rb_namespace add logf

add logging to the namespace file to allow debug of root redirect

Change-Id: I6032aea880998c05dacf3d0d2e0d222205b9376e

firmware/common/file.c

@@ -31,6 +31,15 @@
 #include "rb_namespace.h"
 #include "string-extra.h"
 
+/* Define LOGF_ENABLE to enable logf output in this file */
+//#define LOGF_ENABLE
+#ifdef LOGF_ENABLE
+#include "logf.h"
+#undef DEBUGF
+#define DEBUGF logf
+#endif
+
+
 /**
  * These functions provide a roughly POSIX-compatible file I/O API.
  */
@@ -600,8 +609,10 @@ static inline ssize_t readwrite_partial(struct filestr_desc *file,
 static ssize_t readwrite(struct filestr_desc *file, void *buf, size_t nbyte,
                          bool write)
 {
+#ifndef LOGF_ENABLE /* wipes out log before you can save it */
     DEBUGF("readwrite(%p,%lx,%lu,%s)\n",
            file, (long)buf, (unsigned long)nbyte, write ? "write" : "read");
+#endif
 
     const file_size_t size = *file->sizep;
     file_size_t filerem;
@@ -766,8 +777,9 @@ file_error:;
         /* error or not, update the file offset and size if anything was
            transferred */
         file->offset += done;
+#ifndef LOGF_ENABLE /* wipes out log before you can save it */
         DEBUGF("file offset: %ld\n", file->offset);
-
+#endif
         /* adjust file size to length written */
         if (write && file->offset > size)
             *file->sizep = file->offset;
@@ -901,8 +913,9 @@ file_error:
 /* move the read/write file offset */
 off_t lseek(int fildes, off_t offset, int whence)
 {
+#ifndef LOGF_ENABLE /* wipes out log before you can save it */
     DEBUGF("lseek(fd=%d,ofs=%ld,wh=%d)\n", fildes, (long)offset, whence);
-
+#endif
     struct filestr_desc * const file = GET_FILESTR(READER, fildes);
     if (!file)
         FILE_ERROR_RETURN(ERRNO, -1);

firmware/common/file_internal.c

@@ -30,6 +30,14 @@
 #include "string-extra.h"
 #include "rbunicode.h"
 
+/* Define LOGF_ENABLE to enable logf output in this file */
+//#define LOGF_ENABLE
+#ifdef LOGF_ENABLE
+#include "logf.h"
+#undef DEBUGF
+#define DEBUGF logf
+#endif
+
 /** Internal common filesystem service functions  **/
 
 /* for internal functions' scanning use to save quite a bit of stack space -

firmware/common/rb_namespace.c

@@ -24,6 +24,15 @@
 #include "rb_namespace.h"
 #include "file_internal.h"
 
+/* Define LOGF_ENABLE to enable logf output in this file */
+//#define LOGF_ENABLE
+#include "logf.h"
+
+#if !defined(HAVE_MULTIVOLUME) && defined(LOGF_ENABLE)
+    int volume = 0;
+#endif
+
+
 #define ROOT_CONTENTS_INDEX  (NUM_VOLUMES)
 #define NUM_ROOT_ITEMS   (NUM_VOLUMES+1)
 
@@ -53,12 +62,14 @@ static void get_mount_point_entry(IF_MV(int volume,) struct DIRENT *entry)
     entry->info.wrtdate = 0;
     entry->info.wrttime = 0;
 #endif /* is dirinfo_native */
+    logf("%s: vol:%d, %s", __func__, volume, entry->d_name);
 }
 
 /* unmount the directory that enumerates into the root namespace */
 static void unmount_item(int item)
 {
     unsigned int state = get_root_item_state(item);
+    logf("%s: state: %u", __func__, state);
     if (!state)
         return;
 
@@ -83,13 +94,18 @@ int root_mount_path(const char *path, unsigned int flags)
         return -ENOENT;
 #else
     if (!path_is_absolute(path))
+    {
+        logf("Path not absolute %s", path);
         return -ENOENT;
+    }
 #endif /* HAVE_MULTIVOLUME */
 
     bool contents = flags & NSITEM_CONTENTS;
     int item = contents ? ROOT_CONTENTS_INDEX : IF_MV_VOL(volume);
     unsigned int state = get_root_item_state(item);
 
+    logf("%s: item:%d, st:%u, %s", __func__, item, state, path);
+
     if (state)
         return -EBUSY;
 
@@ -122,6 +138,7 @@ int root_mount_path(const char *path, unsigned int flags)
 /* inform root that an entire volume is being unmounted */
 void root_unmount_volume(IF_MV_NONVOID(int volume))
 {
+    logf("%s: vol: %d", __func__, volume);
     FOR_EACH_VOLUME(volume, item)
     {
     #ifdef HAVE_MULTIVOLUME
@@ -143,6 +160,7 @@ void root_unmount_volume(IF_MV_NONVOID(int volume))
 /* parse the root part of a path */
 int ns_parse_root(const char *path, const char **pathp, uint16_t *lenp)
 {
+    logf("%s: path: %s", __func__, path);
     int volume = ROOT_VOLUME;
 
 #ifdef HAVE_MULTIVOLUME
@@ -151,19 +169,35 @@ int ns_parse_root(const char *path, const char **pathp, uint16_t *lenp)
     const char *p;
     volume = path_strip_volume(path, &p, false);
     if (volume != ROOT_VOLUME && !CHECK_VOL(volume))
+    {
+        logf("vol: %d is not root", volume);
         return -ENOENT;
+    }
 #endif /* HAVE_MULTIVOLUME */
 
     /* set name to start at last leading separator; name of root will
      * be returned as "/", volume specifiers as "/<fooN>" */
     *pathp = GOBBLE_PATH_SEPCH(path) - 1;
     *lenp  = IF_MV( volume < NUM_VOLUMES ? p - *pathp : ) 1;
-
+#ifdef LOGF_ENABLE
+    if (volume == INT_MAX)
+        logf("vol: ROOT(%d) %s", volume, *pathp);
+    else
+        logf("vol: %d %s", volume, *pathp);
+#endif
 #ifdef HAVE_MULTIVOLUME
     if (*lenp > MAX_COMPNAME+1)
+    {
+        logf("%s: path too long %s", __func__, path);
         return -ENAMETOOLONG;
+    }
+#endif
+#ifdef LOGF_ENABLE
+    if (volume == INT_MAX)
+        logf("%s: vol: ROOT(%d) path: %s", __func__, volume, path);
+    else
+        logf("%s: vol: %d path: %s", __func__, volume, path);
 #endif
-
     return volume;
 }
 
@@ -183,7 +217,7 @@ int ns_open_root(IF_MV(int volume,) unsigned int *callflagsp,
 
     int item = sysroot ? ROOT_CONTENTS_INDEX : IF_MV_VOL(volume);
     unsigned int state = get_root_item_state(item);
-
+    logf("%s: Vol:%d St:%d", __func__, item, state);
     if (sysroot)
     {
         *attrp = ATTR_SYSTEM_ROOT;
@@ -191,8 +225,11 @@ int ns_open_root(IF_MV(int volume,) unsigned int *callflagsp,
         if (state)
             *infop = root_bindp->info;
         else
+        {
+            logf("%s: SysRoot Vol:%d St:%d NOT mounted", __func__, item, state);
             *callflagsp = callflags | FF_NOFS;  /* contents not mounted */
         }
+    }
     else
     {
         *attrp = ATTR_MOUNT_POINT;
@@ -201,7 +238,10 @@ int ns_open_root(IF_MV(int volume,) unsigned int *callflagsp,
             return -ENOENT; /* regular open requires having been mounted */
 #if CONFIG_PLATFORM & PLATFORM_NATIVE
         if (fat_open_rootdir(IF_MV(volume,) &infop->fatfile) < 0)
+        {
+            logf("%s: DevPath Vol:%d St:%d NOT mounted", __func__, item, state);
             return -ENOENT; /* not mounted */
+        }
 #endif
         get_rootinfo_internal(infop);
     }
@@ -216,6 +256,7 @@ int root_readdir_dirent(struct filestr_base *stream,
     int rc = 0;
 
     int item = scanp->item;
+    logf("%s: item: %d", __func__, item);
 
     /* skip any not-mounted or hidden items */
     unsigned int state;
@@ -226,7 +267,10 @@ int root_readdir_dirent(struct filestr_base *stream,
 
         state = get_root_item_state(item);
         if ((state & (NSITEM_MOUNTED|NSITEM_HIDDEN)) == NSITEM_MOUNTED)
+        {
+            logf("Found mounted item: %d %s", item, entry->d_name);
             break;
+        }
 
         item++;
     }
@@ -238,13 +282,17 @@ int root_readdir_dirent(struct filestr_base *stream,
             FILE_ERROR(ERRNO, rc * 10 - 1);
 
         if (rc == 0)
+        {
+            logf("Found root item: %d %s", item, entry->d_name);
             item++;
         }
+    }
     else
     {
         get_mount_point_entry(IF_MV(item,) entry);
         item++;
         rc = 1;
+        logf("Found mp item:%d %s", item,  entry->d_name);
     }
 
     scanp->item = item;
@@ -255,6 +303,7 @@ file_eod:
         empty_dirent(entry);
 #endif
 file_error:
+    logf("%s: status: %d", __func__, rc);
     return rc;
 }
 
@@ -262,6 +311,7 @@ file_error:
 int ns_open_stream(const char *path, unsigned int callflags,
                    struct filestr_base *stream, struct ns_scan_info *scanp)
 {
+    logf("%s: path: %s", __func__, path);
     /* stream still needs synchronization even if we don't have a stream */
     static struct mutex no_contents_mtx SHAREDBSS_ATTR;