| Message ID | 20220601112201.15510-6-tinghan.shen@mediatek.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Add support for MT8195 SCP 2nd core | expand |
Il 01/06/22 13:21, Tinghan Shen ha scritto: > MT8195 SCP is a dual-core processor. The mtk_scp.c driver only controls > SCP core 0. This patch adds a basic driver to control the another core. > > Core 1 and core 0 of the SCP are housed in the same subsys.They see > registers and memory in the same way. > > Core 1 of the SCP features its own set of core configuration registers, > interrupt controller, timers, and DMAs. The rest of the peripherals > in this subsystem are shared by core 0 and core 1. > > As for memory, core 1 has its own cache memory, and the SCP SRAM is shared > by core 0 and core 1. > Hello Tinghan, checking all the patches that are introducing support for the secondary SCP core, it's clear that you're practically reusing *most of* mtk_scp in mtk_scp_dual. I don't think that adding a new configuration option for MTK_SCP_DUALCORE (nor a new file just for that) is a good idea... the code is "short enough" so you should really just add support for multi-core SCP in mtk_scp.c instead. After doing so, I have a hunch that we'll be able to reduce the size of this implementation even more, as I see literally too much common code :-) Cheers, Angelo > Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com> > --- > drivers/remoteproc/Makefile | 1 + > drivers/remoteproc/mtk_scp_dual.c | 97 +++++++++++++++++++++++++++++++ > 2 files changed, 98 insertions(+) > create mode 100644 drivers/remoteproc/mtk_scp_dual.c > > diff --git a/drivers/remoteproc/Makefile b/drivers/remoteproc/Makefile > index 5478c7cb9e07..84cb687d28da 100644 > --- a/drivers/remoteproc/Makefile > +++ b/drivers/remoteproc/Makefile > @@ -15,6 +15,7 @@ obj-$(CONFIG_IMX_REMOTEPROC) += imx_rproc.o > obj-$(CONFIG_IMX_DSP_REMOTEPROC) += imx_dsp_rproc.o > obj-$(CONFIG_INGENIC_VPU_RPROC) += ingenic_rproc.o > obj-$(CONFIG_MTK_SCP) += mtk_scp.o mtk_scp_ipi.o > +obj-$(CONFIG_MTK_SCP_DUALCORE) += mtk_scp_dual.o > obj-$(CONFIG_OMAP_REMOTEPROC) += omap_remoteproc.o > obj-$(CONFIG_WKUP_M3_RPROC) += wkup_m3_rproc.o > obj-$(CONFIG_DA8XX_REMOTEPROC) += da8xx_remoteproc.o > diff --git a/drivers/remoteproc/mtk_scp_dual.c b/drivers/remoteproc/mtk_scp_dual.c > new file mode 100644 > index 000000000000..7bc08d26f208 > --- /dev/null > +++ b/drivers/remoteproc/mtk_scp_dual.c > @@ -0,0 +1,97 @@ > +// SPDX-License-Identifier: GPL-2.0 > +// > +// Copyright (c) 2022 MediaTek Inc. > + > +#include <linux/err.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > +#include <linux/of_address.h> > +#include <linux/of_platform.h> > +#include <linux/platform_device.h> > + > +#include "mtk_common.h" > +#include "remoteproc_internal.h" > + > +static const struct rproc_ops scp_ops; > + > +static int scp_dual_probe(struct platform_device *pdev) > +{ > + struct device *dev = &pdev->dev; > + struct device_node *np = dev->of_node; > + struct mtk_scp *scp; > + struct rproc *rproc; > + const char *fw_name = "scp-dual.img"; > + int ret, i; > + struct resource *res; > + > + ret = rproc_of_parse_firmware(dev, 0, &fw_name); > + if (ret < 0 && ret != -EINVAL) > + return ret; > + > + rproc = devm_rproc_alloc(dev, np->name, &scp_ops, fw_name, sizeof(*scp)); > + if (!rproc) { > + dev_err(dev, "unable to allocate remoteproc\n"); > + return -ENOMEM; > + } > + > + scp = (struct mtk_scp *)rproc->priv; > + scp->rproc = rproc; > + scp->dev = dev; > + platform_set_drvdata(pdev, scp); > + > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram"); > + scp->sram_base = devm_ioremap(dev, res->start, resource_size(res)); > + if (IS_ERR(scp->sram_base)) > + return dev_err_probe(dev, PTR_ERR(scp->sram_base), > + "Failed to parse and map sram memory\n"); > + > + scp->sram_size = resource_size(res); > + scp->sram_phys = res->start; > + > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg"); > + scp->reg_base = devm_ioremap(dev, res->start, resource_size(res)); > + if (IS_ERR(scp->reg_base)) > + return dev_err_probe(dev, PTR_ERR(scp->reg_base), > + "Failed to parse and map cfg memory\n"); > + > + mutex_init(&scp->send_lock); > + for (i = 0; i < SCP_IPI_MAX; i++) > + mutex_init(&scp->ipi_desc[i].lock); > + > + init_waitqueue_head(&scp->run.wq); > + init_waitqueue_head(&scp->ack_wq); > + > + return 0; > +} > + > +static int scp_dual_remove(struct platform_device *pdev) > +{ > + struct mtk_scp *scp = platform_get_drvdata(pdev); > + int i; > + > + for (i = 0; i < SCP_IPI_MAX; i++) > + mutex_destroy(&scp->ipi_desc[i].lock); > + mutex_destroy(&scp->send_lock); > + > + return 0; > +} > + > +static const struct of_device_id mtk_scp_dual_of_match[] = { > + { .compatible = "mediatek,mt8195-scp-dual" }, > + {}, > +}; > +MODULE_DEVICE_TABLE(of, mtk_scp_dual_of_match); > + > +static struct platform_driver mtk_scp_dual_driver = { > + .probe = scp_dual_probe, > + .remove = scp_dual_remove, > + .driver = { > + .name = "mtk-scp-dual", > + .of_match_table = mtk_scp_dual_of_match, > + }, > +}; > + > +module_platform_driver(mtk_scp_dual_driver); > + > +MODULE_LICENSE("GPL"); > +MODULE_DESCRIPTION("MediaTek SCP dualcore control driver");
On Mon, 2022-06-06 at 11:15 +0200, AngeloGioacchino Del Regno wrote: > Il 01/06/22 13:21, Tinghan Shen ha scritto: > > MT8195 SCP is a dual-core processor. The mtk_scp.c driver only controls > > SCP core 0. This patch adds a basic driver to control the another core. > > > > Core 1 and core 0 of the SCP are housed in the same subsys.They see > > registers and memory in the same way. > > > > Core 1 of the SCP features its own set of core configuration registers, > > interrupt controller, timers, and DMAs. The rest of the peripherals > > in this subsystem are shared by core 0 and core 1. > > > > As for memory, core 1 has its own cache memory, and the SCP SRAM is shared > > by core 0 and core 1. > > > > Hello Tinghan, > > checking all the patches that are introducing support for the secondary SCP core, > it's clear that you're practically reusing *most of* mtk_scp in mtk_scp_dual. > > I don't think that adding a new configuration option for MTK_SCP_DUALCORE (nor a > new file just for that) is a good idea... the code is "short enough" so you should > really just add support for multi-core SCP in mtk_scp.c instead. > > After doing so, I have a hunch that we'll be able to reduce the size of this > implementation even more, as I see literally too much common code :-) > Hi Angelo, Thanks for your review. This series has 2 new files, mtk_scp_dual.c and mtk_scp_subdev.c. Is your advice to merge both files into mtk_scp.c, or to merely merge mtk_scp_dual.c to mtk_scp.c? Thanks, TingHan
Il 06/06/22 11:52, Tinghan Shen ha scritto: > On Mon, 2022-06-06 at 11:15 +0200, AngeloGioacchino Del Regno wrote: >> Il 01/06/22 13:21, Tinghan Shen ha scritto: >>> MT8195 SCP is a dual-core processor. The mtk_scp.c driver only controls >>> SCP core 0. This patch adds a basic driver to control the another core. >>> >>> Core 1 and core 0 of the SCP are housed in the same subsys.They see >>> registers and memory in the same way. >>> >>> Core 1 of the SCP features its own set of core configuration registers, >>> interrupt controller, timers, and DMAs. The rest of the peripherals >>> in this subsystem are shared by core 0 and core 1. >>> >>> As for memory, core 1 has its own cache memory, and the SCP SRAM is shared >>> by core 0 and core 1. >>> >> >> Hello Tinghan, >> >> checking all the patches that are introducing support for the secondary SCP core, >> it's clear that you're practically reusing *most of* mtk_scp in mtk_scp_dual. >> >> I don't think that adding a new configuration option for MTK_SCP_DUALCORE (nor a >> new file just for that) is a good idea... the code is "short enough" so you should >> really just add support for multi-core SCP in mtk_scp.c instead. >> >> After doing so, I have a hunch that we'll be able to reduce the size of this >> implementation even more, as I see literally too much common code :-) >> > > Hi Angelo, > > Thanks for your review. > > This series has 2 new files, mtk_scp_dual.c and mtk_scp_subdev.c. > Is your advice to merge both files into mtk_scp.c, > or to merely merge mtk_scp_dual.c to mtk_scp.c? > > Thanks, > TingHan > > > I suggest to merge both into mtk_scp.c and commonize/generalize functions inside of there as much as possible... including the removal of #if IS_ENABLED(...) macro usages, as you can simply check that during runtime by setting a bool variable to true when it's dual-core. Let's do this first step. I'll give you a more exhaustive review on v2, when this main step is done. Cheers, Angelo
On Mon, 2022-06-06 at 12:08 +0200, AngeloGioacchino Del Regno wrote: > Il 06/06/22 11:52, Tinghan Shen ha scritto: > > On Mon, 2022-06-06 at 11:15 +0200, AngeloGioacchino Del Regno wrote: > > > Il 01/06/22 13:21, Tinghan Shen ha scritto: > > > > MT8195 SCP is a dual-core processor. The mtk_scp.c driver only controls > > > > SCP core 0. This patch adds a basic driver to control the another core. > > > > > > > > Core 1 and core 0 of the SCP are housed in the same subsys.They see > > > > registers and memory in the same way. > > > > > > > > Core 1 of the SCP features its own set of core configuration registers, > > > > interrupt controller, timers, and DMAs. The rest of the peripherals > > > > in this subsystem are shared by core 0 and core 1. > > > > > > > > As for memory, core 1 has its own cache memory, and the SCP SRAM is shared > > > > by core 0 and core 1. > > > > > > > > > > Hello Tinghan, > > > > > > checking all the patches that are introducing support for the secondary SCP core, > > > it's clear that you're practically reusing *most of* mtk_scp in mtk_scp_dual. > > > > > > I don't think that adding a new configuration option for MTK_SCP_DUALCORE (nor a > > > new file just for that) is a good idea... the code is "short enough" so you should > > > really just add support for multi-core SCP in mtk_scp.c instead. > > > > > > After doing so, I have a hunch that we'll be able to reduce the size of this > > > implementation even more, as I see literally too much common code :-) > > > > > > > Hi Angelo, > > > > Thanks for your review. > > > > This series has 2 new files, mtk_scp_dual.c and mtk_scp_subdev.c. > > Is your advice to merge both files into mtk_scp.c, > > or to merely merge mtk_scp_dual.c to mtk_scp.c? > > > > Thanks, > > TingHan > > > > > > > > I suggest to merge both into mtk_scp.c and commonize/generalize functions inside > of there as much as possible... including the removal of #if IS_ENABLED(...) > macro usages, as you can simply check that during runtime by setting a bool > variable to true when it's dual-core. > > Let's do this first step. > I'll give you a more exhaustive review on v2, when this main step is done. > > Cheers, > Angelo Hi Angelo, Ok, I'll merge these files and send next version. Thanks, TingHan
diff --git a/drivers/remoteproc/Makefile b/drivers/remoteproc/Makefile index 5478c7cb9e07..84cb687d28da 100644 --- a/drivers/remoteproc/Makefile +++ b/drivers/remoteproc/Makefile @@ -15,6 +15,7 @@ obj-$(CONFIG_IMX_REMOTEPROC) += imx_rproc.o obj-$(CONFIG_IMX_DSP_REMOTEPROC) += imx_dsp_rproc.o obj-$(CONFIG_INGENIC_VPU_RPROC) += ingenic_rproc.o obj-$(CONFIG_MTK_SCP) += mtk_scp.o mtk_scp_ipi.o +obj-$(CONFIG_MTK_SCP_DUALCORE) += mtk_scp_dual.o obj-$(CONFIG_OMAP_REMOTEPROC) += omap_remoteproc.o obj-$(CONFIG_WKUP_M3_RPROC) += wkup_m3_rproc.o obj-$(CONFIG_DA8XX_REMOTEPROC) += da8xx_remoteproc.o diff --git a/drivers/remoteproc/mtk_scp_dual.c b/drivers/remoteproc/mtk_scp_dual.c new file mode 100644 index 000000000000..7bc08d26f208 --- /dev/null +++ b/drivers/remoteproc/mtk_scp_dual.c @@ -0,0 +1,97 @@ +// SPDX-License-Identifier: GPL-2.0 +// +// Copyright (c) 2022 MediaTek Inc. + +#include <linux/err.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_address.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> + +#include "mtk_common.h" +#include "remoteproc_internal.h" + +static const struct rproc_ops scp_ops; + +static int scp_dual_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct device_node *np = dev->of_node; + struct mtk_scp *scp; + struct rproc *rproc; + const char *fw_name = "scp-dual.img"; + int ret, i; + struct resource *res; + + ret = rproc_of_parse_firmware(dev, 0, &fw_name); + if (ret < 0 && ret != -EINVAL) + return ret; + + rproc = devm_rproc_alloc(dev, np->name, &scp_ops, fw_name, sizeof(*scp)); + if (!rproc) { + dev_err(dev, "unable to allocate remoteproc\n"); + return -ENOMEM; + } + + scp = (struct mtk_scp *)rproc->priv; + scp->rproc = rproc; + scp->dev = dev; + platform_set_drvdata(pdev, scp); + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "sram"); + scp->sram_base = devm_ioremap(dev, res->start, resource_size(res)); + if (IS_ERR(scp->sram_base)) + return dev_err_probe(dev, PTR_ERR(scp->sram_base), + "Failed to parse and map sram memory\n"); + + scp->sram_size = resource_size(res); + scp->sram_phys = res->start; + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cfg"); + scp->reg_base = devm_ioremap(dev, res->start, resource_size(res)); + if (IS_ERR(scp->reg_base)) + return dev_err_probe(dev, PTR_ERR(scp->reg_base), + "Failed to parse and map cfg memory\n"); + + mutex_init(&scp->send_lock); + for (i = 0; i < SCP_IPI_MAX; i++) + mutex_init(&scp->ipi_desc[i].lock); + + init_waitqueue_head(&scp->run.wq); + init_waitqueue_head(&scp->ack_wq); + + return 0; +} + +static int scp_dual_remove(struct platform_device *pdev) +{ + struct mtk_scp *scp = platform_get_drvdata(pdev); + int i; + + for (i = 0; i < SCP_IPI_MAX; i++) + mutex_destroy(&scp->ipi_desc[i].lock); + mutex_destroy(&scp->send_lock); + + return 0; +} + +static const struct of_device_id mtk_scp_dual_of_match[] = { + { .compatible = "mediatek,mt8195-scp-dual" }, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_scp_dual_of_match); + +static struct platform_driver mtk_scp_dual_driver = { + .probe = scp_dual_probe, + .remove = scp_dual_remove, + .driver = { + .name = "mtk-scp-dual", + .of_match_table = mtk_scp_dual_of_match, + }, +}; + +module_platform_driver(mtk_scp_dual_driver); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("MediaTek SCP dualcore control driver");
MT8195 SCP is a dual-core processor. The mtk_scp.c driver only controls SCP core 0. This patch adds a basic driver to control the another core. Core 1 and core 0 of the SCP are housed in the same subsys.They see registers and memory in the same way. Core 1 of the SCP features its own set of core configuration registers, interrupt controller, timers, and DMAs. The rest of the peripherals in this subsystem are shared by core 0 and core 1. As for memory, core 1 has its own cache memory, and the SCP SRAM is shared by core 0 and core 1. Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com> --- drivers/remoteproc/Makefile | 1 + drivers/remoteproc/mtk_scp_dual.c | 97 +++++++++++++++++++++++++++++++ 2 files changed, 98 insertions(+) create mode 100644 drivers/remoteproc/mtk_scp_dual.c