coreutils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Du feature request - group reporting


From: Daniel Gall
Subject: Re: Du feature request - group reporting
Date: Mon, 5 Feb 2018 19:32:09 -0500

>From c5b3dc83a762fd96c60d7560249a1815386a5efb Mon Sep 17 00:00:00 2001
From: Daniel Gall <address@hidden>
Date: Mon, 5 Feb 2018 14:30:39 -0500
Subject: [PATCH] :Added group reporting to du

---
 NEWS               |   2 +
 doc/coreutils.texi |   5 ++
 src/du.c           | 175 ++++++++++++++++++++++++++++++++++++++++++++++++++---
 3 files changed, 173 insertions(+), 9 deletions(-)

diff --git a/NEWS b/NEWS
index 8a9e09e..97ccc2c 100644
--- a/NEWS
+++ b/NEWS
@@ -56,6 +56,8 @@ GNU coreutils NEWS
 -*- outline -*-

   timeout now supports the --verbose option to diagnose forced termination.

+  du now supports the -g option for group reporting
+
 ** Improvements

   dd now supports iflag=direct with arbitrary sized files on all file systems.
diff --git a/doc/coreutils.texi b/doc/coreutils.texi
index cdde136..d220012 100644
--- a/doc/coreutils.texi
+++ b/doc/coreutils.texi
@@ -11910,6 +11910,11 @@ is at level 0, so @code{du --max-depth=0} is
equivalent to @code{du -s}.
 @c --files0-from=FILE
 @filesZeroFromOption{du,, with the @option{--total} (@option{-c}) option}

+@item -g
+@opindex -g
+@cindex group reporting
+Show group subtotals for each item reported on.
+
 @item -H
 @opindex -H
 Equivalent to @option{--dereference-args} (@option{-D}).
diff --git a/src/du.c b/src/du.c
index ac4489f..a8318ca 100644
--- a/src/du.c
+++ b/src/du.c
@@ -35,6 +35,7 @@
 #include "error.h"
 #include "exclude.h"
 #include "fprintftime.h"
+#include "grp.h"
 #include "human.h"
 #include "mountlist.h"
 #include "quote.h"
@@ -61,6 +62,9 @@ extern bool fts_debug;
 # define FTS_CROSS_CHECK(Fts)
 #endif

+/* If true, display group size info. */
+bool opt_group_sizes = false;
+
 /* A set of dev/ino pairs to help identify files and directories
    whose sizes have already been counted.  */
 static struct di_set *di_files;
@@ -80,7 +84,7 @@ struct duinfo

   /* Number of inodes in directory.  */
   uintmax_t inodes;
-
+  uintmax_t *group_size;
   /* Latest timestamp found.  If tmax.tv_sec == TYPE_MINIMUM (time_t)
      && tmax.tv_nsec < 0, no timestamp has been found.  */
   struct timespec tmax;
@@ -90,28 +94,62 @@ struct duinfo
 static inline void
 duinfo_init (struct duinfo *a)
 {
+  uintmax_t i = 0;
   a->size = 0;
   a->inodes = 0;
   a->tmax.tv_sec = TYPE_MINIMUM (time_t);
   a->tmax.tv_nsec = -1;
+  if (opt_group_sizes)
+    {
+      for (i=0; i<65535; i++)
+        {
+      a->group_size[i] = 0;
+    }
+    }
 }

 /* Set directory data.  */
 static inline void
-duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax)
+duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax,
uintmax_t gid)
 {
+  uintmax_t gid_u = (uintmax_t)gid;
+  uintmax_t gid_s;
   a->size = size;
   a->inodes = 1;
   a->tmax = tmax;
+  if (opt_group_sizes)
+    {
+      if (gid_u > 65534)
+        {
+      gid_s = 65534;
+    }
+      else
+        {
+      gid_s = gid_u;
+    }
+      if (gid_s >=0 && gid_s <=65534)
+        {
+      a->group_size[gid_s] = size;
+    }
+    }
 }

 /* Accumulate directory data.  */
 static inline void
 duinfo_add (struct duinfo *a, struct duinfo const *b)
 {
+  uintmax_t i = 0;
   uintmax_t sum = a->size + b->size;
   a->size = a->size <= sum ? sum : UINTMAX_MAX;
   a->inodes = a->inodes + b->inodes;
+  if (opt_group_sizes)
+    {
+      for (i=0; i<65535; i++)
+        {
+          sum = a->group_size[i] + b->group_size[i];
+          a->group_size[i] = a->group_size[i] <= sum ? sum : UINTMAX_MAX;
+        }
+    }
   if (timespec_cmp (a->tmax, b->tmax) < 0)
     a->tmax = b->tmax;
 }
@@ -226,6 +264,7 @@ static struct option const long_options[] =
   {"exclude", required_argument, NULL, EXCLUDE_OPTION},
   {"exclude-from", required_argument, NULL, 'X'},
   {"files0-from", required_argument, NULL, FILES0_FROM_OPTION},
+  {"group-reporting", no_argument, NULL, 'g'},
   {"human-readable", no_argument, NULL, 'h'},
   {"inodes", no_argument, NULL, INODES_OPTION},
   {"si", no_argument, NULL, HUMAN_SI_OPTION},
@@ -317,6 +356,7 @@ Summarize disk usage of the set of FILEs,
recursively for directories.\n\
       --files0-from=F   summarize disk usage of the\n\
                           NUL-terminated file names specified in file F;\n\
                           if F is -, then read names from standard input\n\
+  -g, --group-reporting also print group subtotals\n\
   -H                    equivalent to --dereference-args (-D)\n\
   -h, --human-readable  print sizes in human readable format (e.g.,
1K 234M 2G)\
 \n\
@@ -411,7 +451,24 @@ print_size (const struct duinfo *pdui, const char *string)
   print_only_size (opt_inodes
                    ? pdui->inodes
                    : pdui->size);
-
+  if (opt_group_sizes)
+    {
+      uintmax_t i=0;
+      struct group *g;
+      for (i=0; i<65535; i++){
+        if (pdui->group_size[i] > 0)
+          {
+            g = getgrgid(i);
+            printf (",");
+            if (g == NULL){
+              printf (" %Ld:", (long long unsigned int)i);
+            }else{
+              printf(" %s:", g->gr_name);
+            }
+            print_only_size(pdui->group_size[i]);
+          }
+      }
+    }
   if (opt_time)
     {
       putchar ('\t');
@@ -506,6 +563,12 @@ process_file (FTS *fts, FTSENT *ent)
   const struct stat *sb = ent->fts_statp;
   int info = ent->fts_info;

+  if(opt_group_sizes)
+    {
+      dui.group_size = xcalloc (65536, sizeof (uintmax_t));
+      dui_to_print.group_size = xcalloc (65536, sizeof (uintmax_t));
+    }
+
   if (info == FTS_DNR)
     {
       /* An error occurred, but the size is known, so count it.  */
@@ -530,7 +593,18 @@ process_file (FTS *fts, FTSENT *ent)
           if (info == FTS_NS || info == FTS_SLNONE)
             {
               error (0, ent->fts_errno, _("cannot access %s"), quoteaf (file));
-              return false;
+              if(opt_group_sizes)
+                {
+                  if (dui.group_size != NULL)
+                    {
+                      free(dui.group_size);
+                    }
+                  if (dui_to_print.group_size != NULL)
+                    {
+                      free(dui_to_print.group_size);
+                    }
+                }
+          return false;
             }

           /* The --one-file-system (-x) option cannot exclude anything
@@ -558,13 +632,34 @@ process_file (FTS *fts, FTSENT *ent)
               FTSENT const *e = fts_read (fts);
               assert (e == ent);
             }
-
+          if(opt_group_sizes)
+            {
+              if (dui.group_size != NULL)
+                {
+                  free(dui.group_size);
+                }
+              if (dui_to_print.group_size != NULL)
+                {
+                  free(dui_to_print.group_size);
+                }
+            }
           return true;
         }

       switch (info)
         {
         case FTS_D:
+      if(opt_group_sizes)
+            {
+              if (dui.group_size != NULL)
+                {
+                  free(dui.group_size);
+                }
+              if (dui_to_print.group_size != NULL)
+                {
+                  free(dui_to_print.group_size);
+                }
+            }
           return true;

         case FTS_ERR:
@@ -574,6 +669,17 @@ process_file (FTS *fts, FTSENT *ent)
           break;

         case FTS_DC:
+      if(opt_group_sizes)
+            {
+              if (dui.group_size != NULL)
+                {
+                  free(dui.group_size);
+                }
+              if (dui_to_print.group_size != NULL)
+                {
+                  free(dui_to_print.group_size);
+                }
+            }
           /* If not following symlinks and not a (bind) mount point.  */
           if (cycle_warning_required (fts, ent)
               && ! mount_point_in_fts_cycle (ent))
@@ -591,15 +697,40 @@ process_file (FTS *fts, FTSENT *ent)
                : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
               (time_type == time_mtime ? get_stat_mtime (sb)
                : time_type == time_atime ? get_stat_atime (sb)
-               : get_stat_ctime (sb)));
+               : get_stat_ctime (sb)),
+           sb->st_gid);

   level = ent->fts_level;
-  dui_to_print = dui;
+
+  if (opt_group_sizes)
+    {
+      duinfo_set (&dui_to_print,
+              (apparent_size
+               ? MAX (0, sb->st_size)
+               : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
+              (time_type == time_mtime ? get_stat_mtime (sb)
+               : time_type == time_atime ? get_stat_atime (sb)
+               : get_stat_ctime (sb)),
+               sb->st_gid);
+    }
+  else
+    {
+      dui_to_print = dui;
+    }

   if (n_alloc == 0)
     {
+      size_t i;
       n_alloc = level + 10;
       dulvl = xcalloc (n_alloc, sizeof *dulvl);
+      if(opt_group_sizes)
+        {
+          for (i=0; i<n_alloc; i++)
+            {
+              dulvl[i].ent.group_size = xcalloc (65536, sizeof (uintmax_t));
+              dulvl[i].subdir.group_size = xcalloc (65536, sizeof (uintmax_t));
+            }
+        }
     }
   else
     {
@@ -613,14 +744,23 @@ process_file (FTS *fts, FTSENT *ent)
              Clear the accumulators for *all* levels between prev_level
              and the current one.  The depth may change dramatically,
              e.g., from 1 to 10.  */
+          size_t i;

           if (n_alloc <= level)
             {
               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
+              if(opt_group_sizes)
+                {
+                  for (i=n_alloc; i<level*2; i++)
+                    {
+                      dulvl[i].ent.group_size = xcalloc (65536,
sizeof (uintmax_t));
+                      dulvl[i].subdir.group_size = xcalloc (65536,
sizeof (uintmax_t));
+                    }
+                }
               n_alloc = level * 2;
             }

-          for (size_t i = prev_level + 1; i <= level; i++)
+          for (i = prev_level + 1; i <= level; i++)
             {
               duinfo_init (&dulvl[i].ent);
               duinfo_init (&dulvl[i].subdir);
@@ -666,6 +806,18 @@ process_file (FTS *fts, FTSENT *ent)
         print_size (&dui_to_print, file);
     }

+  if(opt_group_sizes)
+    {
+      if (dui.group_size != NULL)
+        {
+          free(dui.group_size);
+        }
+      if (dui_to_print.group_size != NULL)
+        {
+          free(dui_to_print.group_size);
+        }
+    }
+
   return ok;
 }

@@ -755,7 +907,7 @@ main (int argc, char **argv)
   while (true)
     {
       int oi = -1;
-      int c = getopt_long (argc, argv, "0abd:chHklmst:xB:DLPSX:",
+      int c = getopt_long (argc, argv, "0abgd:chHklmst:xB:DLPSX:",
                            long_options, &oi);
       if (c == -1)
         break;
@@ -800,6 +952,11 @@ main (int argc, char **argv)
           output_block_size = 1;
           break;

+        case 'g':
+          tot_dui.group_size = xcalloc (65536, sizeof (uintmax_t));
+          opt_group_sizes = true;
+          break;
+
         case 'k':
           human_output_opts = 0;
           output_block_size = 1024;
-- 
2.10.2

On Mon, Feb 5, 2018 at 10:14 AM, Daniel Gall <address@hidden> wrote:
> Thanks. Will resubmit without it.
>
> Sent from my iPhone
>
>> On Feb 5, 2018, at 9:20 AM, Eric Blake <address@hidden> wrote:
>>
>>> On 02/02/2018 09:27 PM, Daniel Gall wrote:
>>> Sorry for the delay; life intervened.  In addition to the feature add,
>>> I found a place where du was calling xcalloc and did not check that
>>> the returned pointer was not NULL.  I added a check.
>>
>> Wrong; the contract of xcalloc() is that it CAN'T return NULL (it will
>> have abort()ed instead, if you are low on memory).
>>
>>>> From 544c581654cd0dcfb363215801245a7c2dd3fcd3 Mon Sep 17 00:00:00 2001
>>> From: Daniel Gall <address@hidden>
>>> Date: Fri, 2 Feb 2018 17:18:44 -0500
>>> Subject: [PATCH] added du group reporting feature and fixed a bug where du
>>> allocated memory and did not check that the target pointer was not NULL 
>>> after
>>> the allocation call.
>>
>> Too long of a subject line; the best commits include a one-line summary
>> (~60 characters or less), then a blank line, then more details.
>>
>>> +++ b/NEWS
>>> @@ -32,6 +32,8 @@ GNU coreutils NEWS
>>> -*- outline -*-
>>>   df no longer hangs when given a fifo argument.
>>>   [bug introduced in coreutils-7.3]
>>>
>>> +  du no longer allocates memory without checking whether the
>>> allocation call succeeded.
>>
>> This change is not needed, as it was not a bug in the first place.
>>
>> --
>> Eric Blake, Principal Software Engineer
>> Red Hat, Inc.           +1-919-301-3266
>> Virtualization:  qemu.org | libvirt.org
>>

Attachment: 0001-Added-group-reporting-to-du.patch
Description: Text Data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]