| Message ID | 20250303-arm-psci-system_reset2-vendor-reboots-v9-0-b2cf4a20feda@oss.qualcomm.com |
|---|---|
| Headers | show |
| Series | Implement vendor resets for PSCI SYSTEM_RESET2 | expand |
On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: > From: Elliot Berman <elliot.berman@oss.qualcomm.com> > > SoC vendors have different types of resets and are controlled through > various registers. For instance, Qualcomm chipsets can reboot to a > "download mode" that allows a RAM dump to be collected. Another example > is they also support writing a cookie that can be read by bootloader > during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these > vendor reset types to be implemented without requiring drivers for every > register/cookie. > > Add support in PSCI to statically map reboot mode commands from > userspace to a vendor reset and cookie value using the device tree. I have managed to discuss a little bit this patchset over the last few days and I think we have defined a plan going forward. A point that was raised is: https://man7.org/linux/man-pages/man2/reboot.2.html LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to represent ? Is it the mode the system should reboot into OR it is the actual command to be issued (which is what this patchset implements) ? LINUX_REBOOT_CMD_RESTART "..a default restart..." It is unclear what "default" means. We wonder whether the reboot_mode variable was introduced to _define_ that "default". So, in short, my aim is trying to decouple reboot_mode from the LINUX_REBOOT_CMD_RESTART2 *arg command. I believe that adding a sysfs interface to reboot-mode driver infrastructure would be useful, so that the commands would be exposed to userspace and userspace can set the *arg command specifically to issue a given reset/mode. I wonder why this is not already in place for eg syscon-reboot-mode resets, how does user space issue a command in those systems if the available commands aren't exposed to userspace ? Is there a kernel entity exposing those "modes" to userspace, somehow ? > A separate initcall is needed to parse the devicetree, instead of using > psci_dt_init because mm isn't sufficiently set up to allocate memory. > > Reboot mode framework is close but doesn't quite fit with the > design and requirements for PSCI SYSTEM_RESET2. Some of these issues can > be solved but doesn't seem reasonable in sum: > 1. reboot mode registers against the reboot_notifier_list, which is too > early to call SYSTEM_RESET2. PSCI would need to remember the reset > type from the reboot-mode framework callback and use it > psci_sys_reset. > 2. reboot mode assumes only one cookie/parameter is described in the > device tree. SYSTEM_RESET2 uses 2: one for the type and one for > cookie. This can be changed and I think it should, so that the reboot modes are exposed to user space and PSCI can use that. > 3. psci cpuidle driver already registers a driver against the > arm,psci-1.0 compatible. Refactoring would be needed to have both a > cpuidle and reboot-mode driver. > > Signed-off-by: Elliot Berman <elliot.berman@oss.qualcomm.com> > --- > drivers/firmware/psci/psci.c | 105 +++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 105 insertions(+) > > diff --git a/drivers/firmware/psci/psci.c b/drivers/firmware/psci/psci.c > index a1ebbe9b73b136218e9d9f9b8daa7756b3ab2fbe..6f8c47deaec0225f26704e1f3bcad52603127a85 100644 > --- a/drivers/firmware/psci/psci.c > +++ b/drivers/firmware/psci/psci.c > @@ -80,6 +80,14 @@ static u32 psci_cpu_suspend_feature; > static bool psci_system_reset2_supported; > static bool psci_system_off2_hibernate_supported; > > +struct psci_reset_param { > + const char *mode; > + u32 reset_type; > + u32 cookie; > +}; > +static struct psci_reset_param *psci_reset_params __ro_after_init; > +static size_t num_psci_reset_params __ro_after_init; > + > static inline bool psci_has_ext_power_state(void) > { > return psci_cpu_suspend_feature & > @@ -306,9 +314,39 @@ static int get_set_conduit_method(const struct device_node *np) > return 0; > } > > +static int psci_vendor_system_reset2(const char *cmd) > +{ > + unsigned long ret; > + size_t i; > + > + for (i = 0; i < num_psci_reset_params; i++) { > + if (!strcmp(psci_reset_params[i].mode, cmd)) { > + ret = invoke_psci_fn(PSCI_FN_NATIVE(1_1, SYSTEM_RESET2), > + psci_reset_params[i].reset_type, > + psci_reset_params[i].cookie, 0); > + /* > + * if vendor reset fails, log it and fall back to > + * architecture reset types That's not what the code does. > + */ > + pr_err("failed to perform reset \"%s\": %ld\n", cmd, > + (long)ret); > + return 0; > + } > + } > + > + return -ENOENT; > +} > + > static int psci_sys_reset(struct notifier_block *nb, unsigned long action, > void *data) > { > + /* > + * try to do the vendor system_reset2 > + * If there wasn't a matching command, fall back to architectural resets > + */ > + if (data && !psci_vendor_system_reset2(data)) > + return NOTIFY_DONE; > + > if ((reboot_mode == REBOOT_WARM || reboot_mode == REBOOT_SOFT) && > psci_system_reset2_supported) { > /* > @@ -795,6 +833,73 @@ static const struct of_device_id psci_of_match[] __initconst = { > {}, > }; > > +#define REBOOT_PREFIX "mode-" > + > +static int __init psci_init_system_reset2_modes(void) > +{ > + const size_t len = strlen(REBOOT_PREFIX); > + struct psci_reset_param *param; > + struct device_node *psci_np __free(device_node) = NULL; > + struct device_node *np __free(device_node) = NULL; > + struct property *prop; > + size_t count = 0; > + u32 magic[2]; > + int num; > + > + if (!psci_system_reset2_supported) > + return 0; > + > + psci_np = of_find_matching_node(NULL, psci_of_match); > + if (!psci_np) > + return 0; > + > + np = of_find_node_by_name(psci_np, "reset-types"); > + if (!np) > + return 0; Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, is the actual reset to be issued, should we add a default mode "cold" and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? It all boils down to what *arg represents - adding "cold" and "warm" modes would remove the dependency on reboot_mode for resets issued through LINUX_REBOOT_CMD_RESTART2, the question is whether this is the correct thing to do. Comments very welcome. Thanks, Lorenzo > + > + for_each_property_of_node(np, prop) { > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > + continue; > + num = of_property_count_u32_elems(np, prop->name); > + if (num != 1 && num != 2) > + continue; > + > + count++; > + } > + > + param = psci_reset_params = > + kcalloc(count, sizeof(*psci_reset_params), GFP_KERNEL); > + if (!psci_reset_params) > + return -ENOMEM; > + > + for_each_property_of_node(np, prop) { > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > + continue; > + > + num = of_property_read_variable_u32_array(np, prop->name, magic, > + 1, ARRAY_SIZE(magic)); > + if (num < 0) { > + pr_warn("Failed to parse vendor reboot mode %s\n", > + param->mode); > + kfree_const(param->mode); > + continue; > + } > + > + param->mode = kstrdup_const(prop->name + len, GFP_KERNEL); > + if (!param->mode) > + continue; > + > + /* Force reset type to be in vendor space */ > + param->reset_type = PSCI_1_1_RESET_TYPE_VENDOR_START | magic[0]; > + param->cookie = num > 1 ? magic[1] : 0; > + param++; > + num_psci_reset_params++; > + } > + > + return 0; > +} > +arch_initcall(psci_init_system_reset2_modes); > + > int __init psci_dt_init(void) > { > struct device_node *np; > > -- > 2.34.1 >
On Fri, Mar 14, 2025, at 12:19, Lorenzo Pieralisi wrote: > On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: >> From: Elliot Berman <elliot.berman@oss.qualcomm.com> >> >> SoC vendors have different types of resets and are controlled through >> various registers. For instance, Qualcomm chipsets can reboot to a >> "download mode" that allows a RAM dump to be collected. Another example >> is they also support writing a cookie that can be read by bootloader >> during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these >> vendor reset types to be implemented without requiring drivers for every >> register/cookie. >> >> Add support in PSCI to statically map reboot mode commands from >> userspace to a vendor reset and cookie value using the device tree. > > I have managed to discuss a little bit this patchset over the last > few days and I think we have defined a plan going forward. > > A point that was raised is: > > https://man7.org/linux/man-pages/man2/reboot.2.html > > LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to > represent ? > > Is it the mode the system should reboot into OR it is the > actual command to be issued (which is what this patchset > implements) ? > > LINUX_REBOOT_CMD_RESTART "..a default restart..." > > It is unclear what "default" means. We wonder whether the > reboot_mode variable was introduced to _define_ that "default". I think the reboot_mode predates the 'cmd' argument: linux-2.1.30 introduced LINUX_REBOOT_CMD_RESTART2 and it already had the warm/cold/bios/hard options for i386 reboot. I think the argument went unused for a while after it got introduced though. > So, in short, my aim is trying to decouple reboot_mode from the > LINUX_REBOOT_CMD_RESTART2 *arg command. > > I believe that adding a sysfs interface to reboot-mode driver > infrastructure would be useful, so that the commands would > be exposed to userspace and userspace can set the *arg command > specifically to issue a given reset/mode. > > I wonder why this is not already in place for eg syscon-reboot-mode > resets, how does user space issue a command in those systems if the > available commands aren't exposed to userspace ? > > Is there a kernel entity exposing those "modes" to userspace, somehow ? Don't know one either. >> A separate initcall is needed to parse the devicetree, instead of using >> psci_dt_init because mm isn't sufficiently set up to allocate memory. >> >> Reboot mode framework is close but doesn't quite fit with the >> design and requirements for PSCI SYSTEM_RESET2. Some of these issues can >> be solved but doesn't seem reasonable in sum: >> 1. reboot mode registers against the reboot_notifier_list, which is too >> early to call SYSTEM_RESET2. PSCI would need to remember the reset >> type from the reboot-mode framework callback and use it >> psci_sys_reset. >> 2. reboot mode assumes only one cookie/parameter is described in the >> device tree. SYSTEM_RESET2 uses 2: one for the type and one for >> cookie. > > This can be changed and I think it should, so that the reboot modes > are exposed to user space and PSCI can use that. Can we try to call them 'arguments' rather than 'modes' while discussing? I think it's way too easy to confuse them otherwise. >> + psci_np = of_find_matching_node(NULL, psci_of_match); >> + if (!psci_np) >> + return 0; >> + >> + np = of_find_node_by_name(psci_np, "reset-types"); >> + if (!np) >> + return 0; > > Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, > is the actual reset to be issued, should we add a default mode "cold" > and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? > > It all boils down to what *arg represents - adding "cold" and "warm" > modes would remove the dependency on reboot_mode for resets issued > through LINUX_REBOOT_CMD_RESTART2, the question is whether this > is the correct thing to do. > > Comments very welcome. It would make some sense to me to treat all psci reboot as "warm" and not do anything here if reboot="cold" is set: those would have to be backed by a hardware specific driver. A related problem is how to determine when to use UEFI reboot: at the moment arm64 tries the UEFI runtime interface before it even attempts any of the notifiers, so PSCI reboot won't ever be used if UEFI is present. Arnd
On Fri, Mar 14, 2025 at 12:19:31PM +0100, Lorenzo Pieralisi wrote: > On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: > > From: Elliot Berman <elliot.berman@oss.qualcomm.com> > > > > SoC vendors have different types of resets and are controlled through > > various registers. For instance, Qualcomm chipsets can reboot to a > > "download mode" that allows a RAM dump to be collected. Another example > > is they also support writing a cookie that can be read by bootloader > > during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these > > vendor reset types to be implemented without requiring drivers for every > > register/cookie. > > > > Add support in PSCI to statically map reboot mode commands from > > userspace to a vendor reset and cookie value using the device tree. > > I have managed to discuss a little bit this patchset over the last > few days and I think we have defined a plan going forward. > > A point that was raised is: > > https://man7.org/linux/man-pages/man2/reboot.2.html > > LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to > represent ? > > Is it the mode the system should reboot into OR it is the > actual command to be issued (which is what this patchset > implements) ? > > LINUX_REBOOT_CMD_RESTART "..a default restart..." > > It is unclear what "default" means. We wonder whether the > reboot_mode variable was introduced to _define_ that "default". > > So, in short, my aim is trying to decouple reboot_mode from the > LINUX_REBOOT_CMD_RESTART2 *arg command. > > I believe that adding a sysfs interface to reboot-mode driver > infrastructure would be useful, so that the commands would > be exposed to userspace and userspace can set the *arg command > specifically to issue a given reset/mode. > > I wonder why this is not already in place for eg syscon-reboot-mode > resets, how does user space issue a command in those systems if the > available commands aren't exposed to userspace ? > > Is there a kernel entity exposing those "modes" to userspace, somehow ? > > > A separate initcall is needed to parse the devicetree, instead of using > > psci_dt_init because mm isn't sufficiently set up to allocate memory. > > > > Reboot mode framework is close but doesn't quite fit with the > > design and requirements for PSCI SYSTEM_RESET2. Some of these issues can > > be solved but doesn't seem reasonable in sum: > > 1. reboot mode registers against the reboot_notifier_list, which is too > > early to call SYSTEM_RESET2. PSCI would need to remember the reset > > type from the reboot-mode framework callback and use it > > psci_sys_reset. > > 2. reboot mode assumes only one cookie/parameter is described in the > > device tree. SYSTEM_RESET2 uses 2: one for the type and one for > > cookie. > > This can be changed and I think it should, so that the reboot modes > are exposed to user space and PSCI can use that. > In the case of a regular reboot or panic, the reboot/panic notifiers run first, followed by the restart notifiers. The PSCI reset/reset2 should be the last call from Linux, and ideally, this call should not fail. Reboot mode notifiers => restart notifiers or Panic notifiers => restart notifiers So, if I understand correctly, you mean that we can change the reboot mode framework to expose the arguments available to user space. We can extend it to accept magic and cookies, save them in the reboot framework, and retrieve them via a call from PSCI during a regular reboot or panic based on the current arguments. Is this leading towards writing an ARM-specific PSCI-reboot-mode driver, which in its reboot notifier callback saves the magic and cookies, and these magic and cookies will be used during psci_sys_reset2()? Or is there something wrong with my understanding? P.S. We appreciate Elliot for his work and follow-up on this while being employed at Qualcomm. > > 3. psci cpuidle driver already registers a driver against the > > arm,psci-1.0 compatible. Refactoring would be needed to have both a > > cpuidle and reboot-mode driver. > > > > Signed-off-by: Elliot Berman <elliot.berman@oss.qualcomm.com> > > --- > > drivers/firmware/psci/psci.c | 105 +++++++++++++++++++++++++++++++++++++++++++ > > 1 file changed, 105 insertions(+) > > > > diff --git a/drivers/firmware/psci/psci.c b/drivers/firmware/psci/psci.c > > index a1ebbe9b73b136218e9d9f9b8daa7756b3ab2fbe..6f8c47deaec0225f26704e1f3bcad52603127a85 100644 > > --- a/drivers/firmware/psci/psci.c > > +++ b/drivers/firmware/psci/psci.c > > @@ -80,6 +80,14 @@ static u32 psci_cpu_suspend_feature; > > static bool psci_system_reset2_supported; > > static bool psci_system_off2_hibernate_supported; > > > > +struct psci_reset_param { > > + const char *mode; > > + u32 reset_type; > > + u32 cookie; > > +}; > > +static struct psci_reset_param *psci_reset_params __ro_after_init; > > +static size_t num_psci_reset_params __ro_after_init; > > + > > static inline bool psci_has_ext_power_state(void) > > { > > return psci_cpu_suspend_feature & > > @@ -306,9 +314,39 @@ static int get_set_conduit_method(const struct device_node *np) > > return 0; > > } > > > > +static int psci_vendor_system_reset2(const char *cmd) > > +{ > > + unsigned long ret; > > + size_t i; > > + > > + for (i = 0; i < num_psci_reset_params; i++) { > > + if (!strcmp(psci_reset_params[i].mode, cmd)) { > > + ret = invoke_psci_fn(PSCI_FN_NATIVE(1_1, SYSTEM_RESET2), > > + psci_reset_params[i].reset_type, > > + psci_reset_params[i].cookie, 0); > > + /* > > + * if vendor reset fails, log it and fall back to > > + * architecture reset types > > That's not what the code does. > Ack. -Mukesh > > + */ > > + pr_err("failed to perform reset \"%s\": %ld\n", cmd, > > + (long)ret); > > + return 0; > > + } > > + } > > + > > + return -ENOENT; > > +} > > + > > static int psci_sys_reset(struct notifier_block *nb, unsigned long action, > > void *data) > > { > > + /* > > + * try to do the vendor system_reset2 > > + * If there wasn't a matching command, fall back to architectural resets > > + */ > > + if (data && !psci_vendor_system_reset2(data)) > > + return NOTIFY_DONE; > > + > > if ((reboot_mode == REBOOT_WARM || reboot_mode == REBOOT_SOFT) && > > psci_system_reset2_supported) { > > /* > > @@ -795,6 +833,73 @@ static const struct of_device_id psci_of_match[] __initconst = { > > {}, > > }; > > > > +#define REBOOT_PREFIX "mode-" > > + > > +static int __init psci_init_system_reset2_modes(void) > > +{ > > + const size_t len = strlen(REBOOT_PREFIX); > > + struct psci_reset_param *param; > > + struct device_node *psci_np __free(device_node) = NULL; > > + struct device_node *np __free(device_node) = NULL; > > + struct property *prop; > > + size_t count = 0; > > + u32 magic[2]; > > + int num; > > + > > + if (!psci_system_reset2_supported) > > + return 0; > > + > > + psci_np = of_find_matching_node(NULL, psci_of_match); > > + if (!psci_np) > > + return 0; > > + > > + np = of_find_node_by_name(psci_np, "reset-types"); > > + if (!np) > > + return 0; > > Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, > is the actual reset to be issued, should we add a default mode "cold" > and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? > > It all boils down to what *arg represents - adding "cold" and "warm" > modes would remove the dependency on reboot_mode for resets issued > through LINUX_REBOOT_CMD_RESTART2, the question is whether this > is the correct thing to do. > > Comments very welcome. > > Thanks, > Lorenzo > > > + > > + for_each_property_of_node(np, prop) { > > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > > + continue; > > + num = of_property_count_u32_elems(np, prop->name); > > + if (num != 1 && num != 2) > > + continue; > > + > > + count++; > > + } > > + > > + param = psci_reset_params = > > + kcalloc(count, sizeof(*psci_reset_params), GFP_KERNEL); > > + if (!psci_reset_params) > > + return -ENOMEM; > > + > > + for_each_property_of_node(np, prop) { > > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > > + continue; > > + > > + num = of_property_read_variable_u32_array(np, prop->name, magic, > > + 1, ARRAY_SIZE(magic)); > > + if (num < 0) { > > + pr_warn("Failed to parse vendor reboot mode %s\n", > > + param->mode); > > + kfree_const(param->mode); > > + continue; > > + } > > + > > + param->mode = kstrdup_const(prop->name + len, GFP_KERNEL); > > + if (!param->mode) > > + continue; > > + > > + /* Force reset type to be in vendor space */ > > + param->reset_type = PSCI_1_1_RESET_TYPE_VENDOR_START | magic[0]; > > + param->cookie = num > 1 ? magic[1] : 0; > > + param++; > > + num_psci_reset_params++; > > + } > > + > > + return 0; > > +} > > +arch_initcall(psci_init_system_reset2_modes); > > + > > int __init psci_dt_init(void) > > { > > struct device_node *np; > > > > -- > > 2.34.1 > >
On Fri, Mar 14, 2025 at 05:31:44PM +0100, Arnd Bergmann wrote: > On Fri, Mar 14, 2025, at 12:19, Lorenzo Pieralisi wrote: > > On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: > >> From: Elliot Berman <elliot.berman@oss.qualcomm.com> > >> > >> SoC vendors have different types of resets and are controlled through > >> various registers. For instance, Qualcomm chipsets can reboot to a > >> "download mode" that allows a RAM dump to be collected. Another example > >> is they also support writing a cookie that can be read by bootloader > >> during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these > >> vendor reset types to be implemented without requiring drivers for every > >> register/cookie. > >> > >> Add support in PSCI to statically map reboot mode commands from > >> userspace to a vendor reset and cookie value using the device tree. > > > > I have managed to discuss a little bit this patchset over the last > > few days and I think we have defined a plan going forward. > > > > A point that was raised is: > > > > https://man7.org/linux/man-pages/man2/reboot.2.html > > > > LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to > > represent ? > > > > Is it the mode the system should reboot into OR it is the > > actual command to be issued (which is what this patchset > > implements) ? > > > > LINUX_REBOOT_CMD_RESTART "..a default restart..." > > > > It is unclear what "default" means. We wonder whether the > > reboot_mode variable was introduced to _define_ that "default". > > I think the reboot_mode predates the 'cmd' argument: linux-2.1.30 > introduced LINUX_REBOOT_CMD_RESTART2 and it already had > the warm/cold/bios/hard options for i386 reboot. I think the > argument went unused for a while after it got introduced though. > > > So, in short, my aim is trying to decouple reboot_mode from the > > LINUX_REBOOT_CMD_RESTART2 *arg command. > > > > I believe that adding a sysfs interface to reboot-mode driver > > infrastructure would be useful, so that the commands would > > be exposed to userspace and userspace can set the *arg command > > specifically to issue a given reset/mode. > > > > I wonder why this is not already in place for eg syscon-reboot-mode > > resets, how does user space issue a command in those systems if the > > available commands aren't exposed to userspace ? > > > > Is there a kernel entity exposing those "modes" to userspace, somehow ? > > Don't know one either. > > >> A separate initcall is needed to parse the devicetree, instead of using > >> psci_dt_init because mm isn't sufficiently set up to allocate memory. > >> > >> Reboot mode framework is close but doesn't quite fit with the > >> design and requirements for PSCI SYSTEM_RESET2. Some of these issues can > >> be solved but doesn't seem reasonable in sum: > >> 1. reboot mode registers against the reboot_notifier_list, which is too > >> early to call SYSTEM_RESET2. PSCI would need to remember the reset > >> type from the reboot-mode framework callback and use it > >> psci_sys_reset. > >> 2. reboot mode assumes only one cookie/parameter is described in the > >> device tree. SYSTEM_RESET2 uses 2: one for the type and one for > >> cookie. > > > > This can be changed and I think it should, so that the reboot modes > > are exposed to user space and PSCI can use that. > > Can we try to call them 'arguments' rather than 'modes' while discussing? > I think it's way too easy to confuse them otherwise. > > >> + psci_np = of_find_matching_node(NULL, psci_of_match); > >> + if (!psci_np) > >> + return 0; > >> + > >> + np = of_find_node_by_name(psci_np, "reset-types"); > >> + if (!np) > >> + return 0; > > > > Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, > > is the actual reset to be issued, should we add a default mode "cold" > > and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? > > > > It all boils down to what *arg represents - adding "cold" and "warm" > > modes would remove the dependency on reboot_mode for resets issued > > through LINUX_REBOOT_CMD_RESTART2, the question is whether this > > is the correct thing to do. > > > > Comments very welcome. > > It would make some sense to me to treat all psci reboot as "warm" and > not do anything here if reboot="cold" is set: those would have to > be backed by a hardware specific driver. The patch implements PSCI SYSTEM_RESET2 vendor reset types which seems independent of reboot_mode (warm/cold/etc) as per specs. As per ARM PSCI SPEC: https://documentation-service.arm.com/static/60feb9bf3d73a34b640e1b67%3Ftoken%3D&ved=2ahUKEwiX7ramjZCGAxUXrlYBHaPaBX4QFnoECBUQAQ&usg=AOvVaw1L35eQniULNRstfGKX5KXp quoted from spec: -- "" 5.1.11 SYSTEM_RESET2 This function extends SYSTEM_RESET. It provides: • Architectural reset definitions. • Support for vendor-specific resets. Bit[31] Reserved must be zero. o Set to 1 for vendor-specific resets. o Set to 0 for architectural resets. Bits[30:0] o For vendor-specific resets, the format of these bits is IMPLEMENTATION DEFINED. o For architectural resets, the following values are defined: - 0x0 SYSTEM_WARM_RESET. - All other values are reserved. "" So, for vendor specific reset types it does not specify warm or cold. -Mukesh > > A related problem is how to determine when to use UEFI reboot: at the > moment arm64 tries the UEFI runtime interface before it even attempts > any of the notifiers, so PSCI reboot won't ever be used if UEFI is > present. > > Arnd
On Tue, Mar 25, 2025 at 7:33 PM Mukesh Ojha <mukesh.ojha@oss.qualcomm.com> wrote: > > On Fri, Mar 14, 2025 at 12:19:31PM +0100, Lorenzo Pieralisi wrote: > > On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: > > > From: Elliot Berman <elliot.berman@oss.qualcomm.com> > > > > > > SoC vendors have different types of resets and are controlled through > > > various registers. For instance, Qualcomm chipsets can reboot to a > > > "download mode" that allows a RAM dump to be collected. Another example > > > is they also support writing a cookie that can be read by bootloader > > > during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these > > > vendor reset types to be implemented without requiring drivers for every > > > register/cookie. > > > > > > Add support in PSCI to statically map reboot mode commands from > > > userspace to a vendor reset and cookie value using the device tree. > > > > I have managed to discuss a little bit this patchset over the last > > few days and I think we have defined a plan going forward. > > > > A point that was raised is: > > > > https://man7.org/linux/man-pages/man2/reboot.2.html > > > > LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to > > represent ? > > > > Is it the mode the system should reboot into OR it is the > > actual command to be issued (which is what this patchset > > implements) ? > > > > LINUX_REBOOT_CMD_RESTART "..a default restart..." > > > > It is unclear what "default" means. We wonder whether the > > reboot_mode variable was introduced to _define_ that "default". > > > > So, in short, my aim is trying to decouple reboot_mode from the > > LINUX_REBOOT_CMD_RESTART2 *arg command. > > > > I believe that adding a sysfs interface to reboot-mode driver > > infrastructure would be useful, so that the commands would > > be exposed to userspace and userspace can set the *arg command > > specifically to issue a given reset/mode. > > > > I wonder why this is not already in place for eg syscon-reboot-mode > > resets, how does user space issue a command in those systems if the > > available commands aren't exposed to userspace ? > > > > Is there a kernel entity exposing those "modes" to userspace, somehow ? > > > > > A separate initcall is needed to parse the devicetree, instead of using > > > psci_dt_init because mm isn't sufficiently set up to allocate memory. > > > > > > Reboot mode framework is close but doesn't quite fit with the > > > design and requirements for PSCI SYSTEM_RESET2. Some of these issues can > > > be solved but doesn't seem reasonable in sum: > > > 1. reboot mode registers against the reboot_notifier_list, which is too > > > early to call SYSTEM_RESET2. PSCI would need to remember the reset > > > type from the reboot-mode framework callback and use it > > > psci_sys_reset. > > > 2. reboot mode assumes only one cookie/parameter is described in the > > > device tree. SYSTEM_RESET2 uses 2: one for the type and one for > > > cookie. > > > > This can be changed and I think it should, so that the reboot modes > > are exposed to user space and PSCI can use that. > > > In the case of a regular reboot or panic, the reboot/panic notifiers run > first, followed by the restart notifiers. The PSCI reset/reset2 should > be the last call from Linux, and ideally, this call should not fail. > > Reboot mode notifiers => restart notifiers or Panic notifiers => restart > notifiers > > So, if I understand correctly, you mean that we can change the reboot > mode framework to expose the arguments available to user space. We can > extend it to accept magic and cookies, save them in the reboot > framework, and retrieve them via a call from PSCI during a regular > reboot or panic based on the current arguments. Is this leading towards > writing an ARM-specific PSCI-reboot-mode driver, which in its reboot > notifier callback saves the magic and cookies, and these magic and > cookies will be used during psci_sys_reset2()? Or is there something > wrong with my understanding? In case this mail got lost in your inbox, I wanted to float this up again as this is a very important ARM feature and its future course of action depends on the common understanding. -Mukesh
On Tue, Mar 25, 2025 at 07:33:36PM +0530, Mukesh Ojha wrote: > On Fri, Mar 14, 2025 at 12:19:31PM +0100, Lorenzo Pieralisi wrote: > > On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: > > > From: Elliot Berman <elliot.berman@oss.qualcomm.com> > > > > > > SoC vendors have different types of resets and are controlled through > > > various registers. For instance, Qualcomm chipsets can reboot to a > > > "download mode" that allows a RAM dump to be collected. Another example > > > is they also support writing a cookie that can be read by bootloader > > > during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these > > > vendor reset types to be implemented without requiring drivers for every > > > register/cookie. > > > > > > Add support in PSCI to statically map reboot mode commands from > > > userspace to a vendor reset and cookie value using the device tree. > > > > I have managed to discuss a little bit this patchset over the last > > few days and I think we have defined a plan going forward. > > > > A point that was raised is: > > > > https://man7.org/linux/man-pages/man2/reboot.2.html > > > > LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to > > represent ? > > > > Is it the mode the system should reboot into OR it is the > > actual command to be issued (which is what this patchset > > implements) ? > > > > LINUX_REBOOT_CMD_RESTART "..a default restart..." > > > > It is unclear what "default" means. We wonder whether the > > reboot_mode variable was introduced to _define_ that "default". > > > > So, in short, my aim is trying to decouple reboot_mode from the > > LINUX_REBOOT_CMD_RESTART2 *arg command. > > > > I believe that adding a sysfs interface to reboot-mode driver > > infrastructure would be useful, so that the commands would > > be exposed to userspace and userspace can set the *arg command > > specifically to issue a given reset/mode. > > > > I wonder why this is not already in place for eg syscon-reboot-mode > > resets, how does user space issue a command in those systems if the > > available commands aren't exposed to userspace ? > > > > Is there a kernel entity exposing those "modes" to userspace, somehow ? > > > > > A separate initcall is needed to parse the devicetree, instead of using > > > psci_dt_init because mm isn't sufficiently set up to allocate memory. > > > > > > Reboot mode framework is close but doesn't quite fit with the > > > design and requirements for PSCI SYSTEM_RESET2. Some of these issues can > > > be solved but doesn't seem reasonable in sum: > > > 1. reboot mode registers against the reboot_notifier_list, which is too > > > early to call SYSTEM_RESET2. PSCI would need to remember the reset > > > type from the reboot-mode framework callback and use it > > > psci_sys_reset. > > > 2. reboot mode assumes only one cookie/parameter is described in the > > > device tree. SYSTEM_RESET2 uses 2: one for the type and one for > > > cookie. > > > > This can be changed and I think it should, so that the reboot modes > > are exposed to user space and PSCI can use that. > > > In the case of a regular reboot or panic, the reboot/panic notifiers run > first, followed by the restart notifiers. The PSCI reset/reset2 should > be the last call from Linux, and ideally, this call should not fail. > > Reboot mode notifiers => restart notifiers or Panic notifiers => restart > notifiers > > So, if I understand correctly, you mean that we can change the reboot > mode framework to expose the arguments available to user space. We can > extend it to accept magic and cookies, save them in the reboot > framework, and retrieve them via a call from PSCI during a regular > reboot or panic based on the current arguments. Is this leading towards > writing an ARM-specific PSCI-reboot-mode driver, which in its reboot > notifier callback saves the magic and cookies, and these magic and > cookies will be used during psci_sys_reset2()? Or is there something > wrong with my understanding? No, you got it right (apologies for the delay in replying) - if the case for making reboot mode available to user space is accepted. > P.S. We appreciate Elliot for his work and follow-up on this while being > employed at Qualcomm. Yes I sincerely do for his patience, thank you. Lorenzo > > > 3. psci cpuidle driver already registers a driver against the > > > arm,psci-1.0 compatible. Refactoring would be needed to have both a > > > cpuidle and reboot-mode driver. > > > > > > Signed-off-by: Elliot Berman <elliot.berman@oss.qualcomm.com> > > > --- > > > drivers/firmware/psci/psci.c | 105 +++++++++++++++++++++++++++++++++++++++++++ > > > 1 file changed, 105 insertions(+) > > > > > > diff --git a/drivers/firmware/psci/psci.c b/drivers/firmware/psci/psci.c > > > index a1ebbe9b73b136218e9d9f9b8daa7756b3ab2fbe..6f8c47deaec0225f26704e1f3bcad52603127a85 100644 > > > --- a/drivers/firmware/psci/psci.c > > > +++ b/drivers/firmware/psci/psci.c > > > @@ -80,6 +80,14 @@ static u32 psci_cpu_suspend_feature; > > > static bool psci_system_reset2_supported; > > > static bool psci_system_off2_hibernate_supported; > > > > > > +struct psci_reset_param { > > > + const char *mode; > > > + u32 reset_type; > > > + u32 cookie; > > > +}; > > > +static struct psci_reset_param *psci_reset_params __ro_after_init; > > > +static size_t num_psci_reset_params __ro_after_init; > > > + > > > static inline bool psci_has_ext_power_state(void) > > > { > > > return psci_cpu_suspend_feature & > > > @@ -306,9 +314,39 @@ static int get_set_conduit_method(const struct device_node *np) > > > return 0; > > > } > > > > > > +static int psci_vendor_system_reset2(const char *cmd) > > > +{ > > > + unsigned long ret; > > > + size_t i; > > > + > > > + for (i = 0; i < num_psci_reset_params; i++) { > > > + if (!strcmp(psci_reset_params[i].mode, cmd)) { > > > + ret = invoke_psci_fn(PSCI_FN_NATIVE(1_1, SYSTEM_RESET2), > > > + psci_reset_params[i].reset_type, > > > + psci_reset_params[i].cookie, 0); > > > + /* > > > + * if vendor reset fails, log it and fall back to > > > + * architecture reset types > > > > That's not what the code does. > > > Ack. > > -Mukesh > > > > + */ > > > + pr_err("failed to perform reset \"%s\": %ld\n", cmd, > > > + (long)ret); > > > + return 0; > > > + } > > > + } > > > + > > > + return -ENOENT; > > > +} > > > + > > > static int psci_sys_reset(struct notifier_block *nb, unsigned long action, > > > void *data) > > > { > > > + /* > > > + * try to do the vendor system_reset2 > > > + * If there wasn't a matching command, fall back to architectural resets > > > + */ > > > + if (data && !psci_vendor_system_reset2(data)) > > > + return NOTIFY_DONE; > > > + > > > if ((reboot_mode == REBOOT_WARM || reboot_mode == REBOOT_SOFT) && > > > psci_system_reset2_supported) { > > > /* > > > @@ -795,6 +833,73 @@ static const struct of_device_id psci_of_match[] __initconst = { > > > {}, > > > }; > > > > > > +#define REBOOT_PREFIX "mode-" > > > + > > > +static int __init psci_init_system_reset2_modes(void) > > > +{ > > > + const size_t len = strlen(REBOOT_PREFIX); > > > + struct psci_reset_param *param; > > > + struct device_node *psci_np __free(device_node) = NULL; > > > + struct device_node *np __free(device_node) = NULL; > > > + struct property *prop; > > > + size_t count = 0; > > > + u32 magic[2]; > > > + int num; > > > + > > > + if (!psci_system_reset2_supported) > > > + return 0; > > > + > > > + psci_np = of_find_matching_node(NULL, psci_of_match); > > > + if (!psci_np) > > > + return 0; > > > + > > > + np = of_find_node_by_name(psci_np, "reset-types"); > > > + if (!np) > > > + return 0; > > > > Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, > > is the actual reset to be issued, should we add a default mode "cold" > > and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? > > > > It all boils down to what *arg represents - adding "cold" and "warm" > > modes would remove the dependency on reboot_mode for resets issued > > through LINUX_REBOOT_CMD_RESTART2, the question is whether this > > is the correct thing to do. > > > > Comments very welcome. > > > > Thanks, > > Lorenzo > > > > > + > > > + for_each_property_of_node(np, prop) { > > > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > > > + continue; > > > + num = of_property_count_u32_elems(np, prop->name); > > > + if (num != 1 && num != 2) > > > + continue; > > > + > > > + count++; > > > + } > > > + > > > + param = psci_reset_params = > > > + kcalloc(count, sizeof(*psci_reset_params), GFP_KERNEL); > > > + if (!psci_reset_params) > > > + return -ENOMEM; > > > + > > > + for_each_property_of_node(np, prop) { > > > + if (strncmp(prop->name, REBOOT_PREFIX, len)) > > > + continue; > > > + > > > + num = of_property_read_variable_u32_array(np, prop->name, magic, > > > + 1, ARRAY_SIZE(magic)); > > > + if (num < 0) { > > > + pr_warn("Failed to parse vendor reboot mode %s\n", > > > + param->mode); > > > + kfree_const(param->mode); > > > + continue; > > > + } > > > + > > > + param->mode = kstrdup_const(prop->name + len, GFP_KERNEL); > > > + if (!param->mode) > > > + continue; > > > + > > > + /* Force reset type to be in vendor space */ > > > + param->reset_type = PSCI_1_1_RESET_TYPE_VENDOR_START | magic[0]; > > > + param->cookie = num > 1 ? magic[1] : 0; > > > + param++; > > > + num_psci_reset_params++; > > > + } > > > + > > > + return 0; > > > +} > > > +arch_initcall(psci_init_system_reset2_modes); > > > + > > > int __init psci_dt_init(void) > > > { > > > struct device_node *np; > > > > > > -- > > > 2.34.1 > > >
On 4/8/2025 8:46 PM, Lorenzo Pieralisi wrote: > On Tue, Mar 25, 2025 at 07:33:36PM +0530, Mukesh Ojha wrote: >> On Fri, Mar 14, 2025 at 12:19:31PM +0100, Lorenzo Pieralisi wrote: >>> On Mon, Mar 03, 2025 at 01:08:31PM -0800, Elliot Berman wrote: >>>> From: Elliot Berman <elliot.berman@oss.qualcomm.com> >>>> >>>> SoC vendors have different types of resets and are controlled through >>>> various registers. For instance, Qualcomm chipsets can reboot to a >>>> "download mode" that allows a RAM dump to be collected. Another example >>>> is they also support writing a cookie that can be read by bootloader >>>> during next boot. PSCI offers a mechanism, SYSTEM_RESET2, for these >>>> vendor reset types to be implemented without requiring drivers for every >>>> register/cookie. >>>> >>>> Add support in PSCI to statically map reboot mode commands from >>>> userspace to a vendor reset and cookie value using the device tree. >>> >>> I have managed to discuss a little bit this patchset over the last >>> few days and I think we have defined a plan going forward. >>> >>> A point that was raised is: >>> >>> https://man7.org/linux/man-pages/man2/reboot.2.html >>> >>> LINUX_REBOOT_CMD_RESTART2 *arg command, what is it supposed to >>> represent ? >>> >>> Is it the mode the system should reboot into OR it is the >>> actual command to be issued (which is what this patchset >>> implements) ? >>> >>> LINUX_REBOOT_CMD_RESTART "..a default restart..." >>> >>> It is unclear what "default" means. We wonder whether the >>> reboot_mode variable was introduced to _define_ that "default". >>> >>> So, in short, my aim is trying to decouple reboot_mode from the >>> LINUX_REBOOT_CMD_RESTART2 *arg command. >>> >>> I believe that adding a sysfs interface to reboot-mode driver >>> infrastructure would be useful, so that the commands would >>> be exposed to userspace and userspace can set the *arg command >>> specifically to issue a given reset/mode. >>> >>> I wonder why this is not already in place for eg syscon-reboot-mode >>> resets, how does user space issue a command in those systems if the >>> available commands aren't exposed to userspace ? >>> >>> Is there a kernel entity exposing those "modes" to userspace, somehow ? >>> >>>> A separate initcall is needed to parse the devicetree, instead of using >>>> psci_dt_init because mm isn't sufficiently set up to allocate memory. >>>> >>>> Reboot mode framework is close but doesn't quite fit with the >>>> design and requirements for PSCI SYSTEM_RESET2. Some of these issues can >>>> be solved but doesn't seem reasonable in sum: >>>> 1. reboot mode registers against the reboot_notifier_list, which is too >>>> early to call SYSTEM_RESET2. PSCI would need to remember the reset >>>> type from the reboot-mode framework callback and use it >>>> psci_sys_reset. >>>> 2. reboot mode assumes only one cookie/parameter is described in the >>>> device tree. SYSTEM_RESET2 uses 2: one for the type and one for >>>> cookie. >>> >>> This can be changed and I think it should, so that the reboot modes >>> are exposed to user space and PSCI can use that. >>> >> In the case of a regular reboot or panic, the reboot/panic notifiers run >> first, followed by the restart notifiers. The PSCI reset/reset2 should >> be the last call from Linux, and ideally, this call should not fail. >> >> Reboot mode notifiers => restart notifiers or Panic notifiers => restart >> notifiers >> >> So, if I understand correctly, you mean that we can change the reboot >> mode framework to expose the arguments available to user space. We can >> extend it to accept magic and cookies, save them in the reboot >> framework, and retrieve them via a call from PSCI during a regular >> reboot or panic based on the current arguments. Is this leading towards >> writing an ARM-specific PSCI-reboot-mode driver, which in its reboot >> notifier callback saves the magic and cookies, and these magic and >> cookies will be used during psci_sys_reset2()? Or is there something >> wrong with my understanding? > > No, you got it right (apologies for the delay in replying) - if the > case for making reboot mode available to user space is accepted. > Agree that the available modes should be exposed to usespace via sysfs interface and we should implement it. Also #1 and #2 can be handled via some changes in the design as mentioned in above discussion. I have one doubt though when we implement this via reboot-mode framework. The current patch implements PSCI ARM PSCI SYSTEM RESET2 vendor reset types. psci driver is initialized very early at boot but potential ARM psci reboot-mode driver will not probe at that stage and the ARM PSCI SYSTEM RESET2 vendor reset types functionality will not be available in psci reset path until the reboot-mode driver probes. Will this cause any limitation on usage of ARM's PSCI vendor-reset types for early device resets? One use-case may be an early device crash or a early reset where a vendor wants to use PSCI SYSTEM RESET2 vendor reset type to a reset the device to a specific state but may not be able to use this driver. (eg: a kernel panic at early boot where a vendor wants to reset device to a specific state using vendor reset. Currently panic passes a NULL (*arg command) while device reset but it may be explored for vendor specific reset). - Shivendra >> P.S. We appreciate Elliot for his work and follow-up on this while being >> employed at Qualcomm. > > Yes I sincerely do for his patience, thank you. > > Lorenzo > >>>> 3. psci cpuidle driver already registers a driver against the >>>> arm,psci-1.0 compatible. Refactoring would be needed to have both a >>>> cpuidle and reboot-mode driver. >>>> >>>> Signed-off-by: Elliot Berman <elliot.berman@oss.qualcomm.com> >>>> --- >>>> drivers/firmware/psci/psci.c | 105 +++++++++++++++++++++++++++++++++++++++++++ >>>> 1 file changed, 105 insertions(+) >>>> >>>> diff --git a/drivers/firmware/psci/psci.c b/drivers/firmware/psci/psci.c >>>> index a1ebbe9b73b136218e9d9f9b8daa7756b3ab2fbe..6f8c47deaec0225f26704e1f3bcad52603127a85 100644 >>>> --- a/drivers/firmware/psci/psci.c >>>> +++ b/drivers/firmware/psci/psci.c >>>> @@ -80,6 +80,14 @@ static u32 psci_cpu_suspend_feature; >>>> static bool psci_system_reset2_supported; >>>> static bool psci_system_off2_hibernate_supported; >>>> >>>> +struct psci_reset_param { >>>> + const char *mode; >>>> + u32 reset_type; >>>> + u32 cookie; >>>> +}; >>>> +static struct psci_reset_param *psci_reset_params __ro_after_init; >>>> +static size_t num_psci_reset_params __ro_after_init; >>>> + >>>> static inline bool psci_has_ext_power_state(void) >>>> { >>>> return psci_cpu_suspend_feature & >>>> @@ -306,9 +314,39 @@ static int get_set_conduit_method(const struct device_node *np) >>>> return 0; >>>> } >>>> >>>> +static int psci_vendor_system_reset2(const char *cmd) >>>> +{ >>>> + unsigned long ret; >>>> + size_t i; >>>> + >>>> + for (i = 0; i < num_psci_reset_params; i++) { >>>> + if (!strcmp(psci_reset_params[i].mode, cmd)) { >>>> + ret = invoke_psci_fn(PSCI_FN_NATIVE(1_1, SYSTEM_RESET2), >>>> + psci_reset_params[i].reset_type, >>>> + psci_reset_params[i].cookie, 0); >>>> + /* >>>> + * if vendor reset fails, log it and fall back to >>>> + * architecture reset types >>> >>> That's not what the code does. >>> >> Ack. >> >> -Mukesh >> >>>> + */ >>>> + pr_err("failed to perform reset \"%s\": %ld\n", cmd, >>>> + (long)ret); >>>> + return 0; >>>> + } >>>> + } >>>> + >>>> + return -ENOENT; >>>> +} >>>> + >>>> static int psci_sys_reset(struct notifier_block *nb, unsigned long action, >>>> void *data) >>>> { >>>> + /* >>>> + * try to do the vendor system_reset2 >>>> + * If there wasn't a matching command, fall back to architectural resets >>>> + */ >>>> + if (data && !psci_vendor_system_reset2(data)) >>>> + return NOTIFY_DONE; >>>> + >>>> if ((reboot_mode == REBOOT_WARM || reboot_mode == REBOOT_SOFT) && >>>> psci_system_reset2_supported) { >>>> /* >>>> @@ -795,6 +833,73 @@ static const struct of_device_id psci_of_match[] __initconst = { >>>> {}, >>>> }; >>>> >>>> +#define REBOOT_PREFIX "mode-" >>>> + >>>> +static int __init psci_init_system_reset2_modes(void) >>>> +{ >>>> + const size_t len = strlen(REBOOT_PREFIX); >>>> + struct psci_reset_param *param; >>>> + struct device_node *psci_np __free(device_node) = NULL; >>>> + struct device_node *np __free(device_node) = NULL; >>>> + struct property *prop; >>>> + size_t count = 0; >>>> + u32 magic[2]; >>>> + int num; >>>> + >>>> + if (!psci_system_reset2_supported) >>>> + return 0; >>>> + >>>> + psci_np = of_find_matching_node(NULL, psci_of_match); >>>> + if (!psci_np) >>>> + return 0; >>>> + >>>> + np = of_find_node_by_name(psci_np, "reset-types"); >>>> + if (!np) >>>> + return 0; >>> >>> Related to my initial question above. If LINUX_REBOOT_CMD_RESTART2 *arg command, >>> is the actual reset to be issued, should we add a default mode "cold" >>> and, if SYSTEM_RESET2 is supported, a "warm" reset mode too ? >>> >>> It all boils down to what *arg represents - adding "cold" and "warm" >>> modes would remove the dependency on reboot_mode for resets issued >>> through LINUX_REBOOT_CMD_RESTART2, the question is whether this >>> is the correct thing to do. >>> >>> Comments very welcome. >>> >>> Thanks, >>> Lorenzo >>> >>>> + >>>> + for_each_property_of_node(np, prop) { >>>> + if (strncmp(prop->name, REBOOT_PREFIX, len)) >>>> + continue; >>>> + num = of_property_count_u32_elems(np, prop->name); >>>> + if (num != 1 && num != 2) >>>> + continue; >>>> + >>>> + count++; >>>> + } >>>> + >>>> + param = psci_reset_params = >>>> + kcalloc(count, sizeof(*psci_reset_params), GFP_KERNEL); >>>> + if (!psci_reset_params) >>>> + return -ENOMEM; >>>> + >>>> + for_each_property_of_node(np, prop) { >>>> + if (strncmp(prop->name, REBOOT_PREFIX, len)) >>>> + continue; >>>> + >>>> + num = of_property_read_variable_u32_array(np, prop->name, magic, >>>> + 1, ARRAY_SIZE(magic)); >>>> + if (num < 0) { >>>> + pr_warn("Failed to parse vendor reboot mode %s\n", >>>> + param->mode); >>>> + kfree_const(param->mode); >>>> + continue; >>>> + } >>>> + >>>> + param->mode = kstrdup_const(prop->name + len, GFP_KERNEL); >>>> + if (!param->mode) >>>> + continue; >>>> + >>>> + /* Force reset type to be in vendor space */ >>>> + param->reset_type = PSCI_1_1_RESET_TYPE_VENDOR_START | magic[0]; >>>> + param->cookie = num > 1 ? magic[1] : 0; >>>> + param++; >>>> + num_psci_reset_params++; >>>> + } >>>> + >>>> + return 0; >>>> +} >>>> +arch_initcall(psci_init_system_reset2_modes); >>>> + >>>> int __init psci_dt_init(void) >>>> { >>>> struct device_node *np; >>>> >>>> -- >>>> 2.34.1 >>>>
The PSCI SYSTEM_RESET2 call allows vendor firmware to define additional reset types which could be mapped to the reboot argument. Setting up reboot on Qualcomm devices can be inconsistent from chipset to chipset. Generally, there is a PMIC register that gets written to decide the reboot type. There is also sometimes a cookie that can be written to indicate that the bootloader should behave differently than a regular boot. These knobs evolve over product generations and require more drivers. Qualcomm firmwares are beginning to expose vendor SYSTEM_RESET2 types to simplify driver requirements from Linux. Add support in PSCI to statically wire reboot mode commands from userspace to a vendor reset and cookie value using the device tree. The DT bindings are similar to reboot mode framework except that 2 integers are accepted (the type and cookie). Also, reboot mode framework is intended to program the cookies, but not actually reboot the host. PSCI SYSTEM_RESET2 does both. I've not added support for reading ACPI tables since I don't have any device which provides them + firmware that supports vendor SYSTEM_RESET2 types. Previous discussions around SYSTEM_RESET2: - https://lore.kernel.org/lkml/20230724223057.1208122-2-quic_eberman@quicinc.com/T/ - https://lore.kernel.org/all/4a679542-b48d-7e11-f33a-63535a5c68cb@quicinc.com/ Signed-off-by: Elliot Berman <quic_eberman@quicinc.com> Changes in v9: - Don't fallback to architecturally defined resets from Lorenzo. - Link to v8: https://lore.kernel.org/r/20241107-arm-psci-system_reset2-vendor-reboots-v8-0-e8715fa65cb5@quicinc.com Changes in v8: - Code style nits from Stephen - Add rb3gen2 - Link to v7: https://lore.kernel.org/r/20241028-arm-psci-system_reset2-vendor-reboots-v7-0-a4c40b0ebc54@quicinc.com Changes in v7: - Code style nits from Stephen - Dropped unnecessary hunk from the sa8775p-ride patch - Link to v6: https://lore.kernel.org/r/20241018-arm-psci-system_reset2-vendor-reboots-v6-0-50cbe88b0a24@quicinc.com Changes in v6: - Rebase to v6.11 and fix trivial conflicts in qcm6490-idp - Add sa8775p-ride support (same as qcm6490-idp) - Link to v5: https://lore.kernel.org/r/20240617-arm-psci-system_reset2-vendor-reboots-v5-0-086950f650c8@quicinc.com Changes in v5: - Drop the nested "items" in prep for future dtschema tools - Link to v4: https://lore.kernel.org/r/20240611-arm-psci-system_reset2-vendor-reboots-v4-0-98f55aa74ae8@quicinc.com Changes in v4: - Change mode- properties from uint32-matrix to uint32-array - Restructure the reset-types node so only the restriction is in the if/then schemas and not the entire definition - Link to v3: https://lore.kernel.org/r/20240515-arm-psci-system_reset2-vendor-reboots-v3-0-16dd4f9c0ab4@quicinc.com Changes in v3: - Limit outer number of items to 1 for mode-* properties - Move the reboot-mode for psci under a subnode "reset-types" - Fix the DT node in qcm6490-idp so it doesn't overwrite the one from sc7820.dtsi - Link to v2: https://lore.kernel.org/r/20240414-arm-psci-system_reset2-vendor-reboots-v2-0-da9a055a648f@quicinc.com Changes in v2: - Fixes to schema as suggested by Rob and Krzysztof - Add qcm6490 idp as first Qualcomm device to support - Link to v1: https://lore.kernel.org/r/20231117-arm-psci-system_reset2-vendor-reboots-v1-0-03c4612153e2@quicinc.com Changes in v1: - Reference reboot-mode bindings as suggeted by Rob. - Link to RFC: https://lore.kernel.org/r/20231030-arm-psci-system_reset2-vendor-reboots-v1-0-dcdd63352ad1@quicinc.com --- Elliot Berman (5): dt-bindings: arm: Document reboot mode magic firmware: psci: Read and use vendor reset types arm64: dts: qcom: qcm6490-idp: Add PSCI SYSTEM_RESET2 types arm64: dts: qcom: qcs6490-rb3gen2: Add PSCI SYSTEM_RESET2 types arm64: dts: qcom: sa8775p-ride: Add PSCI SYSTEM_RESET2 types Documentation/devicetree/bindings/arm/psci.yaml | 43 ++++++++++ arch/arm64/boot/dts/qcom/qcm6490-idp.dts | 7 ++ arch/arm64/boot/dts/qcom/qcs6490-rb3gen2.dts | 7 ++ arch/arm64/boot/dts/qcom/sa8775p-ride.dtsi | 7 ++ arch/arm64/boot/dts/qcom/sa8775p.dtsi | 2 +- arch/arm64/boot/dts/qcom/sc7280.dtsi | 2 +- drivers/firmware/psci/psci.c | 105 ++++++++++++++++++++++++ 7 files changed, 171 insertions(+), 2 deletions(-) --- base-commit: ffd294d346d185b70e28b1a28abe367bbfe53c04 change-id: 20231016-arm-psci-system_reset2-vendor-reboots-cc3ad456c070 Best regards,