[PATCH phosphor-host-ipmid] Adding Boot Policy
Cyril Bur
cyrilbur at gmail.com
Mon Apr 18 18:48:55 AEST 2016
On Mon, 18 Apr 2016 08:38:58 +0000
"Ratan K Gupta" <ratagupt at in.ibm.com> wrote:
> Hi Cyril,
>
> Please find my comments inline.
>
> Regards
> Ratan
> ----- Original message -----
> From: Cyril Bur <cyrilbur at gmail.com>
> To: Ratan K Gupta/India/IBM at IBMIN
> Cc: OpenBMC Patches <openbmc-patches at stwcx.xyz>, openbmc at lists.ozlabs.org
> Subject: Re: [PATCH phosphor-host-ipmid] Adding Boot Policy
> Date: Tue, Apr 5, 2016 7:03 AM
>
> On Mon, 4 Apr 2016 06:30:36 -0500
> OpenBMC Patches <openbmc-patches at stwcx.xyz> wrote:
>
> Hi,
>
> Thanks for the contribution. I have some concerns, most are summarised here:
> https://github.com/openbmc/docs/blob/master/contributing.md
>
> You'll want to note
> https://github.com/openbmc/docs/blob/master/contributing.md#coding-style and
> https://github.com/openbmc/docs/blob/master/contributing.md#submitting-changes
>
> I've added further comments inline.
>
> Thanks,
>
> Cyril
>
> > From: ratagupt <ratagupt at in.ibm.com>
> >
> > ---
> > chassishandler.C | 59 ++++++++++++++++++++++++++++++++++++++++++++------------
> > 1 file changed, 47 insertions(+), 12 deletions(-)
> >
> > diff --git a/chassishandler.C b/chassishandler.C
> > index fca3c79..1a5508f 100644
> > --- a/chassishandler.C
> > +++ b/chassishandler.C
> > @@ -92,7 +92,7 @@ finish:
> > return r;
> > }
> >
> > -int dbus_get_property(char **buf)
> > +int dbus_get_property(const char *name,char **buf)
>
> The style guide recommends spaces between parameters.
>
> >>>>RG:- Done
>
> > {
> > sd_bus_error error = SD_BUS_ERROR_NULL;
> > sd_bus_message *m = NULL;
> > @@ -128,7 +128,7 @@ int dbus_get_property(char **buf)
> > &m, /* return message on success */
> > "ss", /* input signature */
> > host_intf_name, /* first argument */
> > - "boot_flags"); /* second argument */
> > + name); /* second argument */
> >
> > if (r < 0) {
> > fprintf(stderr, "Failed to issue method call: %s\n", error.message);
> > @@ -161,7 +161,7 @@ finish:
> > return r;
> > }
> >
> > -int dbus_set_property(const char *buf)
> > +int dbus_set_property(const char * name,const char *value)
> > {
> > sd_bus_error error = SD_BUS_ERROR_NULL;
> > sd_bus_message *m = NULL;
> > @@ -196,16 +196,16 @@ int dbus_set_property(const char *buf)
> > &m, /* return message on success */
> > "ssv", /* input signature */
> > host_intf_name, /* first argument */
> > - "boot_flags", /* second argument */
> > + name, /* second argument */
> > "s", /* third argument */
> > - buf); /* fourth argument */
> > + value); /* fourth argument */
>
> I'm not really sure why the /* {nth} argument */ comments but since they're
> there please try to keep them aligned...
> >>>>RG:Done
> >
> > if (r < 0) {
> > fprintf(stderr, "Failed to issue method call: %s\n", error.message);
> > goto finish;
> > }
> >
> > - printf("IPMID boot option property set: {%s}.\n", buf);
> > + printf("IPMID boot option property set: {%s}.\n", value);
> >
> > finish:
> > sd_bus_error_free(&error);
> > @@ -374,7 +374,8 @@ char* get_boot_option_by_ipmi(uint8_t p) {
> > }
> >
> > #define SET_PARM_VERSION 1
> > -#define SET_PARM_BOOT_FLAGS_VALID 0x80
> > +#define SET_PARM_BOOT_FLAGS_VALID_ONE_TIME 0x80
> > +#define SET_PARM_BOOT_FLAGS_VALID_PERMANENT 0xC0
> >
> > ipmi_ret_t ipmi_chassis_get_sys_boot_options(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> > ipmi_request_t request, ipmi_response_t response,
> > @@ -391,7 +392,7 @@ ipmi_ret_t ipmi_chassis_get_sys_boot_options(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> > memset(resp,0,sizeof(*resp));
> > resp->version = SET_PARM_VERSION;
> > resp->parm = 5;
> > - resp->data[0] = SET_PARM_BOOT_FLAGS_VALID;
> > + resp->data[0] = SET_PARM_BOOT_FLAGS_VALID_ONE_TIME;
> >
> > *data_len = sizeof(*resp);
> >
> > @@ -401,10 +402,11 @@ ipmi_ret_t ipmi_chassis_get_sys_boot_options(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> > */
> > if (reqptr->parameter == 5) {
> >
> > - int r = dbus_get_property(&p);
> > + /* Get the boot device */
> > + int r = dbus_get_property("boot_flags",&p);
> >
> > if (r < 0) {
> > - fprintf(stderr, "Dbus get property failed for get_sys_boot_options.\n");
> > + fprintf(stderr, "Dbus get property(boot_flags) failed for get_sys_boot_options.\n");
> > rc = IPMI_CC_UNSPECIFIED_ERROR;
> >
> > } else {
> > @@ -412,8 +414,31 @@ ipmi_ret_t ipmi_chassis_get_sys_boot_options(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> > s = get_ipmi_boot_option(p);
> > resp->data[1] = (s << 2);
> > rc = IPMI_CC_OK;
> > +
> > + }
> > +
> > + if (p)
> > + {
> > + free(p);
> > + p = NULL;
> > + }
>
> Just fyi free() will happily accept NULL as an argument and simply do nothing.
> >>>>RG:- I was following the approach which was there before.
> > +
> > + /* Get the boot policy */
> > + r = dbus_get_property("boot_policy",&p);
> > +
> > + if (r < 0) {
> > + fprintf(stderr, "Dbus get property(boot_policy) failed for get_sys_boot_options.\n");
> > + rc = IPMI_CC_UNSPECIFIED_ERROR;
> > +
> > + } else {
> > +
> > + printf("BootPolicy is[%s]", p);
> > + resp->data[0] = (strcmp(p,"ONETIME")==0)?SET_PARM_BOOT_FLAGS_VALID_ONE_TIME:SET_PARM_BOOT_FLAGS_VALID_PERMANENT;
>
> Wow, could you please make use of the spacebar and perhaps the enter key?
> RG:>>>>>>Done
>
> Also, just to be safe it would be best of use strncmp.
> For example strncmp(p, "ONETIME", strlen("ONETIME"))
>
> > + rc = IPMI_CC_OK;
> > +
> > }
> >
> > +
> > } else {
> > fprintf(stderr, "Unsupported parameter 0x%x\n", reqptr->parameter);
> > }
> > @@ -455,13 +480,23 @@ ipmi_ret_t ipmi_chassis_set_sys_boot_options(ipmi_netfn_t netfn, ipmi_cmd_t cmd,
> >
> > } else {
> >
> > - int r = dbus_set_property(s);
> > + int r = dbus_set_property("boot_flags",s);
> >
> > if (r < 0) {
> > - fprintf(stderr, "Dbus set property failed for set_sys_boot_options.\n");
> > + fprintf(stderr, "Dbus set property(boot_flags) failed for set_sys_boot_options.\n");
> > rc = IPMI_CC_UNSPECIFIED_ERROR;
> > }
> > }
> > +
> > + /* setting the boot policy */
> > + s= (char *)(((reqptr->data[0] & 0x40) == 0x40) ?"PERMANENT":"ONETIME");
>
> Firstly if you declare s as a 'const char *' (which by the looks of things it
> probably should be anyway...) then you won't need the cast. Why introduce 0x40?
> Wouldn't it be cleaner to just do
>
> if ((reqptr->data[0] & SET_PARM_BOOT_FLAGS_VALID_PERMANENT) ==
> SET_PARM_BOOT_FLAGS_VALID_PERMANENT)
> s = "PERMANENT";
> else
> s = "ONETIME";
> >>>>>>>RG: would not be able to check with SET_PARM_BOOT_FLAGS_VALID_PERMANENT as by doing this it checks for the two bits(6 and 7).But with the 0x40,We are only checking the one bit(6th bit).
Yes I understand that, which is why I added the
'== SET_PARM_BOOT_FLAGS_VALID_PERMANENT' so that the condition is only true if
both bits 6 and 7 are set.
For example: If only bit 7 is set then (reqptr->data[0] &
SET_PARM_BOOT_FLAGS_VALID_PERMANENT) will not ==
SET_PARM_BOOT_FLAGS_VALID_PERMANENT
> Anyways I have created the macro for better readability.
>
> > + printf ( "\nBoot Policy is %s",s);
> > + int r = dbus_set_property("boot_policy",s);
> > +
> > + if (r < 0) {
> > + fprintf(stderr, "Dbus set property(boot_policy) failed for set_sys_boot_options.\n");
> > + rc = IPMI_CC_UNSPECIFIED_ERROR;
> > + }
> >
> > } else {
> > fprintf(stderr, "Unsupported parameter 0x%x\n", reqptr->parameter);
>
>
>
More information about the openbmc
mailing list