[PATCH phosphor-host-ipmid v3] Add ipmi coldReset command, which call a dbus method, belongs to NETFUN_APP. 1. A method dbus_warm_reset() to talk with the dbus method 'warmRest'; 2. Also get service name by ipmid_get_sd_bus_connection() instead of object_mapper_get_connection(); 3. Register the ipmi command; 4. Add related .o to the Makefile; 5. Add wildcard function.
Nan KX Li
bjlinan at cn.ibm.com
Tue Feb 2 15:15:53 AEDT 2016
Hi,
Thanks Sam.
I will think of it.
Regards,
William Li ( Li Nan, 李楠 )
Firmware Engineering Professional
OpenPower AE Team | IBM System & Technology Lab
Mobile: +86-186-1081 6605, +86-138-1188 5954
Beijing, China
From: Sam Mendoza-Jonas <sam at mendozajonas.com>
To: OpenBMC Patches <openbmc-patches at stwcx.xyz>
Cc: openbmc at lists.ozlabs.org, Nan KX Li/China/IBM at IBMCN
Date: 02/02/2016 11:44
Subject: Re: [PATCH phosphor-host-ipmid v3] Add ipmi coldReset
command, which call a dbus method, belongs to NETFUN_APP. 1. A method
dbus_warm_reset() to talk with the dbus method 'warmRest'; 2. Also get
service name by ipmid_get_sd_bus_connection() instead of
object_mapper_get_connection(); 3. Register the ipmi command; 4. Add
related .o to the Makefile; 5. Add wildcard function.
Hi William, a question below;
On Mon, Feb 01, 2016 at 12:40:25AM -0600, OpenBMC Patches wrote:
> From: William <bjlinan at cn.ibm.com>
>
> ---
> Makefile | 1 +
> globalhandler.C | 171
++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> globalhandler.h | 12 ++++
> 3 files changed, 184 insertions(+)
> create mode 100644 globalhandler.C
> create mode 100644 globalhandler.h
>
> diff --git a/Makefile b/Makefile
> index 472bfaa..58e7310 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -15,6 +15,7 @@ LIB_APP_OBJ = apphandler.o \
> ipmisensor.o \
> storageaddsel.o \
> transporthandler.o \
> + globalhandler.o
>
> LIB_HOST_SRV_OBJ = host-services.o
>
> diff --git a/globalhandler.C b/globalhandler.C
> new file mode 100644
> index 0000000..e68ea79
> --- /dev/null
> +++ b/globalhandler.C
> @@ -0,0 +1,171 @@
> +#include "globalhandler.h"
> +#include "ipmid-api.h"
> +#include <stdio.h>
> +#include <string.h>
> +#include <stdint.h>
> +
> +const char *control_object_name = "/org/openbmc/control/bmc0";
> +const char *control_intf_name = "org.openbmc.control.Bmc";
> +
> +const char *objectmapper_service_name = "org.openbmc.objectmapper";
> +const char *objectmapper_object_name =
"/org/openbmc/objectmapper/objectmapper";
> +const char *objectmapper_intf_name =
"org.openbmc.objectmapper.ObjectMapper";
> +
> +void register_netfn_global_functions() __attribute__((constructor));
> +
> +int obj_mapper_get_connection(char** buf, const char* obj_path)
> +{
> + sd_bus_error error = SD_BUS_ERROR_NULL;
> + sd_bus_message *m = NULL;
> + sd_bus *bus = NULL;
> + char *temp_buf = NULL, *intf = NULL;
> + size_t buf_size = 0;
> + int r;
> +
> + //Get the system bus where most system services are provided.
> + bus = ipmid_get_sd_bus_connection();
> +
> + /*
> + * Bus, service, object path, interface and method are provided to
call
> + * the method.
> + * Signatures and input arguments are provided by the arguments at
the
> + * end.
> + */
> + r = sd_bus_call_method(bus,
> + objectmapper_service_name, /* service
to contact */
> + objectmapper_object_name, /* object
path */
> + objectmapper_intf_name, /*
interface name */
> + "GetObject", /* method
name */
> + &error, /* object
to return error in */
> + &m, /* return
message on success */
> + "s", /* input
signature */
> + obj_path /* first
argument */
> + );
> +
> + if (r < 0) {
> + fprintf(stderr, "Failed to issue method call: %s\n",
error.message);
> + goto finish;
> + }
> +
> + // Get the key, aka, the connection name
> + sd_bus_message_read(m, "a{sas}", 1, &temp_buf, 1, &intf);
> +
> + /*
> + * TODO: check the return code. Currently for no reason the message
> + * parsing of object mapper is always complaining about
> + * "Device or resource busy", but the result seems OK for now. Need
> + * further checks.
> + */
I've seen this comment before and it worries me a bit. Why are we
getting the error? How do you know temp_buf is allocated properly? What
does it mean that the result "seems OK"?
You check return codes before and after this, but if you get unlucky
here you could easily segfault or read nonsense.
I think this issue should be figured out before merging the code, both
here and wherever else it appears.
Thanks,
Sam
> +
> + buf_size = strlen(temp_buf) + 1;
> + printf("IPMID connection name: %s\n", temp_buf);
> + *buf = (char*)malloc(buf_size);
> +
> + if (*buf == NULL) {
> + fprintf(stderr, "Malloc failed for warm reset");
> + r = -1;
> + goto finish;
> + }
> +
> + memcpy(*buf, temp_buf, buf_size);
> +
> +finish:
> + sd_bus_error_free(&error);
> + sd_bus_message_unref(m);
> +
> + return r;
> +}
> +
> +int dbus_warm_reset()
> +{
> + sd_bus_error error = SD_BUS_ERROR_NULL;
> + sd_bus_message *m = NULL;
> + sd_bus *bus = NULL;
> + char* temp_buf = NULL;
> + uint8_t* get_value = NULL;
> + char* connection = NULL;
> + int r, i;
> +
> + r = obj_mapper_get_connection(&connection, control_object_name);
> + if (r < 0) {
> + fprintf(stderr, "Failed to get connection, return value:
%d.\n", r);
> + goto finish;
> + }
> +
> + printf("connection: %s\n", connection);
> +
> + // Open the system bus where most system services are provided.
> + bus = ipmid_get_sd_bus_connection();
> +
> + /*
> + * Bus, service, object path, interface and method are provided to
call
> + * the method.
> + * Signatures and input arguments are provided by the arguments at
the
> + * end.
> + */
> + r = sd_bus_call_method(bus,
> + connection, /* service to
contact */
> + control_object_name, /* object path
*/
> + control_intf_name, /* interface
name */
> + "warmReset", /* method name
*/
> + &error, /* object to
return error in */
> + &m, /* return
message on success */
> + NULL,
> + NULL
> + );
> +
> + if (r < 0) {
> + fprintf(stderr, "Failed to issue method call: %s\n",
error.message);
> + goto finish;
> + }
> +
> +finish:
> + sd_bus_error_free(&error);
> + sd_bus_message_unref(m);
> + free(connection);
> +
> + return r;
> +}
> +
> +ipmi_ret_t ipmi_global_warm_reset(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> + ipmi_request_t request, ipmi_response_t
response,
> + ipmi_data_len_t data_len, ipmi_context_t
context)
> +{
> + printf("Handling GLOBAL warmReset Netfn:[0x%X],
Cmd:[0x%X]\n",netfn, cmd);
> +
> + // TODO: call the correct dbus method for warmReset.
> + dbus_warm_reset();
> +
> + // Status code.
> + ipmi_ret_t rc = IPMI_CC_OK;
> + *data_len = 0;
> + return rc;
> +}
> +
> +ipmi_ret_t ipmi_global_wildcard_handler(ipmi_netfn_t netfn, ipmi_cmd_t
cmd,
> + ipmi_request_t request, ipmi_response_t
response,
> + ipmi_data_len_t data_len, ipmi_context_t
context)
> +{
> + printf("Handling WILDCARD Netfn:[0x%X], Cmd:[0x%X]\n",netfn, cmd);
> +
> + // Status code.
> + ipmi_ret_t rc = IPMI_CC_OK;
> +
> + *data_len = strlen("THIS IS WILDCARD");
> +
> + // Now pack actual response
> + memcpy(response, "THIS IS WILDCARD", *data_len);
> +
> + return rc;
> +}
> +
> +void register_netfn_global_functions()
> +{
> + printf("Registering NetFn:[0x%X], Cmd:[0x%X]\n",NETFUN_APP,
IPMI_CMD_WARM_RESET);
> + ipmi_register_callback(NETFUN_APP, IPMI_CMD_WARM_RESET, NULL,
ipmi_global_warm_reset);
> +
> + printf("Registering NetFn:[0x%X], Cmd:[0x%X]\n",NETFUN_APP,
IPMI_CMD_WILDCARD);
> + ipmi_register_callback(NETFUN_APP, IPMI_CMD_WILDCARD, NULL,
ipmi_global_wildcard_handler);
> +
> + return;
> +}
> diff --git a/globalhandler.h b/globalhandler.h
> new file mode 100644
> index 0000000..608df3b
> --- /dev/null
> +++ b/globalhandler.h
> @@ -0,0 +1,12 @@
> +#ifndef __HOST_IPMI_GLOBAL_HANDLER_H__
> +#define __HOST_IPMI_GLOBAL_HANDLER_H__
> +
> +#include <stdint.h>
> +
> +// Various GLOBAL operations under a single command.
> +enum ipmi_global_control_cmds : uint8_t
> +{
> +IPMI_CMD_WARM_RESET =
0x02,
> +};
> +
> +#endif
> --
> 2.6.4
>
>
> _______________________________________________
> openbmc mailing list
> openbmc at lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/openbmc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openbmc/attachments/20160202/a8321db5/attachment-0001.html>
More information about the openbmc
mailing list