Llewellyn/conky
1
0
Fork 0
mirror of https://github.com/Llewellynvdm/conky.git synced 2024-11-17 18:45:10 +00:00
Code Issues Releases Activity

Nvidia updates, docs and fixes (#278)

Browse Source 
* nvidia: Update the docs and add more arguments

**Updated the docs in variables.xml to reflect the current nvidia object.**

Basically just copied the inline comments into the XML document

**Added the last of the compatible arguments**

The following arguments were added to `get_nvidia_barval`:

* memfree
* ambienttemp
* gpufreq
* memfreq
* mtrfreq
* imagequality

This should leave only incompatible arguments unsupported (e.g. min/max
values).

Decided to base percentages off of zero rather than the specified minimum
values.

Signed-off-by: Matt Sturgeon <matt@sturgeon.me.uk>

* nvidia: Better error/debug printing

* Merged `scan_nvidia_args` and `set_nvidia_type` into new
  `set_nvidia_query`
* Added a translation array for `nvidia`, `nvidiabar`, `nvidiagauge` and
  `nvidiagraph` commands (helps with error printing)
* Added some NORM_ERR prints to `get_nvidia_barval`, `get_nvidia_value` and
  `get_nvidia_string`
* Removed exess newlines in nvidia's core.cc error definitions

Signed-off-by: Matt Sturgeon <matt@sturgeon.me.uk>

* nvidia: Better sanity checks when running queries

Introduced a pair of functions that checks whether the target we are
querying exists and how many of the target there are.

If the target doesn't exist there probably isn't a nvidia card
installed, if the target count query fails then the propietary drivers
probably aren't installed.

Fixes #269

Signed-off-by: Matt Sturgeon <matt@sturgeon.me.uk>
This commit is contained in:
Matt Sturgeon 2016-06-24 19:05:17 +01:00 committed by Brenden Matthews
parent 0d496ad008
commit 3753e2a95b
4 changed files with 356 additions and 65 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

254
doc/variables.xml
View File

@ -2777,49 +2777,251 @@
<command>
<option>nvidia</option>
</command>
<option>threshold</option>
<option>temp</option>
<option>ambient</option>
<option>gpufreq</option>
<option>memfreq</option>
<option>imagequality</option>
<option>argument</option>
</term>
<listitem>Nvidia graficcard support for the XNVCtrl
library. Each option can be shortened to the least
significant part. Temperatures are printed as float, all
other values as integer.
<simplelist>
<listitem>Nvidia graphics card information via the XNVCtrl
library.
<para />
<emphasis>Possible arguments:</emphasis> (Temperatures are
printed as float, all other values as integer. Bracketed
arguments are aliases)
<simplelist type='horiz' columns='2'>
<!-- Temperatures -->
<member>
<command>threshold</command>
<option>The thresholdtemperature at
which the gpu slows down</option>
<command>gputemp</command>
(<command>temp</command>)
<option>GPU temperature</option>
</member>
<member>
<command>temp</command>
<option>Gives the gpu current
temperature</option>
<command>gputempthreshold</command>
(<command>threshold</command>)
<option>Temperature threshold where the GPU will reduce it's clock speed</option>
</member>
<member>
<command>ambient</command>
<option>Gives current air temperature near GPU
case</option>
<command>ambienttemp</command>
(<command>ambient</command>)
<option>Ambient temperature outside the graphics card</option>
</member>
<!-- GPU frequency -->
<member>
<command>gpufreqcur</command>
(<command>gpufreq</command>)
<option>Current GPU clock speed</option>
</member>
<member>
<command>gpufreq</command>
<option>Gives the current gpu frequency</option>
<command>gpufreqmin</command>
<option>Minimum GPU clock speed</option>
</member>
<member>
<command>memfreq</command>
<option>Gives the current mem frequency</option>
<command>gpufreqmax</command>
<option>Maximum GPU clock speed</option>
</member>
<!-- Memory frequency -->
<member>
<command>memfreqcur</command>
(<command>memfreq</command>)
<option>Current memory clock speed</option>
</member>
<member>
<command>memfreqmin</command>
<option>Minimum memory clock speed</option>
</member>
<member>
<command>memfreqmax</command>
<option>Maximum memory clock speed</option>
</member>
<!-- Memory transfer rate frequency -->
<member>
<command>mtrfreqcur</command>
(<command>mtrfreq</command>)
<option>Current memory transfer rate clock speed</option>
</member>
<member>
<command>mtrfreqmin</command>
<option>Minimum memory transfer rate clock speed</option>
</member>
<member>
<command>mtrfreqmax</command>
<option>Maximum memory transfer rate clock speed</option>
</member>
<!-- Performance levels -->
<member>
<command>perflevelcur</command>
(<command>perflevel</command>)
<option>Current performance level</option>
</member>
<member>
<command>perflevelmin</command>
<option>Lowest performance level</option>
</member>
<member>
<command>perflevelmax</command>
<option>Highest performance level</option>
</member>
<member>
<command>perfmode</command>
<option>Performance mode</option>
</member>
<!-- Load/utilization -->
<member>
<command>gpuutil</command>
<option>GPU utilization %</option>
</member>
<member>
<command>membwutil</command>
<option>Memory bandwidth utilization %</option>
</member>
<member>
<command>videoutil</command>
<option>Video engine utilization %</option><!-- ??? -->
</member>
<member>
<command>pcieutil</command>
<option>PCIe bandwidth utilization %</option>
</member>
<!-- RAM statistics -->
<member>
<command>memused</command>
(<command>mem</command>)
<option>Amount of used memory</option>
</member>
<member>
<command>memfree</command>
(<command>memavail</command>)
<option>Amount of free memory</option>
</member>
<member>
<command>memmax</command>
(<command>memtotal</command>)
<option>Total amount of memory</option>
</member>
<member>
<command>memutil</command>
(<command>memperc</command>)
<option>Memory utilization %</option>
</member>
<!-- Fan/cooler -->
<member>
<command>fanspeed</command>
<option>Fan speed</option>
</member>
<member>
<command>fanlevel</command>
<option>Fan level %</option>
</member>
<!-- Miscellaneous -->
<member>
<command>imagequality</command>
<option>Which imagequality should be chosen by
OpenGL applications</option>
<option>Image quality</option>
</member>
</simplelist>
<para /></listitem>
</varlistentry>
<varlistentry>
<term>
<command>
<option>nvidiabar</option>
</command>
<option>(height),(width)</option>
<option>argument</option>
</term>
<listitem>Same as nvidia, except it draws its output in a
horizontal bar. The height and width parameters are optional,
and default to the default_bar_height and default_bar_width
config settings, respectively.
<para />
<emphasis>Note the following arguments are incompatible:</emphasis>
<simplelist type='horiz' columns='3'>
<member>
<command>gputempthreshold</command>
(<command>threshold</command>)
</member>
<member>
<command>gpufreqmin</command>
</member>
<member>
<command>gpufreqmax</command>
</member>
<member>
<command>memfreqmin</command>
</member>
<member>
<command>memfreqmax</command>
</member>
<member>
<command>mtrfreqmin</command>
</member>
<member>
<command>mtrfreqmax</command>
</member>
<member>
<command>perflevelmin</command>
</member>
<member>
<command>perflevelmax</command>
</member>
<member>
<command>perfmode</command>
</member>
<member>
<command>memtotal</command>
(<command>memmax</command>)
</member>
<member>
<command>fanspeed</command>
</member>
</simplelist>
<para /></listitem>
</varlistentry>
<varlistentry>
<term>
<command>
<option>nvidiagauge</option>
</command>
<option>(height),(width)</option>
<option>argument</option>
</term>
<listitem>Same as nvidiabar, except a round gauge
(much like a vehicle speedometer). The height
and width parameters are optional, and default to the
default_gauge_height and default_gauge_width config
settings, respectively.
<para />
For possible arguments see nvidia and nvidiabar.
<para /></listitem>
</varlistentry>
<varlistentry>
<term>
<command>
<option>nvidiagraph</option>
</command>
<option>argument</option>
<option>(height),(width)</option>
<option>(gradient color 1)</option>
<option>(gradient color 2)</option>
<option>(scale)</option>
<option>(-t)</option>
<option>(-l)</option>
</term>
<listitem>Same as nvidiabar, except a horizontally
scrolling graph with values from 0-100 plotted on the
vertical axis. The height and width parameters are
optional, and default to the default_graph_height and
default_graph_width config settings, respectively.
<para />
For possible arguments see nvidia and nvidiabar. To learn more
about the -t -l and gradient color options,
see execgraph.
<para /></listitem>
</varlistentry>
<varlistentry>
<term>
<command>

16
src/core.cc
View File

@ -1802,30 +1802,30 @@ struct text_object *construct_text_object(char *s, const char *arg,
obj->callbacks.free = &free_combine;
#ifdef BUILD_NVIDIA
END OBJ_ARG(nvidia, 0, "nvidia needs an argument")
if (set_nvidia_type(obj, arg)) {
if (set_nvidia_query(obj, arg, NONSPECIAL)) {
CRIT_ERR(obj, free_at_crash, "nvidia: invalid argument"
" specified: '%s'\n", arg);
" specified: '%s'", arg);
}
obj->callbacks.print = &print_nvidia_value;
obj->callbacks.free = &free_nvidia;
END OBJ_ARG(nvidiabar, 0, "nvidiabar needs an argument")
if (scan_nvidia_args(obj, arg, BAR)) {
if (set_nvidia_query(obj, arg, BAR)) {
CRIT_ERR(obj, free_at_crash, "nvidiabar: invalid argument"
" specified: '%s'\n", arg);
" specified: '%s'", arg);
}
obj->callbacks.barval = &get_nvidia_barval;
obj->callbacks.free = &free_nvidia;
END OBJ_ARG(nvidiagraph, 0, "nvidiagraph needs an argument")
if (scan_nvidia_args(obj, arg, GRAPH)) {
if (set_nvidia_query(obj, arg, GRAPH)) {
CRIT_ERR(obj, free_at_crash, "nvidiagraph: invalid argument"
" specified: '%s'\n", arg);
" specified: '%s'", arg);
}
obj->callbacks.graphval = &get_nvidia_barval;
obj->callbacks.free = &free_nvidia;
END OBJ_ARG(nvidiagauge, 0, "nvidiagauge needs an argument")
if (scan_nvidia_args(obj, arg, GAUGE)) {
if (set_nvidia_query(obj, arg, GAUGE)) {
CRIT_ERR(obj, free_at_crash, "nvidiagauge: invalid argument"
" specified: '%s'\n", arg);
" specified: '%s'", arg);
}
obj->callbacks.gaugeval = &get_nvidia_barval;
obj->callbacks.free = &free_nvidia;

154
src/nvidia.cc
View File

@ -43,7 +43,6 @@
* so that all quirks are located there
* - Implement nvs->print_type to allow control over how the value is printed
* (int, float, temperature...)
* - Rename set_nvidia_type() to set_nvidia_query (requires changes in core.cc)
*
* Showcase (conky.conf):
* --==| NVIDIA | ==--
@ -282,9 +281,31 @@ typedef enum _SEARCH_ID {
SEARCH_MAX
} SEARCH_ID;
// Translate special_type into command string
const char* translate_nvidia_special_type[] = {
"nvidia", // NONSPECIAL
"", // HORIZONTAL_LINE
"", // STIPPLED_HR
"nvidiabar", // BAR
"", // FG
"", // BG
"", // OUTLINE
"", // ALIGNR
"", // ALIGNC
"nvidiagague", // GAUGE
"nvidiagraph", // GRAPH
"", // OFFSET
"", // VOFFSET
"", // FONT
"", // GOTO
"" // TAB
};
// Global struct to keep track of queries
struct nvidia_s {
const char *command;
const char *arg;
QUERY_ID query;
TARGET_ID target;
ATTR_ID attribute;
@ -340,34 +361,9 @@ namespace {
nvidia_display_setting nvidia_display;
}
// Extract arguments for nvidiabar, etc, and run set_nvidia_type
int scan_nvidia_args (struct text_object *obj, const char *args, unsigned int special_t) {
const char *arg = args;
switch (special_t) {
case BAR:
arg = scan_bar(obj, arg, 100);
break;
case GRAPH:
arg = scan_graph(obj, arg, 100);
break;
case GAUGE:
arg = scan_gauge(obj, arg, 100);
break;
default:
return 1;
}
// Return error if no argument
// (sometimes scan_graph gets excited and eats the whole string!
if (!arg) return 1;
return set_nvidia_type(obj, arg);
}
// Evaluate module parameters and prepare query
int set_nvidia_type(struct text_object *obj, const char *arg)
int set_nvidia_query(struct text_object *obj, const char *arg, unsigned int special_type)
{
struct nvidia_s *nvs;
int aid;
@ -377,6 +373,23 @@ int set_nvidia_type(struct text_object *obj, const char *arg)
nvs = static_cast<nvidia_s *>(obj->data.opaque);
memset(nvs, 0, sizeof(struct nvidia_s));
// Extract arguments for nvidiabar, etc, and run set_nvidia_query
switch (special_type) {
case BAR:
arg = scan_bar(obj, arg, 100);
break;
case GRAPH:
arg = scan_graph(obj, arg, 100);
break;
case GAUGE:
arg = scan_gauge(obj, arg, 100);
break;
}
// Return error if no argument
// (sometimes scan_graph gets excited and eats the whole string!
if (!arg) return 1;
// Translate parameter to id
for (aid=0; aid < ARG_UNKNOWN; aid++) {
if (strcmp(arg, translate_module_argument[aid]) == 0)
@ -384,6 +397,10 @@ int set_nvidia_type(struct text_object *obj, const char *arg)
}
//fprintf(stderr, "parameter: %s -> aid: %d\n", arg, aid);
// Save pointers to the arg and command strings for dubugging and printing
nvs->arg = translate_module_argument[aid];
nvs->command = translate_nvidia_special_type[special_type];
// Evaluate parameter
switch(aid) {
@ -569,11 +586,37 @@ int set_nvidia_type(struct text_object *obj, const char *arg)
break;
default: // Unknown/invalid argument
// Error printed by core.cc
return 1;
}
return 0;
}
// Return the amount of targets present (or -1 on error)
static inline int get_nvidia_target_count(Display *dpy, TARGET_ID tid)
{
int num_tgts;
if (!XNVCTRLQueryTargetCount(dpy, translate_nvidia_target[tid], &num_tgts)) {
num_tgts = -1;
}
return num_tgts;
}
// Exit if we are unable to get targets of type tid on display dpy
void check_nvidia_target_count(Display *dpy, TARGET_ID tid, ATTR_ID aid)
{
int num_tgts = get_nvidia_target_count(dpy, tid);
if(num_tgts < 1) {
// Print error and exit if there's not enough targets to query
CRIT_ERR(NULL, NULL, "%s:"
"\n Trying to query Nvidia target failed (using the propietary drivers)."
"\n Are you sure they are installed correctly and a Nvidia GPU is in use?"
"\n (display: %d, target_id: %d, target_count: %d, attribute_id: %d)"
, __func__, dpy, tid, num_tgts, aid);
}
}
// Retrieve attribute value via nvidia interface
static int get_nvidia_value(TARGET_ID tid, ATTR_ID aid)
@ -581,8 +624,12 @@ static int get_nvidia_value(TARGET_ID tid, ATTR_ID aid)
Display *dpy = nvdisplay ? nvdisplay : display;
int value;
// Check for issues
check_nvidia_target_count(dpy, tid, aid);
// Query nvidia interface
if(!dpy || !XNVCTRLQueryTargetAttribute(dpy, translate_nvidia_target[tid], 0, 0, translate_nvidia_attribute[aid], &value)){
NORM_ERR("%s: Something went wrong running nvidia query (tid: %d, aid: %d)", __func__, tid, aid);
return -1;
}
@ -603,8 +650,12 @@ static char* get_nvidia_string(TARGET_ID tid, ATTR_ID aid)
Display *dpy = nvdisplay ? nvdisplay : display;
char *str;
// Check for issues
check_nvidia_target_count(dpy, tid, aid);
// Query nvidia interface
if (!dpy || !XNVCTRLQueryTargetStringAttribute(dpy, translate_nvidia_target[tid], 0, 0, translate_nvidia_attribute[aid], &str)) {
NORM_ERR("%s: Something went wrong running nvidia string query (tid: %d, aid: %d)", __func__, tid, aid);
return NULL;
}
//fprintf(stderr, "%s", str);
@ -746,19 +797,28 @@ double get_nvidia_barval(struct text_object *obj) {
// Assume failure
value = 0;
// Convert query_result to a percentage
// Convert query_result to a percentage using ((val-min)÷(max-min)×100)+0.5 if needed.
if (nvs != NULL) {
switch (nvs->attribute) {
case ATTR_UTILS_STRING: // one of the percentage utils (gpuutil, membwutil, videoutil and pcieutil)
value = get_nvidia_string_value(nvs->target, ATTR_UTILS_STRING, nvs->token, nvs->search);
break;
case ATTR_MEM_UTIL: // memutil
case ATTR_MEM_USED:
temp1 = get_nvidia_value(nvs->target, ATTR_MEM_USED);
temp2 = get_nvidia_value(nvs->target, ATTR_MEM_TOTAL);
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
case ATTR_MEM_FREE: // memfree
temp1 = get_nvidia_value(nvs->target, ATTR_MEM_USED);
temp2 = get_nvidia_value(nvs->target, ATTR_MEM_TOTAL);
value = temp2 - temp1;
break;
case ATTR_FAN_SPEED: // fanspeed: Warn user we are using fanlevel
NORM_ERR("%s: invalid argument specified: '%s' (using 'fanlevel' instead).",
nvs->command, nvs->arg);
// No break, continue into fanlevel
case ATTR_FAN_LEVEL: // fanlevel
case ATTR_FAN_SPEED: // TODO warn user to use fanlevel if they use fanspeed
value = get_nvidia_value(nvs->target, ATTR_FAN_LEVEL);
break;
case ATTR_GPU_TEMP: // gputemp (calculate out of gputempthreshold)
@ -766,10 +826,40 @@ double get_nvidia_barval(struct text_object *obj) {
temp2 = get_nvidia_value(nvs->target, ATTR_GPU_TEMP_THRESHOLD);
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
// TODO: calculate gpufreq, memfreq, etc
// can use (val-min)÷(max-min)×100. Perhaps a helper function or macro
case ATTR_AMBIENT_TEMP: // ambienttemp (calculate out of gputempthreshold for consistency)
temp1 = get_nvidia_value(nvs->target, ATTR_AMBIENT_TEMP);
temp2 = get_nvidia_value(nvs->target, ATTR_GPU_TEMP_THRESHOLD);
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
case ATTR_GPU_FREQ: // gpufreq (calculate out of gpufreqmax)
temp1 = get_nvidia_value(nvs->target, ATTR_GPU_FREQ);
temp2 = get_nvidia_string_value(nvs->target, ATTR_PERFMODES_STRING, (char*) "nvclockmax", SEARCH_MAX);
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
case ATTR_MEM_FREQ: // memfreq (calculate out of memfreqmax)
temp1 = get_nvidia_value(nvs->target, ATTR_MEM_FREQ);
temp2 = get_nvidia_string_value(nvs->target, ATTR_PERFMODES_STRING, (char*) "memclockmax", SEARCH_MAX);
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
case ATTR_FREQS_STRING: // mtrfreq (calculate out of memfreqmax)
if (nvs->token == "memTransferRate") {
// Just in case error for silly devs
CRIT_ERR(NULL, NULL, "%s: attribute is 'ATTR_FREQS_STRING' but token is not \"memTransferRate\" (arg: '%s')",
nvs->command, nvs->arg);
return 0;
}
temp1 = get_nvidia_string_value(nvs->target, ATTR_FREQS_STRING, nvs->token, SEARCH_MAX);
temp2 = get_nvidia_string_value(nvs->target, ATTR_PERFMODES_STRING, (char*) "memTransferRatemax", SEARCH_MAX);
if (temp2 > temp1) temp1 = temp2; // extra safe here
value = ((float)temp1 * 100 / (float)temp2) + 0.5;
break;
case ATTR_IMAGE_QUALITY: // imagequality
value = get_nvidia_value(nvs->target, ATTR_IMAGE_QUALITY);
break;
// TODO: throw errors if unsupported args are used
default: // Throw error if unsupported args are used
CRIT_ERR(NULL, NULL, "%s: invalid argument specified: '%s'",
nvs->command, nvs->arg);
}
}

5
src/nvidia.h
View File

@ -32,10 +32,9 @@
#ifndef NVIDIA_CONKY_H
#define NVIDIA_CONKY_H
int scan_nvidia_args (struct text_object *obj, const char *args, unsigned int special_t);
int set_nvidia_type(struct text_object *, const char *);
int set_nvidia_query(struct text_object *, const char *, unsigned int);
void print_nvidia_value(struct text_object *, char *, int);
double get_nvidia_barval(struct text_object *obj);
double get_nvidia_barval(struct text_object *);
void free_nvidia(struct text_object *);
#endif