August 2020 - U-Boot

[PATCH 1/7] wdt: dw: Switch to using fls for log2
by Sean Anderson 16 Aug '20

16 Aug '20

log_2_n_round_up is only found in arm. fls performs the same job and is generic. Signed-off-by: Sean Anderson <seanga2(a)gmail.com> --- drivers/watchdog/designware_wdt.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/watchdog/designware_wdt.c b/drivers/watchdog/designware_wdt.c index 12f09a7a39..5684acefd2 100644 --- a/drivers/watchdog/designware_wdt.c +++ b/drivers/watchdog/designware_wdt.c @@ -9,7 +9,6 @@ #include <reset.h> #include <wdt.h> #include <asm/io.h> -#include <asm/utils.h> #include <linux/bitops.h> #define DW_WDT_CR 0x00 @@ -35,7 +34,7 @@ static int designware_wdt_settimeout(void __iomem *base, unsigned int clk_khz, signed int i; /* calculate the timeout range value */ - i = log_2_n_round_up(timeout * clk_khz) - 16; + i = fls(timeout * clk_khz) - 16; i = clamp(i, 0, 15); writel(i | (i << 4), base + DW_WDT_TORR); -- 2.27.0

4 12

pytest to test reset
by Peng Fan 16 Aug '20

16 Aug '20

Hi All, Any ideas how to test uboot reset cmd many times with pytest? Thanks, Peng.

2 1

Re: [PATCH 1/4] firmware: add new driver for SCMI firmwares
by Simon Glass 16 Aug '20

16 Aug '20

Hi Etienne, On Fri, 17 Jul 2020 at 09:38, Etienne Carriere <etienne.carriere(a)linaro.org> wrote: > > This change introduces SCMI agent driver in U-Boot in the firmware > U-class. > > SCMI agent driver is designed for platforms that embed a SCMI server in > a firmware hosted for example by a companion co-processor or the secure > world of the executing processor. > > SCMI protocols allow an SCMI agent to discover and access external > resources as clock, reset controllers and many more. SCMI agent and > server communicate following the SCMI specification [1]. SCMI agent > complies with the DT bindings defined in the Linux kernel source tree > regarding SCMI agent description since v5.8-rc1. > > These bindings describe 2 supported message transport layer: using > mailbox uclass devices or using Arm SMC invocation instruction. Both > use a piece or shared memory for message data exchange. > > In the current state, the SCMI agent driver does not bind to any SCMI > protocol to a U-Boot device driver. Former changes will implement > dedicated driver (i.e. an SCMI clock driver or an SCMI reset controller > driver) and add bind supported SCMI protocols in scmi_agent_bind(). > > Links: [1] https://developer.arm.com/architectures/system-architectures/software-stand… > Signed-off-by: Etienne Carriere <etienne.carriere(a)linaro.org> > --- > > drivers/firmware/Kconfig | 15 ++ > drivers/firmware/Makefile | 1 + > drivers/firmware/scmi.c | 439 ++++++++++++++++++++++++++++++++++++++ > include/scmi.h | 82 +++++++ > 4 files changed, 537 insertions(+) > create mode 100644 drivers/firmware/scmi.c > create mode 100644 include/scmi.h > > diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig > index b70a2063551..f7c7ee7a5aa 100644 > --- a/drivers/firmware/Kconfig > +++ b/drivers/firmware/Kconfig > @@ -1,6 +1,21 @@ > config FIRMWARE > bool "Enable Firmware driver support" > > +config SCMI_FIRMWARE > + bool "Enable SCMI support" > + select FIRMWARE > + select OF_TRANSLATE > + depends on DM_MAILBOX || ARM_SMCCC > + help > + An SCMI agent communicates with a related SCMI server firmware Please write out SCMI in full somewhere and add a link to the spec. > + located in another sub-system, as a companion micro controller > + or a companion host in the CPU system. > + > + Communications between agent (client) and the SCMI server are > + based on message exchange. Messages can be exchange over tranport > + channels as a mailbox device or an Arm SMCCC service with some > + piece of identified shared memory. > + > config SPL_FIRMWARE > bool "Enable Firmware driver support in SPL" > depends on FIRMWARE > diff --git a/drivers/firmware/Makefile b/drivers/firmware/Makefile > index a0c250a473e..3965838179f 100644 > --- a/drivers/firmware/Makefile > +++ b/drivers/firmware/Makefile > @@ -2,4 +2,5 @@ obj-$(CONFIG_FIRMWARE) += firmware-uclass.o > obj-$(CONFIG_$(SPL_)ARM_PSCI_FW) += psci.o > obj-$(CONFIG_TI_SCI_PROTOCOL) += ti_sci.o > obj-$(CONFIG_SANDBOX) += firmware-sandbox.o > +obj-$(CONFIG_SCMI_FIRMWARE) += scmi.o > obj-$(CONFIG_ZYNQMP_FIRMWARE) += firmware-zynqmp.o > diff --git a/drivers/firmware/scmi.c b/drivers/firmware/scmi.c > new file mode 100644 > index 00000000000..fa8a91c3f3d > --- /dev/null > +++ b/drivers/firmware/scmi.c > @@ -0,0 +1,439 @@ > +// SPDX-License-Identifier: GPL-2.0+ > +/* > + * Copyright (c) 2015-2019, Arm Limited and Contributors. All rights reserved. > + * Copyright (C) 2019-2020 Linaro Limited. > + */ > + > +#include <common.h> > +#include <cpu_func.h> > +#include <dm.h> > +#include <errno.h> > +#include <mailbox.h> > +#include <memalign.h> > +#include <scmi.h> > +#include <asm/system.h> > +#include <asm/types.h> > +#include <dm/device-internal.h> > +#include <dm/devres.h> > +#include <dm/lists.h> > +#include <dm/ofnode.h> > +#include <linux/arm-smccc.h> > +#include <linux/compat.h> > +#include <linux/errno.h> > +#include <linux/io.h> > +#include <linux/ioport.h> > + > +#define TIMEOUT_US_10MS 10000 > + > +struct error_code { Function comment > + int scmi; > + int errno; > +}; > + > +static const struct error_code scmi_linux_errmap[] = { > + { .scmi = SCMI_NOT_SUPPORTED, .errno = -EOPNOTSUPP, }, > + { .scmi = SCMI_INVALID_PARAMETERS, .errno = -EINVAL, }, > + { .scmi = SCMI_DENIED, .errno = -EACCES, }, > + { .scmi = SCMI_NOT_FOUND, .errno = -ENOENT, }, > + { .scmi = SCMI_OUT_OF_RANGE, .errno = -ERANGE, }, > + { .scmi = SCMI_BUSY, .errno = -EBUSY, }, > + { .scmi = SCMI_COMMS_ERROR, .errno = -ECOMM, }, > + { .scmi = SCMI_GENERIC_ERROR, .errno = -EIO, }, > + { .scmi = SCMI_HARDWARE_ERROR, .errno = -EREMOTEIO, }, > + { .scmi = SCMI_PROTOCOL_ERROR, .errno = -EPROTO, }, > +}; > + > +int scmi_to_linux_errno(s32 scmi_code) > +{ > + int n; > + > + if (scmi_code == 0) > + return 0; > + > + for (n = 0; n < ARRAY_SIZE(scmi_linux_errmap); n++) > + if (scmi_code == scmi_linux_errmap[n].scmi) > + return scmi_linux_errmap[1].errno; > + > + return -EPROTO; > +} > + > +struct method_ops { > + int (*process_msg)(struct udevice *dev, struct scmi_msg *msg); > + int (*remove_agent)(struct udevice *dev); Looks like this should be a new uclass and child driver? Also we should have a sandbox implementation so this code can be tested. > +}; > + > +struct scmi_agent { > + struct method_ops *method_ops; > + void *method_priv; > +}; > + > +/* > + * Shared Memory based Transport (SMT) message buffer management > + * > + * SMT uses 28 byte header prior message payload to handle the state of > + * the communication channel realized by the shared memory area and > + * to define SCMI protocol information the payload relates to. > + */ > +struct scmi_smt_header { > + __le32 reserved; > + __le32 channel_status; > +#define SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR BIT(1) > +#define SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE BIT(0) > + __le32 reserved1[2]; > + __le32 flags; > +#define SCMI_SHMEM_FLAG_INTR_ENABLED BIT(0) > + __le32 length; > + __le32 msg_header; > + u8 msg_payload[0]; > +}; > + > +#define SMT_HEADER_TOKEN(token) (((token) << 18) & GENMASK(31, 18)) > +#define SMT_HEADER_PROTOCOL_ID(proto) (((proto) << 10) & GENMASK(17, 10)) > +#define SMT_HEADER_MESSAGE_TYPE(type) (((type) << 18) & GENMASK(9, 8)) > +#define SMT_HEADER_MESSAGE_ID(id) ((id) & GENMASK(7, 0)) > + > +struct scmi_shm_buf { comment > + u8 *buf; > + size_t size; > +}; > + > +static int get_shm_buffer(struct udevice *dev, struct scmi_shm_buf *shm) > +{ > + int rc; Can you use 'ret' as 'rc' is a bit short for a var, and we mostly use 'ret' with driver model. > + struct ofnode_phandle_args args; > + struct resource resource; > + fdt32_t faddr; > + phys_addr_t paddr; > + > + rc = dev_read_phandle_with_args(dev, "shmem", NULL, 0, 0, &args); > + if (rc) > + return rc; > + > + rc = ofnode_read_resource(args.node, 0, &resource); > + if (rc) > + return rc; > + > + faddr = cpu_to_fdt32(resource.start); > + paddr = ofnode_translate_address(args.node, &faddr); > + > + shm->size = resource_size(&resource); > + if (shm->size < sizeof(struct scmi_smt_header)) { > + dev_err(dev, "Shared memory buffer too small\n"); > + return -EINVAL; > + } > + > + shm->buf = devm_ioremap(dev, paddr, shm->size); > + if (!shm->buf) > + return -ENOMEM; > + > + if (dcache_status()) > + mmu_set_region_dcache_behaviour((uintptr_t)shm->buf, > + shm->size, DCACHE_OFF); > + > + return 0; > +} > + > +static int write_msg_to_smt(struct udevice *dev, struct scmi_shm_buf *shm_buf, > + struct scmi_msg *msg) function comment with errors > +{ > + struct scmi_smt_header *hdr = (void *)shm_buf->buf; > + > + if ((!msg->in_msg && msg->in_msg_sz) || > + (!msg->out_msg && msg->out_msg_sz)) > + return -EINVAL; > + > + if (!(hdr->channel_status & SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE)) { > + dev_dbg(dev, "Channel busy\n"); > + return -EBUSY; > + } > + > + if (shm_buf->size < (sizeof(*hdr) + msg->in_msg_sz) || > + shm_buf->size < (sizeof(*hdr) + msg->out_msg_sz)) { > + dev_dbg(dev, "Buffer too small\n"); > + return -ETOOSMALL; > + } > + > + /* Load message in shared memory */ > + hdr->channel_status &= ~SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE; > + hdr->length = msg->in_msg_sz + sizeof(hdr->msg_header); > + hdr->msg_header = SMT_HEADER_TOKEN(0) | > + SMT_HEADER_MESSAGE_TYPE(0) | > + SMT_HEADER_PROTOCOL_ID(msg->protocol_id) | > + SMT_HEADER_MESSAGE_ID(msg->message_id); > + > + memcpy(hdr->msg_payload, msg->in_msg, msg->in_msg_sz); > + > + return 0; > +} > + > +static int read_resp_from_smt(struct udevice *dev, struct scmi_shm_buf *shm_buf, > + struct scmi_msg *msg) > +{ > + struct scmi_smt_header *hdr = (void *)shm_buf->buf; > + > + if (!(hdr->channel_status & SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE)) { > + dev_err(dev, "Channel unexpectedly busy\n"); > + return -EBUSY; > + } > + > + if (hdr->channel_status & SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR) { > + dev_err(dev, "Channel error reported, reset channel\n"); > + return -ECOMM; > + } > + > + if (hdr->length > msg->out_msg_sz + sizeof(hdr->msg_header)) { > + dev_err(dev, "Buffer to small\n"); > + return -ETOOSMALL; > + } > + > + /* Get the data */ > + msg->out_msg_sz = hdr->length - sizeof(hdr->msg_header); > + memcpy(msg->out_msg, hdr->msg_payload, msg->out_msg_sz); > + > + return 0; > +} > + > +static void clear_smt_channel(struct scmi_shm_buf *shm_buf) > +{ > + struct scmi_smt_header *hdr = (void *)shm_buf->buf; > + > + hdr->channel_status &= ~SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR; > +} > + > +struct scmi_mbox_channel { please put structs at top of file > + struct scmi_shm_buf shm_buf; > + struct mbox_chan mbox; > + ulong timeout_us; > +}; > + > +static int mbox_process_msg(struct udevice *dev, struct scmi_msg *msg) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + struct scmi_mbox_channel *chan = agent->method_priv; > + int rc; > + > + rc = write_msg_to_smt(dev, &chan->shm_buf, msg); > + if (rc) > + return rc; > + > + /* Give shm addr to mbox in case it is meaningful */ > + rc = mbox_send(&chan->mbox, chan->shm_buf.buf); > + if (rc) { > + dev_err(dev, "Message send failed: %d\n", rc); > + goto out; > + } > + > + /* Receive the response */ > + rc = mbox_recv(&chan->mbox, chan->shm_buf.buf, chan->timeout_us); > + if (rc) { > + dev_err(dev, "Response failed: %d, abort\n", rc); > + goto out; > + } > + > + rc = read_resp_from_smt(dev, &chan->shm_buf, msg); > + > +out: > + clear_smt_channel(&chan->shm_buf); > + > + return rc; > +} > + > +struct method_ops mbox_channel_ops = { > + .process_msg = mbox_process_msg, > +}; > + > +static int probe_mailbox_channel(struct udevice *dev) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + struct scmi_mbox_channel *chan; > + int rc; > + > + chan = devm_kzalloc(dev, sizeof(*chan), GFP_KERNEL); > + if (!chan) > + return -ENOMEM; > + > + chan->timeout_us = TIMEOUT_US_10MS; > + > + rc = mbox_get_by_index(dev, 0, &chan->mbox); > + if (rc) { > + dev_err(dev, "Failed to find mailbox: %d\n", rc); > + goto out; > + } > + > + rc = get_shm_buffer(dev, &chan->shm_buf); > + if (rc) > + dev_err(dev, "Failed to get shm resources: %d\n", rc); > + > +out: > + if (rc) { > + devm_kfree(dev, chan); return rc > + } else { > + agent->method_ops = &mbox_channel_ops; > + agent->method_priv = (void *)chan; > + } > + > + return rc; return 0 (to keep the normal non-error flow as the main one) > +} > + > +struct scmi_arm_smc_channel { > + ulong func_id; > + struct scmi_shm_buf shm_buf; > +}; > + > +#define SMCCC_RET_NOT_SUPPORTED ((unsigned long)-1) Please struct and #define decls to top of file. > + > +static int arm_smc_process_msg(struct udevice *dev, struct scmi_msg *msg) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + struct scmi_arm_smc_channel *chan = agent->method_priv; > + struct arm_smccc_res res; > + int rc; > + > + rc = write_msg_to_smt(dev, &chan->shm_buf, msg); > + if (rc) > + return rc; > + > + arm_smccc_smc(chan->func_id, 0, 0, 0, 0, 0, 0, 0, &res); > + if (res.a0 == SMCCC_RET_NOT_SUPPORTED) > + rc = -EINVAL; ENOTSUP? > + else > + rc = read_resp_from_smt(dev, &chan->shm_buf, msg); > + > + clear_smt_channel(&chan->shm_buf); > + > + return rc; > +} > + > +struct method_ops arm_smc_channel_ops = { > + .process_msg = arm_smc_process_msg, > +}; > + > +static int probe_arm_smc_channel(struct udevice *dev) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + struct scmi_arm_smc_channel *chan; > + ofnode node = dev_ofnode(dev); > + u32 func_id; > + int rc; > + > + chan = devm_kzalloc(dev, sizeof(*chan), GFP_KERNEL); > + if (!chan) > + return -ENOMEM; > + > + if (ofnode_read_u32(node, "arm,smc-id", &func_id)) { dev_read_u32() > + dev_err(dev, "Missing property func-id\n"); > + return -EINVAL; > + } > + > + chan->func_id = func_id; > + > + rc = get_shm_buffer(dev, &chan->shm_buf); > + if (rc) { > + dev_err(dev, "Failed to get shm resources: %d\n", rc); > + return rc; > + } > + > + agent->method_ops = &arm_smc_channel_ops; > + agent->method_priv = (void *)chan; > + > + return rc; > +} > + > +/* > + * Exported functions by the SCMI agent > + */ > + > +int scmi_send_and_process_msg(struct udevice *dev, struct scmi_msg *msg) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + > + return agent->method_ops->process_msg(dev, msg); > +} > + > +static int scmi_remove(struct udevice *dev) > +{ > + struct scmi_agent *agent = dev_get_priv(dev); > + > + if (agent->method_ops->remove_agent) > + return agent->method_ops->remove_agent(dev); method_ops should be the uclass operations. I think you need a new uclass or add something to UCLASS_FIRMWARE. > + > + return 0; > +} > + > +enum scmi_transport_channel { > + SCMI_MAILBOX_TRANSPORT, > + SCMI_ARM_SMCCC_TRANSPORT, > +}; > + > +static int scmi_probe(struct udevice *dev) > +{ > + switch (dev_get_driver_data(dev)) { > + case SCMI_MAILBOX_TRANSPORT: > + if (IS_ENABLED(CONFIG_DM_MAILBOX)) > + return probe_mailbox_channel(dev); > + break; > + case SCMI_ARM_SMCCC_TRANSPORT: > + if (IS_ENABLED(CONFIG_ARM_SMCCC)) > + return probe_arm_smc_channel(dev); > + break; > + default: > + break; > + } > + > + return -EINVAL; > +} > + > +static int scmi_bind(struct udevice *dev) > +{ > + int rc = 0; > + ofnode node; > + struct driver *drv; > + > + dev_for_each_subnode(node, dev) { > + u32 protocol_id; > + > + if (!ofnode_is_available(node)) > + continue; > + > + if (ofnode_read_u32(node, "reg", &protocol_id)) > + continue; > + > + switch (protocol_id) { > + default: > + dev_info(dev, "Ignore unsupported SCMI protocol %u\n", > + protocol_id); > + continue; > + } > + > + rc = device_bind_ofnode(dev, drv, ofnode_get_name(node), > + NULL, node, NULL); Can we instead add a compatible string to the device tree so driver model creates these children automatically? > + if (rc) > + break; > + } > + > + if (rc) > + device_unbind(dev); You should not unbind the device within its bind() method. Just make sure it refuses to probe. > + > + return rc; > +} > + > +static const struct udevice_id scmi_ids[] = { > +#ifdef CONFIG_DM_MAILBOX > + { .compatible = "arm,scmi", .data = SCMI_MAILBOX_TRANSPORT }, > +#endif > +#ifdef CONFIG_ARM_SMCCC > + { .compatible = "arm,scmi-smc", .data = SCMI_ARM_SMCCC_TRANSPORT }, > +#endif Seems like you need a uclass for the transport? > + { } > +}; > + > +U_BOOT_DRIVER(scmi) = { > + .name = "scmi", > + .id = UCLASS_FIRMWARE, > + .of_match = scmi_ids, > + .priv_auto_alloc_size = sizeof(struct scmi_agent), > + .bind = scmi_bind, > + .probe = scmi_probe, > + .remove = scmi_remove, > + .flags = DM_FLAG_OS_PREPARE, > +}; > diff --git a/include/scmi.h b/include/scmi.h > new file mode 100644 > index 00000000000..e12e322991d > --- /dev/null > +++ b/include/scmi.h > @@ -0,0 +1,82 @@ > +/* SPDX-License-Identifier: GPL-2.0+ OR BSD-3-Clause */ > +/* > + * Copyright (c) 2015-2019, Arm Limited and Contributors. All rights reserved. > + * Copyright (C) 2019, Linaro Limited > + */ > +#ifndef SCMI_H > +#define SCMI_H > + > +#include <asm/types.h> > + > +/** > + * An SCMI agent represent on communication path from a device driver to > + * the remote SCMI server which driver sends messages to and receives > + * response messages from. > + */ > +struct scmi_agent; > + > +enum scmi_std_protocol { > + SCMI_PROTOCOL_ID_BASE = 0x10, > + SCMI_PROTOCOL_ID_POWER_DOMAIN = 0x11, > + SCMI_PROTOCOL_ID_SYSTEM = 0x12, > + SCMI_PROTOCOL_ID_PERF = 0x13, > + SCMI_PROTOCOL_ID_CLOCK = 0x14, > + SCMI_PROTOCOL_ID_SENSOR = 0x15, > + SCMI_PROTOCOL_ID_RESET_DOMAIN = 0x16, > +}; > + > +enum scmi_status_code { > + SCMI_SUCCESS = 0, > + SCMI_NOT_SUPPORTED = -1, > + SCMI_INVALID_PARAMETERS = -2, > + SCMI_DENIED = -3, > + SCMI_NOT_FOUND = -4, > + SCMI_OUT_OF_RANGE = -5, > + SCMI_BUSY = -6, > + SCMI_COMMS_ERROR = -7, > + SCMI_GENERIC_ERROR = -8, > + SCMI_HARDWARE_ERROR = -9, > + SCMI_PROTOCOL_ERROR = -10, > +}; > + > +/* > + * struct scmi_msg - Context of a SCMI message sent and the response received > + * > + * @protocol_id: SCMI protocol ID > + * @message_id: SCMI message ID for a defined protocol ID > + * @in_msg: pointer to the message payload sent by the driver > + * @in_msg_sz: byte size of the message payload sent > + * @out_msg: pointer to buffer to store response message payload > + * @out_msg_size: Byte size of the response buffer or payload sz > + */ > +struct scmi_msg { > + unsigned int protocol_id; > + unsigned int message_id; Does this need a particular size, e,g. u32? > + u8 *in_msg; > + size_t in_msg_sz; > + u8 *out_msg; > + size_t out_msg_sz; > +}; > + > +/** > + * scmi_send_and_process_msg() - send and process a SCMI message > + * > + * Send a message to a SCMI server through a target SCMI agent device. > + * Caller sets scmi_msg::out_msg_sz to the output message buffer size. > + * On return, scmi_msg::out_msg_sz stores the response payload size. > + * > + * @dev: SCMI agent device > + * @msg: Message structure reference > + * @return 0 on success, a negative errno otherwise > + */ > +int scmi_send_and_process_msg(struct udevice *dev, struct scmi_msg *msg); > + > +/** > + * scmi_to_linux_errno() - Convert an SCMI error code into a Linux errno code > + * > + * @scmi_errno: SCMI error code value > + * @return 0 for successful status and a negative errno otherwise > + */ > +int scmi_to_linux_errno(s32 scmi_errno); > + > +#endif /* SCMI_H */ > -- > 2.17.1 > Regards, Simon

2 2

[PATCH v2 0/2] reset: Add a managed API
by Pratyush Yadav 16 Aug '20

16 Aug '20

Hi, This is the 4th of a few series that are re-rolls of Jean-Jacques' earlier efforts. The goal is to facilitate porting drivers from the Linux kernel. This particular series is about reset controllers. It adds a managed API, close to that of Linux. The main difference is that bulk and reset_ctl are handled with different functions. Changes in v2: - Fix test build failing because of missing function prototypes and includes. - Drop NULL checks from the reset core. For now let drivers handle them. We can always add the checks back if we think they are a better idea. - Rebase on latest master. Jean-Jacques Hiblot (2): drivers: reset: Add a managed API to get reset controllers from the DT test: reset: Add tests for the managed API arch/sandbox/include/asm/reset.h | 3 + drivers/reset/reset-uclass.c | 118 ++++++++++++++++++++++++- drivers/reset/sandbox-reset-test.c | 51 +++++++++-- drivers/reset/sandbox-reset.c | 19 ++++ include/reset.h | 135 ++++++++++++++++++++++++++++- test/dm/reset.c | 60 +++++++++++++ 6 files changed, 375 insertions(+), 11 deletions(-) -- 2.27.0

2 5

[PATCH v2 0/8] regmap: Add managed API, regmap fields, regmap config
by Pratyush Yadav 16 Aug '20

16 Aug '20

Hi, This series is a re-spin of Jean-Jacques' earlier effort [0], the goal of which was to facilitate porting drivers from the Linux kernel. It adds the managed API, using the same API as Linux. It also adds support for regmap fields. Jean-Jacques' series added support for custom regmap read/write callbacks. The design was argued against by Simon [1]. He argued that using the driver model was a better idea instead of the custom functions. That would mean slightly more memory usage and a more involved change. The main aim of adding the custom functions is to support the Cadence Sierra PHY driver from Linux [2]. The driver's custom functions aren't very complicated, so they can be replaced by some simple formatting options. So, add the struct regmap_config which contains fields to alter the behaviour of the regmap. This includes specifying the size of the read/write operations via 'width', specifying the start and size of range from code instead of device tree via 'r_start' and 'r_size', and specifying a left shift of the register offset before access via 'reg_offset_shift'. The driver can't be ported verbatim now, but this allows the changes to be very isolated and minimal. These config options allow us to avoid converting to driver model until we really need it. The patches are based on [3] which fixes a segmentation fault in sandbox which didn't allow the tests to complete. [0] https://patchwork.ozlabs.org/project/uboot/cover/20191105114700.24989-1-jjh… [1] https://patchwork.ozlabs.org/comment/2426186/ [2] https://elixir.bootlin.com/linux/latest/source/drivers/phy/cadence/phy-cade… [3] https://patchwork.ozlabs.org/project/uboot/patch/20200526120557.26240-1-p.y… Changes in v2: - Add comments explaining the need for regmap_field. - Set regmap->width to a default of REGMAP_SIZE_32 in regmap_alloc() to avoid the checks in regmap_{read,write}(). - Use calloc() instead of using malloc() and memset() to initialize a regmap to zero. - Update comments on regmap_{read,write}() and regmap_raw_{read,write}(). - Drop comments explaining two non-existent fields in struct reg_field. - Add a comment with example explaining REG_FIELD(). - Add Simon's Reviewed-by trailers. Jean-Jacques Hiblot (3): regmap: Add devm_regmap_init() regmap: Add support for regmap fields test: dm: Add tests for regmap managed API and regmap fields Pratyush Yadav (5): regmap: zero out the regmap on allocation regmap: Allow specifying read/write width regmap: Allow left shifting register offset before access regmap: Add regmap_init_mem_range() regmap: Allow devices to specify regmap range start and size in config arch/sandbox/dts/test.dts | 13 +++ drivers/core/regmap.c | 163 +++++++++++++++++++++++++++++- include/regmap.h | 205 +++++++++++++++++++++++++++++++++++--- test/dm/regmap.c | 198 ++++++++++++++++++++++++++++++++++++ 4 files changed, 562 insertions(+), 17 deletions(-) -- 2.27.0

3 11

[PATCH 1/2] cmd: clk: cosmetic: correct code alignment in show_clks
by Patrick Delaunay 16 Aug '20

16 Aug '20

Correct code alignment in show_clks() function. Signed-off-by: Patrick Delaunay <patrick.delaunay(a)st.com> --- cmd/clk.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/cmd/clk.c b/cmd/clk.c index 439736d072..ba4540334a 100644 --- a/cmd/clk.c +++ b/cmd/clk.c @@ -25,24 +25,24 @@ static void show_clks(struct udevice *dev, int depth, int last_flag) if (device_get_uclass_id(dev) == UCLASS_CLK && clkp) { rate = clk_get_rate(clkp); - printf(" %-12u %8d ", rate, clkp->enable_count); - - for (i = depth; i >= 0; i--) { - is_last = (last_flag >> i) & 1; - if (i) { - if (is_last) - printf(" "); - else - printf("| "); - } else { - if (is_last) - printf("`-- "); - else - printf("|-- "); + printf(" %-12u %8d ", rate, clkp->enable_count); + + for (i = depth; i >= 0; i--) { + is_last = (last_flag >> i) & 1; + if (i) { + if (is_last) + printf(" "); + else + printf("| "); + } else { + if (is_last) + printf("`-- "); + else + printf("|-- "); + } } - } - printf("%s\n", dev->name); + printf("%s\n", dev->name); } list_for_each_entry(child, &dev->child_head, sibling_node) { -- 2.17.1

2 3

[RFC PATCH v2 0/3] RFC: tiny-dm: Proposal for using driver model in SPL
by Simon Glass 16 Aug '20

16 Aug '20

This series provides a proposed enhancement to driver model to reduce overhead in SPL. These patches should not be reviewed other than to comment on the approach. The code is all lumped together in a few patches and so cannot be applied as is. For now, the source tree is available at: https://gitlab.denx.de/u-boot/custodians/u-boot-dm/-/tree/dtoc-working Comments welcome! Benefits (good news) -------------------- As an example of the impact of tiny-dm, chromebook_jerry is converted to use it. This shows approximately a 30% reduction in code and data size and a 85% reduction in malloc() space over of-platdata: text data bss dec hex filename 25248 1836 12 27096 69d8 spl/u-boot-spl (original with DT) 19727 3436 12 23175 5a87 spl/u-boot-spl (OF_PLATDATA) 78% 187% 100% 86% as %age of original 13784 1408 12 15204 3b64 spl/u-boot-spl (SPL_TINY) 70% 41% 100% 66% as %age of platdata 55% 77% 100% 56% as %age of original SPL malloc() usage drops from 944 bytes (OF_PLATDATA) to 116 (SPL_TINY). Overall the 'overhead' of tiny-dm is much less than the full driver model. Code size is currently about 600 bytes for these functions on Thumb2: 00000054 T tiny_dev_probe 00000034 T tiny_dev_get_by_drvdata 00000024 T tiny_dev_find 0000001a T tiny_dev_get 0000003c T tinydev_alloc_data 0000002a t tinydev_lookup_data 00000022 T tinydev_ensure_data 00000014 T tinydev_get_data 00000004 T tinydev_get_parent Effort (bad news) ----------------- Unfortunately it is quite a bit of work to convert drivers over to tiny-dm. First, the of-platdata conversion must be done. But on top of that, tiny-dm needs entirely separate code for dealing with devices. This means that instead of 'struct udevice' and 'struct uclass' there is just 'struct tinydev'. Each driver and uclass must be modified to support both, pulling common code into internal static functions. Another option -------------- Note: It is assumed that any board that is space-contrained should use of-platdata in SPL (see doc/driver-model/of-plat.rst). This is shown to reduce device-tree overhead by approximately 4KB. Designing tiny-dm has suggested a number of things that could be changed in the current driver model to make it more space-efficient for TPL and SPL. The ones with least impact on driver code are (CS=reduces code size, DS=reduces data size): CS - drop driver_bind() and create devices (struct udevice) at build-time CS - allocate all device- and uclass-private data at build-time CS - remove all but the basic operations for each uclass (e.g. SPI flash only supports reading) DS - use 8-bit indexes instead of 32/64-bit pointers for device pointers possible since these are created at build-time) DS - use singly-linked lists DS - use 16-bit offsets to private data, instead of 32/64-bit pointers (possible since it is all in SRAM relative to malloc() base, presumably word-aligned and < 256KB) DS - move private pointers into a separate data structure so that NULLs are not stored CS / DS - Combine req_seq and seq and calculate the new value at build-time More difficult are: DS - drop some of the lesser-used driver and uclass methods DS - drop all uclass methods except init() DS - drop all driver methods except probe() CS / DS - drop uclasses and require drivers to manually call uclass functions Even with all of this we would not reach tiny-dm and it would muddy up the driver-model datas structures. But we might come close to tiny-dm on size and there are some advantages: - much of the benefit would apply to all boards that use of-platdata (i.e. with very little effort on behalf of board maintainers) - the impact on the code base is much less (we keep a single, unified driver mode in SPL and U-Boot proper) Overall I think it is worth looking at this option. While it doesn't have the 'nuclear' impact of tiny-dm, neither does it mess with the U-Boot driver code as much and it is easier to learn. Changes in v2: - Various updates, and ported to chromebook_jerry (rockchip) Simon Glass (3): dm: Driver and uclass changes for tiny-dm dm: Arch-specific changes for tiny-dm dm: Core changes for tiny-dm arch/arm/dts/rk3288-u-boot.dtsi | 17 +- arch/arm/dts/rk3288-veyron.dtsi | 26 +- arch/arm/dts/rk3288.dtsi | 3 + arch/arm/include/asm/arch-rockchip/clock.h | 9 + .../include/asm/arch-rockchip/sdram_rk3288.h | 21 ++ arch/arm/include/asm/arch-rockchip/spi.h | 1 + arch/arm/include/asm/io.h | 18 + arch/arm/mach-rockchip/rk3288/clk_rk3288.c | 46 +++ arch/arm/mach-rockchip/rk3288/rk3288.c | 2 + arch/arm/mach-rockchip/rk3288/syscon_rk3288.c | 45 +-- arch/arm/mach-rockchip/sdram.c | 33 +- arch/sandbox/cpu/u-boot-spl.lds | 12 + arch/sandbox/dts/sandbox.dts | 3 +- arch/sandbox/dts/sandbox.dtsi | 2 +- arch/x86/cpu/apollolake/cpu_spl.c | 5 +- arch/x86/cpu/apollolake/uart.c | 56 ++++ arch/x86/dts/chromebook_coral.dts | 1 + arch/x86/lib/tpl.c | 4 +- board/Synology/ds109/ds109.c | 3 +- common/console.c | 2 +- common/log.c | 36 +- common/malloc_simple.c | 31 ++ common/spl/spl.c | 17 +- common/spl/spl_spi.c | 91 +++-- configs/chromebook_coral_defconfig | 1 + configs/chromebook_jerry_defconfig | 11 + configs/rock2_defconfig | 3 + doc/develop/debugging.rst | 35 ++ doc/driver-model/tiny-dm.rst | 315 +++++++++++++++++ drivers/clk/Kconfig | 54 +++ drivers/clk/Makefile | 4 +- drivers/clk/clk-uclass.c | 53 ++- drivers/clk/rockchip/clk_rk3288.c | 106 ++++-- drivers/core/Kconfig | 106 ++++++ drivers/core/Makefile | 3 + drivers/core/of_extra.c | 49 ++- drivers/core/regmap.c | 3 +- drivers/core/syscon-uclass.c | 68 +++- drivers/core/tiny.c | 249 ++++++++++++++ drivers/mtd/spi/Kconfig | 18 + drivers/mtd/spi/sf-uclass.c | 76 +++++ drivers/mtd/spi/sf_probe.c | 4 + drivers/mtd/spi/spi-nor-tiny.c | 166 ++++++++- drivers/ram/Kconfig | 18 + drivers/ram/ram-uclass.c | 12 + drivers/ram/rockchip/sdram_rk3188.c | 2 +- drivers/ram/rockchip/sdram_rk322x.c | 2 +- drivers/ram/rockchip/sdram_rk3288.c | 231 ++++++++----- drivers/ram/rockchip/sdram_rk3328.c | 2 +- drivers/ram/rockchip/sdram_rk3399.c | 2 +- drivers/reset/reset-rockchip.c | 4 +- drivers/serial/Kconfig | 38 +++ drivers/serial/ns16550.c | 195 +++++++++-- drivers/serial/sandbox.c | 59 +++- drivers/serial/serial-uclass.c | 77 +++++ drivers/serial/serial_omap.c | 2 +- drivers/serial/serial_rockchip.c | 59 ++++ drivers/spi/Kconfig | 18 + drivers/spi/Makefile | 2 + drivers/spi/rk_spi.c | 301 ++++++++++++----- drivers/spi/spi-uclass.c | 77 +++++ drivers/sysreset/Kconfig | 18 + drivers/sysreset/sysreset-uclass.c | 124 ++++--- drivers/sysreset/sysreset_rockchip.c | 61 +++- include/asm-generic/global_data.h | 7 +- include/clk-uclass.h | 11 + include/clk.h | 32 +- include/dm/device.h | 121 +++++++ include/dm/of_extra.h | 6 + include/dm/platdata.h | 20 +- include/dm/tiny_struct.h | 42 +++ include/linker_lists.h | 6 + include/linux/mtd/mtd.h | 23 +- include/linux/mtd/spi-nor.h | 22 ++ include/log.h | 6 + include/malloc.h | 3 + include/ns16550.h | 7 +- include/ram.h | 25 ++ include/regmap.h | 4 +- include/serial.h | 45 ++- include/spi.h | 31 ++ include/spi_flash.h | 7 + include/spl.h | 8 +- include/syscon.h | 2 + include/sysreset.h | 9 + scripts/Makefile.spl | 6 +- tools/dtoc/dtb_platdata.py | 316 +++++++++++++++--- tools/dtoc/dtoc_test_simple.dts | 12 +- tools/dtoc/fdt.py | 7 +- tools/dtoc/main.py | 9 +- tools/dtoc/test_dtoc.py | 91 ++++- tools/patman/tools.py | 4 +- 92 files changed, 3454 insertions(+), 540 deletions(-) create mode 100644 doc/develop/debugging.rst create mode 100644 doc/driver-model/tiny-dm.rst create mode 100644 drivers/core/tiny.c create mode 100644 include/dm/tiny_struct.h -- 2.27.0.212.ge8ba1cc988-goog

3 10

[PATCH 1/1] configs: defconfig for Sipeed Maix in S-mode
by Heinrich Schuchardt 15 Aug '20

15 Aug '20

Provide a defconfig that can be used to build U-Boot for the Maix boards running upon OpenSBI. Update the documentation. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- configs/sipeed_maix_smode_defconfig | 10 ++++++ doc/board/sipeed/maix.rst | 49 +++++++++++++++++++++++++++++ 2 files changed, 59 insertions(+) create mode 100644 configs/sipeed_maix_smode_defconfig diff --git a/configs/sipeed_maix_smode_defconfig b/configs/sipeed_maix_smode_defconfig new file mode 100644 index 0000000000..2516bb7258 --- /dev/null +++ b/configs/sipeed_maix_smode_defconfig @@ -0,0 +1,10 @@ +CONFIG_RISCV=y +CONFIG_SYS_TEXT_BASE=0x80020000 +CONFIG_TARGET_SIPEED_MAIX=y +CONFIG_ARCH_RV64I=y +CONFIG_RISCV_SMODE=y +CONFIG_STACK_SIZE=0x100000 +# CONFIG_NET is not set +# CONFIG_INPUT is not set +# CONFIG_DM_ETH is not set +# CONFIG_EFI_UNICODE_CAPITALIZATION is not set diff --git a/doc/board/sipeed/maix.rst b/doc/board/sipeed/maix.rst index b1894f3a6f..efcde9aebf 100644 --- a/doc/board/sipeed/maix.rst +++ b/doc/board/sipeed/maix.rst @@ -75,6 +75,49 @@ console shall be opened immediately. Boot output should look like the following: Err: serial@38000000 => +OpenSBI +^^^^^^^ + +OpenSBI is an open source supervisor execution environment implementing the +RISC-V Supervisor Binary Interface Specification [1]. One of its features is +to intercept run-time exceptions, e.g. for unaligned access or illegal +instructions, and to emulate the failing instructions. + +The OpenSBI source can be downloaded via: + +.. code-block:: bash + + git clone https://github.com/riscv/opensbi + +As OpenSBI will be loaded at 0x80000000 we have to adjust the U-Boot text base. +Furthermore we have to enable building U-Boot for S-mode:: + + CONFIG_SYS_TEXT_BASE=0x80020000 + CONFIG_RISCV_SMODE=y + +Both settings are contained in sipeed_maix_smode_defconfig so we can build +U-Boot with: + +.. code-block:: bash + + make sipeed_maix_smode_defconfig + make + +To build OpenSBI with U-Boot as a payload: + +.. code-block:: bash + + cd opensbi + make \ + PLATFORM=kendryte/k210 \ + FW_PAYLOAD=y \ + FW_PAYLOAD_OFFSET=0x20000 \ + FW_PAYLOAD_PATH=<path to U-Boot>/u-boot-dtb.bin + +The value of FW_PAYLOAD_OFFSET must match CONFIG_SYS_TEXT_BASE - 0x80000000. + +The file to flash is build/platform/kendryte/k210/firmware/fw_payload.bin. + Loading Images ^^^^^^^^^^^^^^ @@ -363,3 +406,9 @@ Address Size Description interrupts) 0x8801f000 0x1000 credits ========== ========= =========== + +Links +----- + +[1] https://github.com/riscv/riscv-sbi-doc + RISC-V Supervisor Binary Interface Specification -- 2.28.0

1 0

[PATCH 1/1] efi_loader: remove empty comment line
by Heinrich Schuchardt 15 Aug '20

15 Aug '20

Remove a line leading to a warning in make htmldocs. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- include/efi_variable.h | 1 - 1 file changed, 1 deletion(-) diff --git a/include/efi_variable.h b/include/efi_variable.h index 60491cb640..4704a3c16e 100644 --- a/include/efi_variable.h +++ b/include/efi_variable.h @@ -274,7 +274,6 @@ efi_get_next_variable_name_mem(efi_uintn_t *variable_name_size, u16 *variable_na * @data: buffer to which the variable value is copied * @timep: authentication time (seconds since start of epoch) * Return: status code - */ efi_status_t __efi_runtime efi_get_variable_mem(u16 *variable_name, const efi_guid_t *vendor, u32 *attributes, -- 2.28.0

1 0

[PATCH 1/1] efi_loader: document parameters of do_bootefi_exec()
by Heinrich Schuchardt 15 Aug '20

15 Aug '20

Add the missing description of the load_options parameter. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- cmd/bootefi.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/cmd/bootefi.c b/cmd/bootefi.c index fbfed54e85..06563d28ca 100644 --- a/cmd/bootefi.c +++ b/cmd/bootefi.c @@ -304,7 +304,11 @@ efi_status_t efi_install_fdt(void *fdt) /** * do_bootefi_exec() - execute EFI binary * + * The image indicated by @handle is started. When it returns the allocated + * memory for the @load_options is freed. + * * @handle: handle of loaded image + * @load_options: load options * Return: status code * * Load the EFI binary into a newly assigned memory unwinding the relocation -- 2.28.0

1 0