[PATCH v4 7/7] soc: renesas: Add L2 cache management for RZ/Five SoC

Heiko Stübner heiko at sntech.de
Thu Nov 24 10:30:11 PST 2022


Am Donnerstag, 24. November 2022, 18:22:07 CET schrieb Prabhakar:
> From: Lad Prabhakar <prabhakar.mahadev-lad.rj at bp.renesas.com>
> 
> On the AX45MP core, cache coherency is a specification option so it may
> not be supported. In this case DMA will fail. As a workaround, firstly we
> allocate a global dma coherent pool from which DMA allocations are taken
> and marked as non-cacheable + bufferable using the PMA region as specified
> in the device tree. Synchronization callbacks are implemented to
> synchronize when doing DMA transactions.
> 
> The Andes AX45MP core has a Programmable Physical Memory Attributes (PMA)
> block that allows dynamic adjustment of memory attributes in the runtime.
> It contains a configurable amount of PMA entries implemented as CSR
> registers to control the attributes of memory locations in interest.
> 
> Below are the memory attributes supported:
> * Device, Non-bufferable
> * Device, bufferable
> * Memory, Non-cacheable, Non-bufferable
> * Memory, Non-cacheable, Bufferable
> * Memory, Write-back, No-allocate
> * Memory, Write-back, Read-allocate
> * Memory, Write-back, Write-allocate
> * Memory, Write-back, Read and Write-allocate
> 
> This patch adds support to configure the memory attributes of the memory
> regions as passed from the l2 cache node and exposes the cache management
> ops.
> 
> More info about PMA (section 10.3):
> Link: http://www.andestech.com/wp-content/uploads/AX45MP-1C-Rev.-5.0.0-Datasheet.pdf
> 
> Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj at bp.renesas.com>
> ---
> RFC v3 -> v4
> * Made use of runtime patching instead of compile time
> * Now just exposing single function ax45mp_no_iocp_cmo() for CMO handling
> * Added a check to make sure cache line size is always 64 bytes
> * Renamed folder rzf -> rzfive
> * Improved Kconfig description
> * Dropped L2 cache configuration
> * Dropped unnecessary casts
> * Fixed comments pointed by Geert, apart from use of PTR_ALIGN_XYZ() macros.
> ---
>  arch/riscv/include/asm/cacheflush.h       |   8 +
>  arch/riscv/include/asm/errata_list.h      |  32 +-
>  drivers/soc/renesas/Kconfig               |   7 +
>  drivers/soc/renesas/Makefile              |   2 +
>  drivers/soc/renesas/rzfive/Kconfig        |   6 +
>  drivers/soc/renesas/rzfive/Makefile       |   3 +
>  drivers/soc/renesas/rzfive/ax45mp_cache.c | 415 ++++++++++++++++++++++
>  drivers/soc/renesas/rzfive/ax45mp_sbi.h   |  29 ++
>  8 files changed, 496 insertions(+), 6 deletions(-)
>  create mode 100644 drivers/soc/renesas/rzfive/Kconfig
>  create mode 100644 drivers/soc/renesas/rzfive/Makefile
>  create mode 100644 drivers/soc/renesas/rzfive/ax45mp_cache.c
>  create mode 100644 drivers/soc/renesas/rzfive/ax45mp_sbi.h
> 
> diff --git a/arch/riscv/include/asm/cacheflush.h b/arch/riscv/include/asm/cacheflush.h
> index 4a04d1be7c67..3226f3aceafe 100644
> --- a/arch/riscv/include/asm/cacheflush.h
> +++ b/arch/riscv/include/asm/cacheflush.h
> @@ -61,6 +61,14 @@ static inline void riscv_noncoherent_supported(void) {}
>  #define SYS_RISCV_FLUSH_ICACHE_LOCAL 1UL
>  #define SYS_RISCV_FLUSH_ICACHE_ALL   (SYS_RISCV_FLUSH_ICACHE_LOCAL)
>  
> +#ifdef CONFIG_AX45MP_L2_CACHE
> +extern asmlinkage void ax45mp_no_iocp_cmo(unsigned int cache_size, void *vaddr,
> +					  size_t size, int dir, int ops);
> +#else
> +inline void ax45mp_no_iocp_cmo(unsigned int cache_size, void *vaddr,
> +			       size_t size, int dir, int ops) {}
> +#endif
> +
>  #include <asm-generic/cacheflush.h>
>  
>  #endif /* _ASM_RISCV_CACHEFLUSH_H */
> diff --git a/arch/riscv/include/asm/errata_list.h b/arch/riscv/include/asm/errata_list.h
> index 48e899a8e7a9..300fed3bfd80 100644
> --- a/arch/riscv/include/asm/errata_list.h
> +++ b/arch/riscv/include/asm/errata_list.h
> @@ -125,8 +125,8 @@ asm volatile(ALTERNATIVE(						\
>  #define THEAD_SYNC_S	".long 0x0190000b"
>  
>  #define ALT_CMO_OP(_op, _start, _size, _cachesize, _dir, _ops)		\
> -asm volatile(ALTERNATIVE_2(						\
> -	__nops(6),							\
> +asm volatile(ALTERNATIVE_3(						\
> +	__nops(14),							\
>  	"mv a0, %1\n\t"							\
>  	"j 2f\n\t"							\
>  	"3:\n\t"							\
> @@ -134,7 +134,7 @@ asm volatile(ALTERNATIVE_2(						\
>  	"add a0, a0, %0\n\t"						\
>  	"2:\n\t"							\
>  	"bltu a0, %2, 3b\n\t"						\
> -	"nop", 0, CPUFEATURE_ZICBOM, CONFIG_RISCV_ISA_ZICBOM,		\
> +	__nops(8), 0, CPUFEATURE_ZICBOM, CONFIG_RISCV_ISA_ZICBOM,	\
>  	"mv a0, %1\n\t"							\
>  	"j 2f\n\t"							\
>  	"3:\n\t"							\
> @@ -142,8 +142,28 @@ asm volatile(ALTERNATIVE_2(						\
>  	"add a0, a0, %0\n\t"						\
>  	"2:\n\t"							\
>  	"bltu a0, %2, 3b\n\t"						\
> -	THEAD_SYNC_S, THEAD_VENDOR_ID,					\
> -			ERRATA_THEAD_CMO, CONFIG_ERRATA_THEAD_CMO)	\
> +	THEAD_SYNC_S "\n\t"						\
> +	__nops(8), THEAD_VENDOR_ID,					\
> +			ERRATA_THEAD_CMO, CONFIG_ERRATA_THEAD_CMO,	\
> +	".option push\n\t\n\t"						\
> +	".option norvc\n\t"						\
> +	".option norelax\n\t">						\

alternatives already do the norvc + norelax options anyway for old and new instructions,
so the .option stuff shouldn't be necessary I guess?


> +	"addi sp,sp,-16\n\t"						\
> +	"sd s0,0(sp)\n\t"						\
> +	"sd ra,8(sp)\n\t"						\
> +	"addi s0,sp,16\n\t"						\
> +	"mv a4,%6\n\t"							\
> +	"mv a3,%5\n\t"							\
> +	"mv a2,%4\n\t"							\
> +	"mv a1,%3\n\t"							\
> +	"mv a0,%0\n\t"							\
> +	"call ax45mp_no_iocp_cmo\n\t"					\
> +	"ld ra,8(sp)\n\t"						\
> +	"ld s0,0(sp)\n\t"						\
> +	"addi sp,sp,16\n\t"						\
> +	".option pop\n\t",						\
> +	ANDESTECH_VENDOR_ID, ERRATA_ANDESTECH_NO_IOCP,			\
> +	CONFIG_ERRATA_ANDES_CMO)					\
>  	: : "r"(_cachesize),						\
>  	    "r"((unsigned long)(_start) & ~((_cachesize) - 1UL)),	\
>  	    "r"((unsigned long)(_start) + (_size)),			\
> @@ -151,7 +171,7 @@ asm volatile(ALTERNATIVE_2(						\
>  	    "r"((unsigned long)(_size)),				\
>  	    "r"((unsigned long)(_dir)),					\
>  	    "r"((unsigned long)(_ops))					\
> -	: "a0")
> +	: "a0", "a1", "a2", "a3", "a4", "memory")
>  
>  #define THEAD_C9XX_RV_IRQ_PMU			17
>  #define THEAD_C9XX_CSR_SCOUNTEROF		0x5c5

[...]

> +static int ax45mp_configure_l2_cache(struct device_node *np)
> +{
> +	int ret;
> +
> +	ret = of_property_read_u32(np, "cache-line-size", &ax45mp_priv->ax45mp_cache_line_size);
> +	if (ret) {
> +		pr_err("Failed to get cache-line-size defaulting to 64 bytes\n");
> +		ax45mp_priv->ax45mp_cache_line_size = SZ_64;
> +	}
> +
> +	if (ax45mp_priv->ax45mp_cache_line_size != SZ_64) {
> +		pr_err("Expected cache-line-size to 64 bytes (found:%u). Defaulting to 64 bytes\n",
> +		       ax45mp_priv->ax45mp_cache_line_size);
> +		ax45mp_priv->ax45mp_cache_line_size = SZ_64;
> +	}
> +
> +	ax45mp_priv->ucctl_ok = ax45mp_cpu_cache_controlable();
> +	ax45mp_priv->l2cache_enabled = ax45mp_cpu_l2c_ctl_status() & AX45MP_L2_CACHE_CTL_CEN_MASK;
> +
> +	return 0;
> +}
> +
> +static int ax45mp_l2c_probe(struct platform_device *pdev)
> +{
> +	struct device_node *np = pdev->dev.of_node;
> +	int ret;
> +
> +	ax45mp_priv = devm_kzalloc(&pdev->dev, sizeof(*ax45mp_priv), GFP_KERNEL);
> +	if (!ax45mp_priv)
> +		return -ENOMEM;
> +
> +	ax45mp_priv->l2c_base = devm_of_iomap(&pdev->dev, pdev->dev.of_node, 0, NULL);
> +	if (!ax45mp_priv->l2c_base) {
> +		ret = -ENOMEM;
> +		goto l2c_err;
> +	}
> +
> +	ret = ax45mp_configure_l2_cache(np);
> +	if (ret)
> +		goto l2c_err;
> +
> +	ret = ax45mp_configure_pma_regions(np);
> +	if (ret)
> +		goto l2c_err;
> +
> +	static_branch_disable(&ax45mp_l2c_configured);
> +
> +	return 0;
> +
> +l2c_err:
> +	devm_kfree(&pdev->dev, ax45mp_priv);
> +	ax45mp_priv = NULL;
> +	return ret;
> +}
> +
> +static const struct of_device_id ax45mp_cache_ids[] = {
> +	{ .compatible = "andestech,ax45mp-cache" },
> +	{ /* sentinel */ }
> +};
> +
> +static struct platform_driver ax45mp_l2c_driver = {
> +	.driver = {
> +		.name = "ax45mp-l2c",
> +		.of_match_table = ax45mp_cache_ids,
> +	},
> +	.probe = ax45mp_l2c_probe,
> +};
> +
> +static int __init ax45mp_cache_init(void)
> +{
> +	static_branch_enable(&ax45mp_l2c_configured);
> +	return platform_driver_register(&ax45mp_l2c_driver);

the ordering is racy I think.

I.e. in the function called from the cmo operations (ax45mp*_range)
you need to access ax45mp_priv and its line-size element.

But when you enable the static branch the driver is not yet registered
but even more important, also not probed yet.

So I guess the static-branch-enable should be living at the end of
ax45mp_l2c_probe()


Heiko





More information about the linux-riscv mailing list