minor trinity patch

Hi,
I'm not sure if you would like to change this....
In devices.c you scan for the device name in /proc/devices using the %as string.
On bionic this doesn't seem to work and therefore no devices are found and because of this no ioctl group ever matches.

(quoted diff) +6 -5

        size_t n = 0;
        int block, major;
        void *new;

@@ -33,30 +35,31 @@ static void parse_proc_devices(void)
        block = 0;
 
        while (getline(&line, &n, fp) >= 0) {
+               memset(name, 0, sizeof(name));
                if (strcmp("Block devices:\n", line) == 0)
                        block = 1;
-               else if (sscanf(line, "%d %as", &major, &name) == 2) {
+               else if (strcmp("Character devices:\n", line) == 0)
+                       block = 0;
+               else if (sscanf(line, "%d %s", &major, &name) == 2) {
                        if (block) {
                                new = realloc(block_devs, (bldevs+1)*sizeof(*block_devs));
                                if (!new) {
-                                       free(name);
                                        continue;
                                }
                                block_devs = new;
                                block_devs[bldevs].major = major;
                                block_devs[bldevs].minor = 0;
-                               block_devs[bldevs].name = name;
+                               block_devs[bldevs].name = strdup(name);
                                bldevs++;
                        } else {
                                new = realloc(char_devs, (chrdevs+1)*sizeof(*char_devs));
                                if (!new) {
-                                       free(name);
                                        continue;
                                }
                                char_devs = new;
                                char_devs[chrdevs].major = major;
                                char_devs[chrdevs].minor = 0;
-                               char_devs[chrdevs].name = name;
+                               char_devs[chrdevs].name = strdup(name);
                                chrdevs++;
                        }
                }

I also added the additional strcmp just to make sure that the code works also if the order of character devices/block devices in /proc/devices changes at some point.

Cheers
Nico

2013/5/22 Nico Golde <[email protected]>:

Review static size_t bldevs, chrdevs, miscdevs; · line 26

Quoted 8 lines

I changed the code the following way:

(quoted diff) +2 -1

@@ -21,7 +22,8 @@ static size_t bldevs, chrdevs, miscdevs;
 static void parse_proc_devices(void)
 {
        FILE *fp;
-       char *name, *line = NULL;
+       char *line = NULL;
+       char name[32];

It would be IMO preferable to not hardcode the max length.

Hi,

Quoted 11 lines · Tommi Rantala

2013/5/22 Nico Golde <[email protected]>:

Review static size_t bldevs, chrdevs, miscdevs; · line 26

Quoted 8 lines

I changed the code the following way:

(quoted diff) +2 -1

@@ -21,7 +22,8 @@ static size_t bldevs, chrdevs, miscdevs;
 static void parse_proc_devices(void)
 {
        FILE *fp;
-       char *name, *line = NULL;
+       char *line = NULL;
+       char name[32];

It would be IMO preferable to not hardcode the max length.

I agree, this is actually left over from quickly changing this.
Are you aware of a proper maximum defined in the kernel headers or do we want to allocate this completely dynamically?

Cheers
Nico
--
To unsubscribe from this list: send the line "unsubscribe trinity" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

2013/5/24 Nico Golde <[email protected]>:

Quoted 13 lines · Tommi Rantala

Quoted 11 lines

2013/5/22 Nico Golde <[email protected]>:

Review static size_t bldevs, chrdevs, miscdevs; · line 26

Quoted 8 lines

I changed the code the following way:

(quoted diff) +2 -1

@@ -21,7 +22,8 @@ static size_t bldevs, chrdevs, miscdevs;
 static void parse_proc_devices(void)
 {
        FILE *fp;
-       char *name, *line = NULL;
+       char *line = NULL;
+       char name[32];

It would be IMO preferable to not hardcode the max length.

Quoted 3 lines

I agree, this is actually left over from quickly changing this.
Are you aware of a proper maximum defined in the kernel headers or do we want to allocate this completely dynamically?

dynamic please!

Sorry I had no time to look into this for a while. Patch attached.