:bionic-20200807-02Sep2020
+image: trini/u-boot-gitlab-ci-runner:bionic-20210416-29Apr2021
 
 # We run some tests in different order, to catch some failures quicker.
 stages:
@@ -38,7 +38,6 @@ stages:
     - cp ~/grub_x86.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp ~/grub_x64.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp /opt/grub/grubriscv64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv64.efi
-    - cp /opt/grub/grubriscv32.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv32.efi
     - cp /opt/grub/grubaa64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm64.efi
     - cp /opt/grub/grubarm.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm.efi
     - virtualenv -p /usr/bin/python3 /tmp/venv
tools/docker/Dockerfile +30 -44

122 lines not on this page · raw message

Signature · 1 line
2.17.1
author
#19 lore raw

RFC v2 1/1 TEST: Move to gcc-11.0 prerelease and LLVM-11

- gcc-11.0 proper is coming soon to kernel.org, use this snapshot for now to see what fails.
- grub-2.04 does not build with gcc-11.0 due to -Werror, grub-2.06-rc1 fails on riscv32-linux as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html

Cc Heinrich Schuchardt <[email protected]>
Cc Simon Glass <[email protected]>
Cc Bin Meng <[email protected]>

---
Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

---
 .azure-pipelines.yml    |  3 +-
 .gitlab-ci.yml          |  3 +-
 tools/docker/Dockerfile | 74 +++++++++++++++++------------------------
 3 files changed, 32 insertions(+), 48 deletions(-)
.azure-pipelines.yml +1 -2
index 59e99b8894c5..28e4d53ece71 100644
@@ -2,7 +2,7 @@ variables:
   windows_vm: vs2017-win2016
   ubuntu_vm: ubuntu-18.04
   macos_vm: macOS-10.15
-  ci_runner_image: trini/u-boot-gitlab-ci-runner:bionic-20200807-02Sep2020
+  ci_runner_image: trini/u-boot-gitlab-ci-runner:bionic-20210416-29Apr2021
   # Add '-u 0' options for Azure pipelines, otherwise we get "permission
   # denied" error when it tries to "useradd -m -u 1001 vsts_azpcontainer",
   # since our $(ci_runner_image) user is not root.
@@ -312,7 +312,6 @@ jobs:
           cp ~/grub_x86.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp ~/grub_x64.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp /opt/grub/grubriscv64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv64.efi
-          cp /opt/grub/grubriscv32.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv32.efi
           cp /opt/grub/grubaa64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm64.efi
           cp /opt/grub/grubarm.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm.efi
           virtualenv -p /usr/bin/python3 /tmp/venv
.gitlab-ci.yml +1 -2
index bff487404f30..02dc8449ecf9 100644
@@ -2,7 +2,7 @@
 
 # Grab our configured image.  The source for this is found at:
 # https://source.denx.de/u-boot/gitlab-ci-runner
-image: trini/u-boot-gitlab-ci-runner:bionic-20200807-02Sep2020
+image: trini/u-boot-gitlab-ci-runner:bionic-20210416-29Apr2021
 
 # We run some tests in different order, to catch some failures quicker.
 stages:
@@ -38,7 +38,6 @@ stages:
     - cp ~/grub_x86.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp ~/grub_x64.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp /opt/grub/grubriscv64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv64.efi
-    - cp /opt/grub/grubriscv32.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv32.efi
     - cp /opt/grub/grubaa64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm64.efi
     - cp /opt/grub/grubarm.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm.efi
     - virtualenv -p /usr/bin/python3 /tmp/venv
tools/docker/Dockerfile +30 -44

122 lines not on this page · raw message

Signature · 1 line
2.17.1
author
#20 lore raw

RFC v3 1/1 Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

In reply to Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Changes in v2:
- None, I re-sent v1 by accident

---
 .azure-pipelines.yml    |  3 +-
 .gitlab-ci.yml          |  3 +-
 tools/docker/Dockerfile | 83 ++++++++++++++++-------------------------
 3 files changed, 35 insertions(+), 54 deletions(-)
.azure-pipelines.yml +1 -2
index 5ac4d648ec9e..3d4cd76c6061 100644
@@ -2,7 +2,7 @@ variables:
   windows_vm: vs2017-win2016
   ubuntu_vm: ubuntu-18.04
   macos_vm: macOS-10.15
-  ci_runner_image: trini/u-boot-gitlab-ci-runner:focal-20210416-10Jun2021
+  ci_runner_image: trini/u-boot-gitlab-ci-runner:focal-20210416-14Jun2021
   # Add '-u 0' options for Azure pipelines, otherwise we get "permission
   # denied" error when it tries to "useradd -m -u 1001 vsts_azpcontainer",
   # since our $(ci_runner_image) user is not root.
@@ -300,7 +300,6 @@ jobs:
           cp ~/grub_x86.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp ~/grub_x64.efi ${UBOOT_TRAVIS_BUILD_DIR}/
           cp /opt/grub/grubriscv64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv64.efi
-          cp /opt/grub/grubriscv32.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_riscv32.efi
           cp /opt/grub/grubaa64.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm64.efi
           cp /opt/grub/grubarm.efi ${UBOOT_TRAVIS_BUILD_DIR}/grub_arm.efi
           virtualenv -p /usr/bin/python3 /tmp/venv
.gitlab-ci.yml +1 -2
index 41a2c18fa180..b7c049825102 100644
@@ -2,7 +2,7 @@
 
 # Grab our configured image.  The source for this is found at:
 # https://source.denx.de/u-boot/gitlab-ci-runner
-image: trini/u-boot-gitlab-ci-runner:focal-20210416-10Jun2021
+image: trini/u-boot-gitlab-ci-runner:focal-20210416-14Jun2021
 
 # We run some tests in different order, to catch some failures quicker.
 stages:
@@ -38,7 +38,6 @@ stages:
     - cp ~/grub_x86.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp ~/grub_x64.efi $UBOOT_TRAVIS_BUILD_DIR/
     - cp /opt/grub/grubriscv64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv64.efi
-    - cp /opt/grub/grubriscv32.efi $UBOOT_TRAVIS_BUILD_DIR/grub_riscv32.efi
     - cp /opt/grub/grubaa64.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm64.efi
     - cp /opt/grub/grubarm.efi $UBOOT_TRAVIS_BUILD_DIR/grub_arm.efi
     - virtualenv -p /usr/bin/python3 /tmp/venv
tools/docker/Dockerfile +33 -50

133 lines not on this page · raw message

Signature · 1 line
2.17.1
#21 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Hi Tom,

Quoted 13 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>
Quoted 8 lines

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Did you mean x86 fails to link with LLVM 11, not GCC 11?

Quoted 9 lines

Changes in v2:
- None, I re-sent v1 by accident

---
 .azure-pipelines.yml    |  3 +-
 .gitlab-ci.yml          |  3 +-
 tools/docker/Dockerfile | 83 ++++++++++++++++-------------------------
 3 files changed, 35 insertions(+), 54 deletions(-)

Regards,
Bin

author
#22 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 13 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>
Quoted 8 lines

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Signature · 1 line
Tom
#23 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 27 lines · Tom Rini
Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 21 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 8 lines
Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Just had a quick look. The u-boot.rom (qemu-x86_defconfig) built with GCC 11.1.0 does not boot. It fails at the very early stage at fdtdec_prepare_fdt(). It turns out gd->fdt_blob which is set to _end does not point to a valid DT.

This may have something to do with:

x86_64-linux-ld.bfd: arch/x86/cpu/start.o: warning: relocation in read-only section `.text.start'
x86_64-linux-ld.bfd: warning: creating DT_TEXTREL in a PIE

Regards,
Bin

author
#24 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 42 lines · Bin Meng
Quoted 27 lines · Tom Rini
Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 21 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 8 lines
Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Just had a quick look. The u-boot.rom (qemu-x86_defconfig) built with GCC 11.1.0 does not boot. It fails at the very early stage at fdtdec_prepare_fdt(). It turns out gd->fdt_blob which is set to _end does not point to a valid DT.

Quoted 5 lines

This may have something to do with:

x86_64-linux-ld.bfd: arch/x86/cpu/start.o: warning: relocation in read-only section `.text.start'
x86_64-linux-ld.bfd: warning: creating DT_TEXTREL in a PIE

Yeah, I suspect that's the problem. I just don't know where to start looking for what to change to address that.

Signature · 1 line
Tom
#25 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 44 lines · Tom Rini
Quoted 42 lines · Bin Meng
Quoted 27 lines · Tom Rini
Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 21 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 8 lines
Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Just had a quick look. The u-boot.rom (qemu-x86_defconfig) built with GCC 11.1.0 does not boot. It fails at the very early stage at fdtdec_prepare_fdt(). It turns out gd->fdt_blob which is set to _end does not point to a valid DT.

Quoted 8 lines
Quoted 5 lines

This may have something to do with:

x86_64-linux-ld.bfd: arch/x86/cpu/start.o: warning: relocation in read-only section `.text.start'
x86_64-linux-ld.bfd: warning: creating DT_TEXTREL in a PIE

Yeah, I suspect that's the problem. I just don't know where to start looking for what to change to address that.

The following changes seem to fix the issue:

arch/x86/cpu/u-boot.lds +1 -0
index a283c290ee..22fde01e74 100644
@@ -105,6 +105,7 @@ SECTIONS
        /DISCARD/ : { *(.plt*) }
        /DISCARD/ : { *(.interp*) }
        /DISCARD/ : { *(.gnu*) }
+       /DISCARD/ : { *(.note.gnu.property) }
 
 #ifdef CONFIG_X86_16BIT_INIT
        /*

Not sure if this is something introduced by using --whole-archive?

Regards,
Bin

author
#26 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 69 lines · Bin Meng
Quoted 44 lines · Tom Rini
Quoted 42 lines · Bin Meng
Quoted 27 lines · Tom Rini
Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 21 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 8 lines
Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Just had a quick look. The u-boot.rom (qemu-x86_defconfig) built with GCC 11.1.0 does not boot. It fails at the very early stage at fdtdec_prepare_fdt(). It turns out gd->fdt_blob which is set to _end does not point to a valid DT.

Quoted 8 lines
Quoted 5 lines

This may have something to do with:

x86_64-linux-ld.bfd: arch/x86/cpu/start.o: warning: relocation in read-only section `.text.start'
x86_64-linux-ld.bfd: warning: creating DT_TEXTREL in a PIE

Yeah, I suspect that's the problem. I just don't know where to start looking for what to change to address that.

The following changes seem to fix the issue:

arch/x86/cpu/u-boot.lds +1 -0
index a283c290ee..22fde01e74 100644
@@ -105,6 +105,7 @@ SECTIONS
        /DISCARD/ : { *(.plt*) }
        /DISCARD/ : { *(.interp*) }
        /DISCARD/ : { *(.gnu*) }
+       /DISCARD/ : { *(.note.gnu.property) }
 
 #ifdef CONFIG_X86_16BIT_INIT
        /*
Quoted 1 line

Not sure if this is something introduced by using --whole-archive?

Could be. Thanks for digging in to this part! Do you have any comments / concerns about the dropping grub for riscv32 part of this?

Signature · 1 line
Tom
#27 lore raw

Re: [RFCv3] Azure/GitLab: Move to gcc-11.1.0 and LLVM-11

Quoted 68 lines · Tom Rini
Quoted 66 lines · Bin Meng
Quoted 44 lines · Tom Rini
Quoted 42 lines · Bin Meng
Quoted 27 lines · Tom Rini
Quoted 25 lines · Bin Meng

Hi Tom,

Quoted 21 lines · Tom Rini

- Move to gcc-11.1.0 builds from kernel.org for supported platforms and LLVM-11 for those tests.
- grub-2.06 does not build with gcc-11.1.0 on riscv32 as seen here:
https://www.mail-archive.com/[email protected]/msg30736.html
so drop it.
- Update to newer Xtensa (gcc-9.2.0) and ARC (gcc-10.2) toolchains

Cc Heinrich Schuchardt <[email protected]>
Cc Bin Meng <[email protected]>
Cc Simon Glass <[email protected]>
Cc Rick Chen <[email protected]>

---
Changes in v3:
- Post the right patch this time.
- Switch to gcc-11.1.0 builds rather than pre-release
- grub-2.06 is out, use that rather than -rc1
- Update Xtensa and ARC toolchains.
- Note that right now this causes x86 to fail to link, which I did not see with gcc-11.0 tests.

Quoted 8 lines
Quoted 1 line

Did you mean x86 fails to link with LLVM 11, not GCC 11?

It fails to run for qemu-x86:
https://source.denx.de/u-boot/u-boot/-/jobs/279345#L39
And fails to link on other configs:
https://dev.azure.com/u-boot/u-boot/_build/results?buildId=2378&view=logs&j=8a1d3be7-a4c9-55b6-774d-e7f1a8f80847&t=b2f224a7-1103-5b52-edbc-3784ae727e03

Sorry I forgot to add the CI links.

Just had a quick look. The u-boot.rom (qemu-x86_defconfig) built with GCC 11.1.0 does not boot. It fails at the very early stage at fdtdec_prepare_fdt(). It turns out gd->fdt_blob which is set to _end does not point to a valid DT.

Quoted 8 lines
Quoted 5 lines

This may have something to do with:

x86_64-linux-ld.bfd: arch/x86/cpu/start.o: warning: relocation in read-only section `.text.start'
x86_64-linux-ld.bfd: warning: creating DT_TEXTREL in a PIE

Yeah, I suspect that's the problem. I just don't know where to start looking for what to change to address that.

The following changes seem to fix the issue:

arch/x86/cpu/u-boot.lds +1 -0
index a283c290ee..22fde01e74 100644
@@ -105,6 +105,7 @@ SECTIONS
        /DISCARD/ : { *(.plt*) }
        /DISCARD/ : { *(.interp*) }
        /DISCARD/ : { *(.gnu*) }
+       /DISCARD/ : { *(.note.gnu.property) }
 
Quoted 6 lines
Quoted 4 lines
arch/x86/cpu/u-boot.lds +0 -0
 #ifdef CONFIG_X86_16BIT_INIT
        /*

Not sure if this is something introduced by using --whole-archive?

Could be. Thanks for digging in to this part!

You are welcome :)
I will continue the investigation on this.

Quoted 1 line

Do you have any comments / concerns about the dropping grub for riscv32 part of this?

It looks like the grub community does not want to fix the build issue for riscv32. I am not sure if we can have a solution for grub soon. Heinrich could comment whether it's okay to drop riscv32 EFI testing in U-Boot.

Regards,
Bin