Skip to contentSkip to conversation

Introduce MEMORY uclass and TI GPMC driver

PATCH Superseded Roger Quadros 3 participants 15 messages Oct 6, 2022 – Oct 26, 2022

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover Introduce MEMORY uclass and TI GPMC driver

Hi,

This series introduces the MEMORY controller uclass for the drivers that exist in drivers/memory directory.

With that, we add the TI GPMC Memory controller driver as the first user of this uclass.

The GPMC is a unified memory controller dedicated for interfacing with external memory devices like
- Asynchronous SRAM-like memories and ASICs
- Asynchronous, synchronous, and page mode burst NOR flash
- NAND flash
- Pseudo-SRAM devices

The driver is pulled straight from the Linux kernel and adapted for u-boot.

This driver will take care of setting up the GPMC based on the settings specified in the Device tree and then probe its children.

Roger Quadros (4):
dm: memory: Introduce new uclass
scripts: Makefile.spl: Enable memory drivers to be built for SPL dt/bindings: memory: Add bindings for TI GPMC driver
memory: Add TI GPMC driver

 arch/sandbox/dts/test.dts                     |    4 +
 .../memory/ti,gpmc-child.yaml                 |  252 ++++
 doc/device-tree-bindings/memory/ti,gpmc.yaml  |  190 +++
 drivers/memory/Kconfig                        |   36 +
 drivers/memory/Makefile                       |    3 +
 drivers/memory/memory-sandbox.c               |   18 +
 drivers/memory/memory-uclass.c                |   13 +
 drivers/memory/ti-gpmc.c                      | 1240 +++++++++++++++++
 drivers/memory/ti-gpmc.h                      |  298 ++++
 include/dm/uclass-id.h                        |    1 +
 include/linux/mtd/omap_gpmc.h                 |    3 +
 scripts/Makefile.spl                          |    1 +
 test/dm/Makefile                              |    1 +
 test/dm/memory.c                              |   21 +
 14 files changed, 2081 insertions(+)
 create mode 100644 doc/device-tree-bindings/memory/ti,gpmc-child.yaml
 create mode 100644 doc/device-tree-bindings/memory/ti,gpmc.yaml
 create mode 100644 drivers/memory/memory-sandbox.c
 create mode 100644 drivers/memory/memory-uclass.c
 create mode 100644 drivers/memory/ti-gpmc.c
 create mode 100644 drivers/memory/ti-gpmc.h
 create mode 100644 test/dm/memory.c
Signature · 1 line
2.17.1
#3 lore raw

PATCH v2 2/4 scripts: Makefile.spl: Enable memory drivers to be built for SPL

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +1 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/
Signature · 1 line
2.17.1
#1 lore raw

PATCH v2 1/4 dm: memory: Introduce new uclass

In reply to Roger Quadros

Introduce UCLASS_MEMORY for future Memory Controller device drivers.

---
 arch/sandbox/dts/test.dts       |  4 ++++
 drivers/memory/Kconfig          | 17 +++++++++++++++++
 drivers/memory/Makefile         |  2 ++
 drivers/memory/memory-sandbox.c | 18 ++++++++++++++++++
 drivers/memory/memory-uclass.c  | 13 +++++++++++++
 include/dm/uclass-id.h          |  1 +
 test/dm/Makefile                |  1 +
 test/dm/memory.c                | 21 +++++++++++++++++++++
 8 files changed, 77 insertions(+)
 create mode 100644 drivers/memory/memory-sandbox.c
 create mode 100644 drivers/memory/memory-uclass.c
 create mode 100644 test/dm/memory.c
arch/sandbox/dts/test.dts +4 -0
index 2761588f0d..d65b2d2dcb 100644
@@ -926,6 +926,10 @@
 		};
 	};
 
+	memory-controller {
+		compatible = "sandbox,memory";
+	};
+
 	misc-test {
 		#address-cells = <1>;
 		#size-cells = <1>;
drivers/memory/Kconfig +17 -0
index 7271892763..632feb3aaa 100644
@@ -4,6 +4,23 @@
 
 menu "Memory Controller drivers"
 
+config DM_MEMORY
+	bool "Enable Driver Model for Memory Controller drivers"
+	depends on DM
+	help
+	  Enable driver model for Memory Controller devices.
+	  These devices provide Memory bus interface to various devices like
+	  SRAM, Ethernet adapters, FPGAs, etc.
+	  For now this uclass has no methods yet.
+
+config SANDBOX_MEMORY
+	bool "Enable Sandbox Memory Controller driver"
+	depends on SANDBOX && DM_MEMORY
+	help
+	  This is a driver model based Memory Controller driver for sandbox.
+	  Currently it is a stub only, as there are no usable uclass methods
+	  yet.
+
 config STM32_FMC2_EBI
 	bool "Support for FMC2 External Bus Interface on STM32MP SoCs"
 	depends on ARCH_STM32MP
drivers/memory/Makefile +2 -0
index fec52efb60..1d68b2b296 100644
@@ -1,3 +1,5 @@
 
+obj-$(CONFIG_DM_MEMORY) += memory-uclass.o
+obj-$(CONFIG_SANDBOX_MEMORY) += memory-sandbox.o
 obj-$(CONFIG_STM32_FMC2_EBI) += stm32-fmc2-ebi.o
 obj-$(CONFIG_TI_AEMIF) += ti-aemif.o
drivers/memory/memory-sandbox.c new file +18 -0
new file mode 100644
index 0000000000..f2ede50863
@@ -0,0 +1,18 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2022
+ *     Texas Instruments Incorporated, <www.ti.com>
+ */
+
+#include <dm.h>
+
+static const struct udevice_id sandbox_memory_match[] = {
+	{ .compatible = "sandbox,memory" },
+	{ /* sentinel */ }
+};
+
+U_BOOT_DRIVER(sandbox_memory) = {
+	.name	= "sandbox_memory",
+	.id	= UCLASS_MEMORY,
+	.of_match = sandbox_memory_match,
+};
drivers/memory/memory-uclass.c new file +13 -0
new file mode 100644
index 0000000000..d6d37fe777
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2022
+ *     Texas Instruments Incorporated, <www.ti.com>
+ */
+
+#include <dm.h>
+
+UCLASS_DRIVER(memory) = {
+	.name = "memory",
+	.id = UCLASS_MEMORY,
+	.post_bind = dm_scan_fdt_dev,
+};
include/dm/uclass-id.h +1 -0
index a432e43871..936a16c5d9 100644
@@ -76,6 +76,7 @@ enum uclass_id {
 	UCLASS_MASS_STORAGE,	/* Mass storage device */
 	UCLASS_MDIO,		/* MDIO bus */
 	UCLASS_MDIO_MUX,	/* MDIO MUX/switch */
+	UCLASS_MEMORY,		/* Memory Controller device */
 	UCLASS_MISC,		/* Miscellaneous device */
 	UCLASS_MMC,		/* SD / MMC card or chip */
 	UCLASS_MOD_EXP,		/* RSA Mod Exp device */
test/dm/Makefile +1 -0
index 7543df8823..01275db2cd 100644
@@ -47,6 +47,7 @@ ifneq ($(CONFIG_EFI_PARTITION),)
 obj-$(CONFIG_FASTBOOT_FLASH_MMC) += fastboot.o
 endif
 obj-$(CONFIG_FIRMWARE) += firmware.o
+obj-$(CONFIG_DM_MEMORY) += memory.o
 obj-$(CONFIG_DM_HWSPINLOCK) += hwspinlock.o
 obj-$(CONFIG_DM_I2C) += i2c.o
 obj-$(CONFIG_SOUND) += i2s.o
test/dm/memory.c new file +21 -0
new file mode 100644
index 0000000000..7d9500aa91
@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2022
+ *     Texas Instruments Incorporated, <www.ti.com>
+ */
+
+#include <dm.h>
+#include <dm/test.h>
+#include <test/test.h>
+#include <test/ut.h>
+
+static int dm_test_memory(struct unit_test_state *uts)
+{
+	struct udevice *dev;
+
+	ut_assertok(uclass_first_device_err(UCLASS_MEMORY, &dev));
+
+	return 0;
+}
+
+DM_TEST(dm_test_memory, UT_TESTF_SCAN_FDT);
Signature · 1 line
2.17.1
#13 lore raw

PATCH v2 3/4 dt/bindings: memory: Add bindings for TI GPMC driver

In reply to Roger Quadros

GPMC stands for General Purpose Memory Controller and it is present on many Texas Instruments SoCs.

It supports a number of Asynchronous and Synchronous interfaces and has various settings to configure the bus interface.

The DT bindings define all the various GPMC settings.

As the GPMC supports multiple devices on the bus, each device is represented as a child and the respective GPMC settings are situated there. (see ti,gpmc-child.yaml)

These binding docs are picked up from the Linux kernel.

---
 .../memory/ti,gpmc-child.yaml                 | 252 ++++++++++++++++++
 doc/device-tree-bindings/memory/ti,gpmc.yaml  | 190 +++++++++++++
 2 files changed, 442 insertions(+)
 create mode 100644 doc/device-tree-bindings/memory/ti,gpmc-child.yaml
 create mode 100644 doc/device-tree-bindings/memory/ti,gpmc.yaml
doc/device-tree-bindings/memory/ti,gpmc-child.yaml new file +252 -0
new file mode 100644
index 0000000000..8e541acdb1
@@ -0,0 +1,252 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/memory-controllers/ti,gpmc-child.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: device tree bindings for children of the Texas Instruments GPMC
+
+maintainers:
+  - Tony Lindgren <[email protected]>
+  - Roger Quadros <[email protected]>
+
+description:
+  This binding is meant for the child nodes of the GPMC node. The node
+  represents any device connected to the GPMC bus. It may be a Flash chip,
+  RAM chip or Ethernet controller, etc. These properties are meant for
+  configuring the GPMC settings/timings and will accompany the bindings
+  supported by the respective device.
+
+properties:
+  reg: true
+
+# GPMC Timing properties for child nodes. All are optional and default to 0.
+  gpmc,sync-clk-ps:
+    description: Minimum clock period for synchronous mode
+    default: 0
+
+# Chip-select signal timings corresponding to GPMC_CONFIG2:
+  gpmc,cs-on-ns:
+    description: Assertion time
+    default: 0
+
+  gpmc,cs-rd-off-ns:
+    description: Read deassertion time
+    default: 0
+
+  gpmc,cs-wr-off-ns:
+    description: Write deassertion time
+    default: 0
+
+# ADV signal timings corresponding to GPMC_CONFIG3:
+  gpmc,adv-on-ns:
+    description: Assertion time
+    default: 0
+
+  gpmc,adv-rd-off-ns:
+    description: Read deassertion time
+    default: 0
+
+  gpmc,adv-wr-off-ns:
+    description: Write deassertion time
+    default: 0
+
+  gpmc,adv-aad-mux-on-ns:
+    description: Assertion time for AAD
+    default: 0
+
+  gpmc,adv-aad-mux-rd-off-ns:
+    description: Read deassertion time for AAD
+    default: 0
+
+  gpmc,adv-aad-mux-wr-off-ns:
+    description: Write deassertion time for AAD
+    default: 0
+
+# WE signals timings corresponding to GPMC_CONFIG4:
+  gpmc,we-on-ns:
+    description: Assertion time
+    default: 0
+
+  gpmc,we-off-ns:
+    description: Deassertion time
+    default: 0
+
+# OE signals timings corresponding to GPMC_CONFIG4:
+  gpmc,oe-on-ns:
+    description: Assertion time
+    default: 0
+
+  gpmc,oe-off-ns:
+    description: Deassertion time
+    default: 0
+
+  gpmc,oe-aad-mux-on-ns:
+    description: Assertion time for AAD
+    default: 0
+
+  gpmc,oe-aad-mux-off-ns:
+    description: Deassertion time for AAD
+    default: 0
+
+# Access time and cycle time timings (in nanoseconds) corresponding to
+# GPMC_CONFIG5:
+  gpmc,page-burst-access-ns:
+    description: Multiple access word delay
+    default: 0
+
+  gpmc,access-ns:
+    description: Start-cycle to first data valid delay
+    default: 0
+
+  gpmc,rd-cycle-ns:
+    description: Total read cycle time
+    default: 0
+
+  gpmc,wr-cycle-ns:
+    description: Total write cycle time
+    default: 0
+
+  gpmc,bus-turnaround-ns:
+    description: Turn-around time between successive accesses
+    default: 0
+
+  gpmc,cycle2cycle-delay-ns:
+    description: Delay between chip-select pulses
+    default: 0
+
+  gpmc,clk-activation-ns:
+    description: GPMC clock activation time
+    default: 0
+
+  gpmc,wait-monitoring-ns:
+    description: Start of wait monitoring with regard to valid data
+    default: 0
+
+# Boolean timing parameters. If property is present, parameter is enabled
+# otherwise disabled.
+  gpmc,adv-extra-delay:
+    description: ADV signal is delayed by half GPMC clock
+    type: boolean
+
+  gpmc,cs-extra-delay:
+    description: CS signal is delayed by half GPMC clock
+    type: boolean
+
+  gpmc,cycle2cycle-diffcsen:
+    description: |
+      Add "cycle2cycle-delay" between successive accesses
+      to a different CS
+    type: boolean
+
+  gpmc,cycle2cycle-samecsen:
+    description: |
+      Add "cycle2cycle-delay" between successive accesses
+      to the same CS
+    type: boolean
+
+  gpmc,oe-extra-delay:
+    description: OE signal is delayed by half GPMC clock
+    type: boolean
+
+  gpmc,we-extra-delay:
+    description: WE signal is delayed by half GPMC clock
+    type: boolean
+
+  gpmc,time-para-granularity:
+    description: Multiply all access times by 2
+    type: boolean
+
+# The following two properties are applicable only to OMAP3+ and AM335x:
+  gpmc,wr-access-ns:
+    description: |
+      In synchronous write mode, for single or
+      burst accesses, defines the number of
+      GPMC_FCLK cycles from start access time
+      to the GPMC_CLK rising edge used by the
+      memory device for the first data capture.
+    default: 0
+
+  gpmc,wr-data-mux-bus-ns:
+    description: |
+      In address-data multiplex mode, specifies
+      the time when the first data is driven on
+      the address-data bus.
+    default: 0
+
+# GPMC chip-select settings properties for child nodes. All are optional.
+  gpmc,burst-length:
+    description: Page/burst length.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [0, 4, 8, 16]
+    default: 0
+
+  gpmc,burst-wrap:
+    description: Enables wrap bursting
+    type: boolean
+
+  gpmc,burst-read:
+    description: Enables read page/burst mode
+    type: boolean
+
+  gpmc,burst-write:
+    description: Enables write page/burst mode
+    type: boolean
+
+  gpmc,device-width:
+    description: |
+      Total width of device(s) connected to a GPMC
+      chip-select in bytes. The GPMC supports 8-bit
+      and 16-bit devices and so this property must be
+      1 or 2.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [1, 2]
+    default: 1
+
+  gpmc,mux-add-data:
+    description: |
+      Address and data multiplexing configuration.
+      Valid values are
+      0 for Non multiplexed mode
+      1 for address-address-data multiplexing mode and
+      2 for address-data multiplexing mode.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [0, 1, 2]
+
+  gpmc,sync-read:
+    description: |
+      Enables synchronous read. Defaults to asynchronous
+      is this is not set.
+    type: boolean
+
+  gpmc,sync-write:
+    description: |
+      Enables synchronous writes. Defaults to asynchronous
+      is this is not set.
+    type: boolean
+
+  gpmc,wait-pin:
+    description: |
+      Wait-pin used by client. Must be less than "gpmc,num-waitpins".
+    $ref: /schemas/types.yaml#/definitions/uint32
+
+  gpmc,wait-pin-polarity:
+    description: |
+      Set the desired polarity for the selected wait pin.
+      0 for active low, 1 for active high.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [0, 1]
+
+  gpmc,wait-on-read:
+    description: Enables wait monitoring on reads.
+    type: boolean
+
+  gpmc,wait-on-write:
+    description: Enables wait monitoring on writes.
+    type: boolean
+
+required:
+  - reg
+
+# the GPMC child will have its own native properties
+additionalProperties: true
doc/device-tree-bindings/memory/ti,gpmc.yaml new file +190 -0

190 lines not on this page · raw message

Signature · 1 line
2.17.1
#14 lore raw

PATCH v2 4/4 memory: Add TI GPMC driver

In reply to Roger Quadros

The GPMC is a unified memory controller dedicated for interfacing with external memory devices like
- Asynchronous SRAM-like memories and ASICs
- Asynchronous, synchronous, and page mode burst NOR flash
- NAND flash
- Pseudo-SRAM devices

This driver will take care of setting up the GPMC based on the settings specified in the Device tree and then probe its children.

---
 drivers/memory/Kconfig        |   19 +
 drivers/memory/Makefile       |    1 +
 drivers/memory/ti-gpmc.c      | 1240 +++++++++++++++++++++++++++++++++
 drivers/memory/ti-gpmc.h      |  298 ++++++++
 include/linux/mtd/omap_gpmc.h |    3 +
 5 files changed, 1561 insertions(+)
 create mode 100644 drivers/memory/ti-gpmc.c
 create mode 100644 drivers/memory/ti-gpmc.h
drivers/memory/Kconfig +19 -0
index 632feb3aaa..d057c9883b 100644
@@ -41,4 +41,23 @@ config TI_AEMIF
 	  of 256M bytes of any of these memories can be accessed at a given
 	  time via four chip selects with 64M byte access per chip select.
 
+config TI_GPMC
+	bool "Texas Instruments GPMC driver"
+	depends on ARCH_OMAP2PLUS || ARCH_KEYSTONE || ARCH_K3
+	depends on DM_MEMORY
+	help
+	  This driver is for the General Purpose Memory Controller (GPMC)
+          present on Texas Instruments SoCs (e.g. OMAP2+). GPMC allows
+          interfacing to a variety of asynchronous as well as synchronous
+          memory drives like NOR, NAND, OneNAND, SRAM.
+
+if TI_GPMC
+config TI_GPMC_DEBUG
+	bool "Debug Texas Instruments GPMC timings"
+	default n
+	help
+	  Enable this to print GPMC timings before and after the GPMC registers
+	  are programmed. This should not be left enabled on production systems.
+endif
+
 endmenu
drivers/memory/Makefile +1 -0
index 1d68b2b296..7ce0b98e42 100644
@@ -3,3 +3,4 @@ obj-$(CONFIG_DM_MEMORY) += memory-uclass.o
 obj-$(CONFIG_SANDBOX_MEMORY) += memory-sandbox.o
 obj-$(CONFIG_STM32_FMC2_EBI) += stm32-fmc2-ebi.o
 obj-$(CONFIG_TI_AEMIF) += ti-aemif.o
+obj-$(CONFIG_TI_GPMC) += ti-gpmc.o
drivers/memory/ti-gpmc.c new file +1240 -0

1,240 lines not on this page · raw message

drivers/memory/ti-gpmc.h new file +298 -0

298 lines not on this page · raw message

include/linux/mtd/omap_gpmc.h +3 -0
index 864b05e432..f08e700a1d 100644
@@ -8,6 +8,9 @@
 #ifndef __ASM_OMAP_GPMC_H
 #define __ASM_OMAP_GPMC_H
 
+/* Maximum Number of Chip Selects */
+#define GPMC_CS_NUM	8
+
 #define GPMC_BUF_EMPTY	0
 #define GPMC_BUF_FULL	1
 #define GPMC_MAX_SECTORS	8
Signature · 1 line
2.17.1
Reviewed-by on v2 1/4
#2 lore raw

Re: [u-boot][PATCH v2 1/4] dm: memory: Introduce new uclass

In reply to Roger Quadros

Hi Roger,

Quoted 5 lines · Roger Quadros

Introduce UCLASS_MEMORY for future Memory Controller device drivers.

Quoted 14 lines
---
 arch/sandbox/dts/test.dts       |  4 ++++
 drivers/memory/Kconfig          | 17 +++++++++++++++++
 drivers/memory/Makefile         |  2 ++
 drivers/memory/memory-sandbox.c | 18 ++++++++++++++++++
 drivers/memory/memory-uclass.c  | 13 +++++++++++++
 include/dm/uclass-id.h          |  1 +
 test/dm/Makefile                |  1 +
 test/dm/memory.c                | 21 +++++++++++++++++++++
 8 files changed, 77 insertions(+)
 create mode 100644 drivers/memory/memory-sandbox.c
 create mode 100644 drivers/memory/memory-uclass.c
 create mode 100644 test/dm/memory.c

(with change below)

Quoted 20 lines
arch/sandbox/dts/test.dts +4 -0
index 2761588f0d..d65b2d2dcb 100644
@@ -926,6 +926,10 @@
                };
        };
 
+       memory-controller {
+               compatible = "sandbox,memory";
+       };
+
        misc-test {
                #address-cells = <1>;
                #size-cells = <1>;
drivers/memory/Kconfig +0 -0
index 7271892763..632feb3aaa 100644

drivers/memory/Kconfig line 7

Quoted 4 lines
drivers/memory/Kconfig +1 -0
@@ -4,6 +4,23 @@
 
 menu "Memory Controller drivers"
 
+config DM_MEMORY

Please use MEMORY since there is no non-DM memory option.

Quoted 9 lines
drivers/memory/Kconfig +9 -0
+       bool "Enable Driver Model for Memory Controller drivers"
+       depends on DM
+       help
+         Enable driver model for Memory Controller devices.
+         These devices provide Memory bus interface to various devices like
+         SRAM, Ethernet adapters, FPGAs, etc.
+         For now this uclass has no methods yet.
+
+config SANDBOX_MEMORY

drivers/memory/Kconfig line 26

Quoted 10 lines
drivers/memory/Kconfig +7 -0
+       bool "Enable Sandbox Memory Controller driver"
+       depends on SANDBOX && DM_MEMORY
+       help
+         This is a driver model based Memory Controller driver for sandbox.
+         Currently it is a stub only, as there are no usable uclass methods
+         yet.
+
 config STM32_FMC2_EBI
        bool "Support for FMC2 External Bus Interface on STM32MP SoCs"
        depends on ARCH_STM32MP

[..]

Regards,
Simon

#4 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

In reply to Roger Quadros
Quoted 13 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 7 lines
scripts/Makefile.spl +1 -0
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

Signature · 1 line
Tom
#5 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

That's right. I'll fix it up. Will wait for your comments on the rest of the series before re-spin.

cheers,
-roger

#6 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Quoted 31 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

Quoted 2 lines

That's right. I'll fix it up. Will wait for your comments on the rest of the series before re-spin.

Everything else is fine, I was about to merge it (with a %s/DM_MEMORY/MEMORY) when I saw the stm32 platforms increasing in size and dug in.

Signature · 1 line
Tom
#7 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

So drivers/memory/Makefile remains as it is?

Quoted 8 lines
Quoted 3 lines

That's right. I'll fix it up. Will wait for your comments on the rest of the series before re-spin.

Everything else is fine, I was about to merge it (with a %s/DM_MEMORY/MEMORY) when I saw the stm32 platforms increasing in size and dug in.

cheers,
-roger

#8 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Quoted 35 lines · Roger Quadros

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

Quoted 7 lines

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

So drivers/memory/Makefile remains as it is?

Well, for consistency code should use IS_ENABLED(MEMORY) which will be true for CONFIG_MEMORY or CONFIG_SPL_MEMORY.

Signature · 1 line
Tom
#9 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Hi Tom,

Quoted 39 lines · Tom Rini
Quoted 37 lines · Roger Quadros

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Quoted 8 lines
Quoted 5 lines

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

So drivers/memory/Makefile remains as it is?

Well, for consistency code should use IS_ENABLED(MEMORY) which will be true for CONFIG_MEMORY or CONFIG_SPL_MEMORY.

nit: CONFIG_IS_ENABLED(MEMORY)

Perhaps we can look at my Kconfig series so it can become CONFIG(MEMORY) and we can drop all the SPL_TPL_ stuff in Makefiles?

Regards,
SImon

#10 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Tom,

Quoted 46 lines · Simon Glass

Hi Tom,

Quoted 43 lines · Tom Rini
Quoted 41 lines · Roger Quadros

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

So drivers/memory/Makefile remains as it is?

Well, for consistency code should use IS_ENABLED(MEMORY) which will be true for CONFIG_MEMORY or CONFIG_SPL_MEMORY.

nit: CONFIG_IS_ENABLED(MEMORY)

CONFIG_IS_ENABLED* is used in code but not by any Makefile.
My question was on what basis do we gate drivers/memory inclusion in build from Makefile.

Can CONFIG_MEMORY and CONFIG_SPL_MEMORY co-exist or that should never happen?

Quoted 6 lines

Perhaps we can look at my Kconfig series so it can become CONFIG(MEMORY) and we can drop all the SPL_TPL_ stuff in Makefiles?

Regards,
SImon

cheers,
-roger

#11 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

Quoted 54 lines · Roger Quadros

Tom,

Quoted 51 lines · Simon Glass

Hi Tom,

Quoted 47 lines · Tom Rini
Quoted 35 lines · Roger Quadros

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

Quoted 7 lines

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

So drivers/memory/Makefile remains as it is?

Well, for consistency code should use IS_ENABLED(MEMORY) which will be true for CONFIG_MEMORY or CONFIG_SPL_MEMORY.

Quoted 6 lines
Quoted 1 line

nit: CONFIG_IS_ENABLED(MEMORY)

CONFIG_IS_ENABLED* is used in code but not by any Makefile.
My question was on what basis do we gate drivers/memory inclusion in build from Makefile.

Can CONFIG_MEMORY and CONFIG_SPL_MEMORY co-exist or that should never happen?

What we have for now in your patch is fine and we can tweak slightly if needed if other cases arise.

Signature · 1 line
Tom
#12 lore raw

Re: [u-boot][PATCH v2 2/4] scripts: Makefile.spl: Enable memory drivers to be built for SPL

In reply to Simon Glass
Quoted 48 lines · Simon Glass

Hi Tom,

Quoted 45 lines · Tom Rini
Quoted 42 lines · Roger Quadros

Hi Tom,

Quoted 20 lines · Tom Rini
Quoted 19 lines · Roger Quadros
Quoted 16 lines · Tom Rini
Quoted 15 lines · Roger Quadros

We will need ti-gpmc driver for SPL. Allow memory drivers do be built for SPL.

---
 scripts/Makefile.spl | 1 +
 1 file changed, 1 insertion(+)
scripts/Makefile.spl +0 -0
index 3bafeb4fe9..110076b22f 100644
@@ -114,6 +114,7 @@ libs-$(CONFIG_PARTITIONS) += disk/
 endif
 
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

scripts/Makefile.spl libs-$(CONFIG_PARTITIONS) += disk/ · line 120

Quoted 5 lines
scripts/Makefile.spl +1 -0
 libs-y += drivers/
+libs-y += drivers/memory/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/dwc3/
 libs-$(CONFIG_SPL_USB_GADGET) += drivers/usb/cdns3/
 libs-y += dts/

This ends up being the wrong approach as it then pulls in drivers/memory/stm32-fmc2-ebi.o on all of those platforms, in SPL, which is not what's intended. We need an SPL_MEMORY symbol and then gate the directory on that.

I have a question about how CONFIG_SPL_MEMORY works together with CONFIG_MEMORY.

Do we use CONFIG_SPL_MEMORY only to gate the drivers/memory directory inclusion? Then continue to use CONFIG_MEMORY and others to enable/disable driver build for both non-SPL and SPL case?

So drivers/memory/Makefile remains as it is?

Quoted 7 lines
Quoted 2 lines

Well, for consistency code should use IS_ENABLED(MEMORY) which will be true for CONFIG_MEMORY or CONFIG_SPL_MEMORY.

nit: CONFIG_IS_ENABLED(MEMORY)

Perhaps we can look at my Kconfig series so it can become CONFIG(MEMORY) and we can drop all the SPL_TPL_ stuff in Makefiles?

I think that was the part I had some reservations about, but we should re-visit that idea as I wasn't disagreeing (I believe) with making the logic easier, but I think that we need to re-evaluate how we manage the SPL/TPL/full U-Boot idea.

Signature · 1 line
Tom