fix(linux): improved container support

We make use of the /proc/root of the procfs to improve the chances
that we can infer the required information, such as the location
of the runtime structure and the interpreter version from binaries
pointed at by the actual process root. This is relevant for
processes running inside containers, since their FS root does not
coincide with the host root.

Author: P403n1x87

ChangeLog

@@ -1,5 +1,7 @@
 2024-xx-xx v3.6.1
 
+  Improve support for Python processes running in containers.
+
   Bugfix: fixed a bug with the MOJO binary format that caused the line end
   position to wrongly be set to a non-zero value for CPython < 3.11, where line
   end information is not actually available.

src/linux/common.h

@@ -26,9 +26,11 @@
 #include <pthread.h>
 #include <stdint.h>
 #include <stdio.h>
+#include <string.h>
 #include <sys/ptrace.h>
 
 #include "../error.h"
+#include "../hints.h"
 #include "../stats.h"
 
 
@@ -109,3 +111,24 @@ _procfs(pid_t pid, char * file) {
 
   return fp;
 }
+
+
+// ----------------------------------------------------------------------------
+static inline char *
+proc_root(pid_t pid, char * file) {
+  if (file[0] != '/') {
+    log_e("File path is not absolute");
+    return NULL;
+  }
+
+  char * proc_root = calloc(1, strlen(file) + 24);
+  if (!isvalid(proc_root))
+    return NULL;
+
+  if (sprintf(proc_root, "/proc/%d/root%s", pid, file) < 0) {
+    free(proc_root);
+    return NULL;
+  }
+  
+  return proc_root;
+}

src/linux/py_proc.h

@@ -455,16 +455,16 @@ _py_proc__parse_maps_file(py_proc_t * self) {
     // The first memory map of the executable
     if (!isvalid(pd->maps[MAP_BIN].path) && strcmp(pd->exe_path, pathname) == 0) {
       map = &(pd->maps[MAP_BIN]);
-      map->path = strndup(pathname, strlen(pathname));
+      map->path = proc_root(self->pid, pathname);
       if (!isvalid(map->path)) {
-        log_ie("Cannot duplicate path name");
+        log_e("Cannot get proc root path for %s", pathname);
         set_error(EPROC);
         FAIL;
       }
-      map->file_size = _file_size(pathname);
+      map->file_size = _file_size(map->path);
       map->base = (void *) lower;
       map->size = upper - lower;
-      map->has_symbols = success(_py_proc__analyze_elf(self, pathname, (void *) lower));
+      map->has_symbols = success(_py_proc__analyze_elf(self, map->path, (void *) lower));
       if (map->has_symbols) {
         map->bss_base = self->map.bss.base;
         map->bss_size = self->map.bss.size;
@@ -479,13 +479,13 @@ _py_proc__parse_maps_file(py_proc_t * self) {
       int has_symbols = success(_py_proc__analyze_elf(self, pathname, (void *) lower));
       if (has_symbols) {
         map = &(pd->maps[MAP_LIBSYM]);
-        map->path = strndup(pathname, strlen(pathname));
+        map->path = proc_root(self->pid, pathname);
         if (!isvalid(map->path)) {
-          log_ie("Cannot duplicate path name");
+          log_e("Cannot get proc root path for %s", pathname);
           set_error(EPROC);
           FAIL;
         }
-        map->file_size = _file_size(pathname);
+        map->file_size = _file_size(map->path);
         map->base = (void *) lower;
         map->size = upper - lower;
         map->has_symbols = TRUE;
@@ -503,13 +503,13 @@ _py_proc__parse_maps_file(py_proc_t * self) {
           unsigned int v;
           if (sscanf(needle, "libpython%u.%u", &v, &v) == 2) {
             map = &(pd->maps[MAP_LIBNEEDLE]);
-            map->path = needle_path = strndup(pathname, strlen(pathname));
+            map->path = needle_path = proc_root(self->pid, pathname);
             if (!isvalid(map->path)) {
-              log_ie("Cannot duplicate path name");
+              log_e("Cannot get proc root path for %s", pathname);
               set_error(EPROC);
               FAIL;
             }
-            map->file_size = _file_size(pathname);
+            map->file_size = _file_size(map->path);
             map->base = (void *) lower;
             map->size = upper - lower;
             map->has_symbols = FALSE;

src/py_proc.c

@@ -110,17 +110,14 @@ _get_version_from_executable(char * binary, int * major, int * minor, int * patc
   #endif
 
   fp = _popen(cmd, "r");
-  if (!isvalid(fp)) {
-    set_error(EPROC);
+  if (!isvalid(fp))
     FAIL;
-  }
 
   while (fgets(version, sizeof(version) - 1, fp) != NULL) {
     if (sscanf(version, "Python %d.%d.%d", major, minor, patch) == 3)
       SUCCESS;
   }
 
-  set_error(EPROC);
   FAIL;
 } /* _get_version_from_executable */
 
@@ -129,37 +126,34 @@ _get_version_from_filename(char * filename, const char * needle, int * major, in
   #if defined PL_LINUX                                               /* LINUX */
   char         * base       = filename;
   char         * end        = base + strlen(base);
+  size_t         needle_len = strlen(needle);
 
   while (base < end) {
     base = strstr(base, needle);
     if (!isvalid(base)) {
       break;
     }
-    if (sscanf(base + strlen(needle), "%u.%u", major, minor) == 2) {
+    base += needle_len;
+    if (sscanf(base, "%u.%u", major, minor) == 2) {
       SUCCESS;
     }
   }
 
   #elif defined PL_WIN                                                 /* WIN */
   // Assume the library path is of the form *.python3[0-9]+[.]dll
   int n = strlen(filename);
-  if (n < 10) {
-    set_error(EPROC);
+  if (n < 10)
     FAIL;
-  }
 
   char * p = filename + n - 1;
   while (*(p--) != 'n' && p > filename);
   p++;
   *major = *(p++) - '0';
-  if (*major != 3) {
-    set_error(EPROC);
+  if (*major != 3)
     FAIL;
-  }
 
-  if (sscanf(p,"%d.dll", minor) == 1) {
+  if (sscanf(p,"%d.dll", minor) == 1)
     SUCCESS;
-  }
 
   #elif defined PL_MACOS                                               /* MAC */
   char * ver_needle = strstr(filename, "3.");
@@ -169,7 +163,6 @@ _get_version_from_filename(char * filename, const char * needle, int * major, in
 
   #endif
 
-  set_error(EPROC);
   FAIL;
 } /* _get_version_from_filename */
 
@@ -648,6 +641,9 @@ _py_proc__run(py_proc_t * self) {
       TIMER_STOP;
     }
 
+    if (is_fatal(austin_errno))
+      FAIL;
+
   TIMER_END
 
   log_d("_py_proc__init timer loop terminated");