Skip to contentSkip to conversation

fs/ext3: memory leak by calling set_qf_name or clear_qf_name, many times.

Suggestion Chen Gang 2 participants 7 messages Dec 26, 2012 – Jan 14, 2013

lore
Share & tools
lore

Chen Gang

author
#0 lore raw

Hello Theodore Ts'o

in fs/ext3/supper.c
for function set_qf_name:

    sbi->s_qf_names[qtype] may already have owned a memory (line 919..925)
    we set sbi->s_qf_names[qtype] = qname directly without checking (line 926)

for function clear_qf_name:

    we set sbi->s_qf_names[qtype] = NULL (line 942..952)

for function parse_options:
we can call set_qf_name or clear_qf_name with USR or GRP many times.
we find parameters not mind whether they are repeated. (line 975..985) so we may call set_qf_name or clear_qf_name several times.
also may first call set_qf_name, then call clear_qf_name.

in this situation, we will get memory leak.

please help check this suggestion whether valid (I find it by code review).

regards.

gchen.

901 static int set_qf_name(struct super_block *sb, int qtype, substring_t *args) 902 {
903 struct ext3_sb_info *sbi = EXT3_SB(sb);
904 char *qname;
905
906 if (sb_any_quota_loaded(sb) &&
907 !sbi->s_qf_names[qtype]) {
908 ext3_msg(sb, KERN_ERR,
909 "Cannot change journaled "
910 "quota options when quota turned on");
911 return 0;

 912         }
 913         qname = match_strdup(args);
 914         if (!qname) {
 915                 ext3_msg(sb, KERN_ERR,
 916                         "Not enough memory for storing quotafile name");
 917                 return 0;
 918         }
 919         if (sbi->s_qf_names[qtype] &&
 920                 strcmp(sbi->s_qf_names[qtype], qname)) {
 921                 ext3_msg(sb, KERN_ERR,
 922                         "%s quota file already specified", QTYPE2NAME(qtype));
 923                 kfree(qname);
 924                 return 0;
 925         }
 926         sbi->s_qf_names[qtype] = qname;
 927         if (strchr(sbi->s_qf_names[qtype], '/')) {
 928                 ext3_msg(sb, KERN_ERR,
 929                         "quotafile must be on filesystem root");
 930                 kfree(sbi->s_qf_names[qtype]);
 931                 sbi->s_qf_names[qtype] = NULL;
 932                 return 0;
 933         }
 934         set_opt(sbi->s_mount_opt, QUOTA);
 935         return 1;
 936 }
 937
 938 static int clear_qf_name(struct super_block *sb, int qtype) {
 939 
 940         struct ext3_sb_info *sbi = EXT3_SB(sb);
 941 
 942         if (sb_any_quota_loaded(sb) &&
 943                 sbi->s_qf_names[qtype]) {
 944                 ext3_msg(sb, KERN_ERR, "Cannot change journaled quota options"
 945                         " when quota turned on");
 946                 return 0;
 947         }
 948         /*
 949          * The space will be released later when all options are confirmed
 950          * to be correct
 951          */
 952         sbi->s_qf_names[qtype] = NULL;
 953         return 1;
 954 }
 955 #endif
 956 
 957 static int parse_options (char *options, struct super_block *sb,
 958                           unsigned int *inum, unsigned long *journal_devnum,
 959                           ext3_fsblk_t *n_blocks_count, int is_remount)
 960 {
 961         struct ext3_sb_info *sbi = EXT3_SB(sb);
 962         char * p;
 963         substring_t args[MAX_OPT_ARGS];
 964         int data_opt = 0;
 965         int option;
 966         kuid_t uid;
 967         kgid_t gid;
 968 #ifdef CONFIG_QUOTA
 969         int qfmt;
 970 #endif
 971 
 972         if (!options)
 973                 return 1;
 974 
 975         while ((p = strsep (&options, ",")) != NULL) {
 976                 int token;
 977                 if (!*p)
 978                         continue;
 979                 /*
 980                  * Initialize args struct so we know whether arg was
 981                  * found; some options take optional arguments.
 982                  */
 983                 args[0].to = args[0].from = NULL;
 984                 token = match_token(p, tokens, args);
 985                 switch (token) {
 986                 case Opt_bsd_df:
 987                         clear_opt (sbi->s_mount_opt, MINIX_DF);
 988                         break;
 989                 case Opt_minix_df:
 990                         set_opt (sbi->s_mount_opt, MINIX_DF);
 991                         break;
 992                 case Opt_grpid:
 993                         set_opt (sbi->s_mount_opt, GRPID);
 955 #endif
 956 
 957 static int parse_options (char *options, struct super_block *sb,
 958                           unsigned int *inum, unsigned long *journal_devnum,
 959                           ext3_fsblk_t *n_blocks_count, int is_remount)
 960 {
 961         struct ext3_sb_info *sbi = EXT3_SB(sb);
 962         char * p;
 963         substring_t args[MAX_OPT_ARGS];
 964         int data_opt = 0;
 965         int option;
 966         kuid_t uid;
 967         kgid_t gid;
 968 #ifdef CONFIG_QUOTA
 969         int qfmt;
 970 #endif
 971 
 972         if (!options)
 973                 return 1;
 974 
 975         while ((p = strsep (&options, ",")) != NULL) {
 976                 int token;
 977                 if (!*p)
 978                         continue;
 979                 /*
 980                  * Initialize args struct so we know whether arg was
 981                  * found; some options take optional arguments.
 982                  */
 983                 args[0].to = args[0].from = NULL;
 984                 token = match_token(p, tokens, args);
 985                 switch (token) {
 986                 case Opt_bsd_df:

987 clear_opt (sbi->s_mount_opt, MINIX_DF);
988 break;
989 case Opt_minix_df:
990 set_opt (sbi->s_mount_opt, MINIX_DF);
991 break;
992 case Opt_grpid:
993 set_opt (sbi->s_mount_opt, GRPID);
994 break;
995 case Opt_nogrpid:
996 clear_opt (sbi->s_mount_opt, GRPID);
997 break;
998 case Opt_resuid:
999 if (match_int(&args[0], &option))
1000 return 0;
1001 uid = make_kuid(current_user_ns(), option);
1002 if (!uid_valid(uid)) {
1003 ext3_msg(sb, KERN_ERR, "Invalid uid value %d", option);
1004 return 0;
1005

1006                         }
1007                         sbi->s_resuid = uid;
1008                         break;

1009 case Opt_resgid:
1010 if (match_int(&args[0], &option))
1011 return 0;
1012 gid = make_kgid(current_user_ns(), option);
1013 if (!gid_valid(gid)) {
1014 ext3_msg(sb, KERN_ERR, "Invalid gid value %d", option);
1015 return 0;

1016                         }
1017                         sbi->s_resgid = gid;
1018                         break;
1019                 case Opt_sb:
1020                         /* handled by get_sb_block() instead of here */
1021                         /* *sb_block = match_int(&args[0]); */
1022                         break;
1023                 case Opt_err_panic:
1024                         clear_opt (sbi->s_mount_opt, ERRORS_CONT);
1025                         clear_opt (sbi->s_mount_opt, ERRORS_RO);
1026                         set_opt (sbi->s_mount_opt, ERRORS_PANIC);
1027                         break;
1028                 case Opt_err_ro:
1029                         clear_opt (sbi->s_mount_opt, ERRORS_CONT);
1030                         clear_opt (sbi->s_mount_opt, ERRORS_PANIC);
1031                         set_opt (sbi->s_mount_opt, ERRORS_RO);
1032                         break;
1033                 case Opt_err_cont:
1034                         clear_opt (sbi->s_mount_opt, ERRORS_RO);
1035                         clear_opt (sbi->s_mount_opt, ERRORS_PANIC);
1036                         set_opt (sbi->s_mount_opt, ERRORS_CONT);
1037                         break;

1038 case Opt_nouid32:
1039 set_opt (sbi->s_mount_opt, NO_UID32);
1040 break;
1041 case Opt_nocheck:
1042 clear_opt (sbi->s_mount_opt, CHECK);
1043 break;
1044 case Opt_debug:
1045 set_opt (sbi->s_mount_opt, DEBUG);
1046 break;
1047 case Opt_oldalloc:
1048 ext3_msg(sb, KERN_WARNING,
1049 "Ignoring deprecated oldalloc option");
1050 break;
1051 case Opt_orlov:
1052 ext3_msg(sb, KERN_WARNING,
1053 "Ignoring deprecated orlov option");
1054 break;
1055 #ifdef CONFIG_EXT3_FS_XATTR
1056 case Opt_user_xattr:
1057 set_opt (sbi->s_mount_opt, XATTR_USER);
1058 break;
1059 case Opt_nouser_xattr:
1060 clear_opt (sbi->s_mount_opt, XATTR_USER);
1061 break;
1062 #else
1063 case Opt_user_xattr:
1064 case Opt_nouser_xattr:
1065 ext3_msg(sb, KERN_INFO,
1066 "(no)user_xattr options not supported");
1067 break;
1068 #endif
1069 #ifdef CONFIG_EXT3_FS_POSIX_ACL
1070 case Opt_acl:
1071 set_opt(sbi->s_mount_opt, POSIX_ACL);
1072 break;
1073 case Opt_noacl:
1074 clear_opt(sbi->s_mount_opt, POSIX_ACL);
1075 break;
1076 #else
1077 case Opt_acl:
1078 case Opt_noacl:
1079 ext3_msg(sb, KERN_INFO,
1080 "(no)acl options not supported");
1081 break;
1082 #endif
1083 case Opt_reservation:
1084 set_opt(sbi->s_mount_opt, RESERVATION);
1085 break;
1086 case Opt_noreservation:
1087 clear_opt(sbi->s_mount_opt, RESERVATION);
1088 break;
1089 case Opt_journal_update:
1090 /* @@@ FIXME */
1091 /* Eventually we will want to be able to create

1092                            a journal file here.  For now, only allow the
1093                            user to specify an existing inode to be the
1094                            journal file. */

1095 if (is_remount) {
1096 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1097 "journal on remount");
1098 return 0;

1099                         }
1100                         set_opt (sbi->s_mount_opt, UPDATE_JOURNAL);
1101                         break;

1102 case Opt_journal_inum:
1103 if (is_remount) {
1104 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1105 "journal on remount");
1106 return 0;
1107 }
1108 if (match_int(&args[0], &option))
1109 return 0;
1110 *inum = option;
1111 break;
1112 case Opt_journal_dev:
1113 if (is_remount) {
1114 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1115 "journal on remount");
1116 return 0;
1117 }
1118 if (match_int(&args[0], &option))
1119 return 0;
1120 *journal_devnum = option;
1121 break;
1122 case Opt_noload:
1123 set_opt (sbi->s_mount_opt, NOLOAD);
1124 break;
1125 case Opt_commit:
1126 if (match_int(&args[0], &option))
1127 return 0;
1128 if (option < 0)
1129 return 0;
1130 if (option == 0)
1131 option = JBD_DEFAULT_MAX_COMMIT_AGE;
1132 sbi->s_commit_interval = HZ * option;
1133 break;
1134 case Opt_data_journal:
1135 data_opt = EXT3_MOUNT_JOURNAL_DATA;
1136 goto datacheck;
1137 case Opt_data_ordered:
1138 data_opt = EXT3_MOUNT_ORDERED_DATA;
1139 goto datacheck;
1140 case Opt_data_writeback:
1141 data_opt = EXT3_MOUNT_WRITEBACK_DATA;
1142 datacheck:
1143 if (is_remount) {
1144 if (test_opt(sb, DATA_FLAGS) == data_opt)
1145 break;
1146 ext3_msg(sb, KERN_ERR,

1147                                         "error: cannot change "
1148                                         "data mode on remount. The filesystem "
1149                                         "is mounted in data=%s mode and you "
1150                                         "try to remount it in data=%s mode.",
1151                                         data_mode_string(test_opt(sb,

1152 DATA_FLAGS)),
1153 data_mode_string(data_opt));
1154 return 0;
1155 } else {
1156 clear_opt(sbi->s_mount_opt, DATA_FLAGS);
1157 sbi->s_mount_opt |= data_opt;
1158 }
1159 break;
1160 case Opt_data_err_abort:
1161 set_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1162 break;
1163 case Opt_data_err_ignore:
1164 clear_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1165 break;
1166 #ifdef CONFIG_QUOTA
1167 case Opt_usrjquota:
1168 if (!set_qf_name(sb, USRQUOTA, &args[0]))
1169 return 0;
1170 break;
1171 case Opt_grpjquota:
1172 if (!set_qf_name(sb, GRPQUOTA, &args[0]))
1173 return 0;
1174 break;
1175 case Opt_offusrjquota:
1176 if (!clear_qf_name(sb, USRQUOTA))
1177 return 0;
1178 break;
1179 case Opt_offgrpjquota:
1180 if (!clear_qf_name(sb, GRPQUOTA))
1181 return 0;
1182 break;

author
#1 lore raw

Really (just as Theodore Ts'o said), for ext4 is the same thing as ext3 !

so please help check it by ext4 maintainers, or ext3 maintainers

if want to let me to send relative patch, please tell me, I will try.

:-)

thanks.

gchen

Quoted 348 lines · Chen Gang

Hello Theodore Ts'o

in fs/ext3/supper.c
for function set_qf_name:

    sbi->s_qf_names[qtype] may already have owned a memory (line 919..925)
    we set sbi->s_qf_names[qtype] = qname directly without checking (line 926)

for function clear_qf_name:

    we set sbi->s_qf_names[qtype] = NULL (line 942..952)

for function parse_options:
we can call set_qf_name or clear_qf_name with USR or GRP many times.
we find parameters not mind whether they are repeated. (line 975..985) so we may call set_qf_name or clear_qf_name several times.
also may first call set_qf_name, then call clear_qf_name.

in this situation, we will get memory leak.

please help check this suggestion whether valid (I find it by code review).

regards.

gchen.

901 static int set_qf_name(struct super_block *sb, int qtype, substring_t *args) 902 {
903 struct ext3_sb_info *sbi = EXT3_SB(sb);
904 char *qname;
905
906 if (sb_any_quota_loaded(sb) &&
907 !sbi->s_qf_names[qtype]) {
908 ext3_msg(sb, KERN_ERR,
909 "Cannot change journaled "
910 "quota options when quota turned on");
911 return 0;

 912         }
 913         qname = match_strdup(args);
 914         if (!qname) {
 915                 ext3_msg(sb, KERN_ERR,
 916                         "Not enough memory for storing quotafile name");
 917                 return 0;
 918         }
 919         if (sbi->s_qf_names[qtype] &&
 920                 strcmp(sbi->s_qf_names[qtype], qname)) {
 921                 ext3_msg(sb, KERN_ERR,
 922                         "%s quota file already specified", QTYPE2NAME(qtype));
 923                 kfree(qname);
 924                 return 0;
 925         }
 926         sbi->s_qf_names[qtype] = qname;
 927         if (strchr(sbi->s_qf_names[qtype], '/')) {
 928                 ext3_msg(sb, KERN_ERR,
 929                         "quotafile must be on filesystem root");
 930                 kfree(sbi->s_qf_names[qtype]);
 931                 sbi->s_qf_names[qtype] = NULL;
 932                 return 0;
 933         }
 934         set_opt(sbi->s_mount_opt, QUOTA);
 935         return 1;
 936 }
 937
 938 static int clear_qf_name(struct super_block *sb, int qtype) {
 939 
 940         struct ext3_sb_info *sbi = EXT3_SB(sb);
 941 
 942         if (sb_any_quota_loaded(sb) &&
 943                 sbi->s_qf_names[qtype]) {
 944                 ext3_msg(sb, KERN_ERR, "Cannot change journaled quota options"
 945                         " when quota turned on");
 946                 return 0;
 947         }
 948         /*
 949          * The space will be released later when all options are confirmed
 950          * to be correct
 951          */
 952         sbi->s_qf_names[qtype] = NULL;
 953         return 1;
 954 }
 955 #endif
 956 
 957 static int parse_options (char *options, struct super_block *sb,
 958                           unsigned int *inum, unsigned long *journal_devnum,
 959                           ext3_fsblk_t *n_blocks_count, int is_remount)
 960 {
 961         struct ext3_sb_info *sbi = EXT3_SB(sb);
 962         char * p;
 963         substring_t args[MAX_OPT_ARGS];
 964         int data_opt = 0;
 965         int option;
 966         kuid_t uid;
 967         kgid_t gid;
 968 #ifdef CONFIG_QUOTA
 969         int qfmt;
 970 #endif
 971 
 972         if (!options)
 973                 return 1;
 974 
 975         while ((p = strsep (&options, ",")) != NULL) {
 976                 int token;
 977                 if (!*p)
 978                         continue;
 979                 /*
 980                  * Initialize args struct so we know whether arg was
 981                  * found; some options take optional arguments.
 982                  */
 983                 args[0].to = args[0].from = NULL;
 984                 token = match_token(p, tokens, args);
 985                 switch (token) {
 986                 case Opt_bsd_df:
 987                         clear_opt (sbi->s_mount_opt, MINIX_DF);
 988                         break;
 989                 case Opt_minix_df:
 990                         set_opt (sbi->s_mount_opt, MINIX_DF);
 991                         break;
 992                 case Opt_grpid:
 993                         set_opt (sbi->s_mount_opt, GRPID);
 955 #endif
 956 
 957 static int parse_options (char *options, struct super_block *sb,
 958                           unsigned int *inum, unsigned long *journal_devnum,
 959                           ext3_fsblk_t *n_blocks_count, int is_remount)
 960 {
 961         struct ext3_sb_info *sbi = EXT3_SB(sb);
 962         char * p;
 963         substring_t args[MAX_OPT_ARGS];
 964         int data_opt = 0;
 965         int option;
 966         kuid_t uid;
 967         kgid_t gid;
 968 #ifdef CONFIG_QUOTA
 969         int qfmt;
 970 #endif
 971 
 972         if (!options)
 973                 return 1;
 974 
 975         while ((p = strsep (&options, ",")) != NULL) {
 976                 int token;
 977                 if (!*p)
 978                         continue;
 979                 /*
 980                  * Initialize args struct so we know whether arg was
 981                  * found; some options take optional arguments.
 982                  */
 983                 args[0].to = args[0].from = NULL;
 984                 token = match_token(p, tokens, args);
 985                 switch (token) {
 986                 case Opt_bsd_df:

987 clear_opt (sbi->s_mount_opt, MINIX_DF);
988 break;
989 case Opt_minix_df:
990 set_opt (sbi->s_mount_opt, MINIX_DF);
991 break;
992 case Opt_grpid:
993 set_opt (sbi->s_mount_opt, GRPID);
994 break;
995 case Opt_nogrpid:
996 clear_opt (sbi->s_mount_opt, GRPID);
997 break;
998 case Opt_resuid:
999 if (match_int(&args[0], &option))
1000 return 0;
1001 uid = make_kuid(current_user_ns(), option);
1002 if (!uid_valid(uid)) {
1003 ext3_msg(sb, KERN_ERR, "Invalid uid value %d", option);
1004 return 0;
1005

1006                         }
1007                         sbi->s_resuid = uid;
1008                         break;

1009 case Opt_resgid:
1010 if (match_int(&args[0], &option))
1011 return 0;
1012 gid = make_kgid(current_user_ns(), option);
1013 if (!gid_valid(gid)) {
1014 ext3_msg(sb, KERN_ERR, "Invalid gid value %d", option);
1015 return 0;

1016                         }
1017                         sbi->s_resgid = gid;
1018                         break;
1019                 case Opt_sb:
1020                         /* handled by get_sb_block() instead of here */
1021                         /* *sb_block = match_int(&args[0]); */
1022                         break;
1023                 case Opt_err_panic:
1024                         clear_opt (sbi->s_mount_opt, ERRORS_CONT);
1025                         clear_opt (sbi->s_mount_opt, ERRORS_RO);
1026                         set_opt (sbi->s_mount_opt, ERRORS_PANIC);
1027                         break;
1028                 case Opt_err_ro:
1029                         clear_opt (sbi->s_mount_opt, ERRORS_CONT);
1030                         clear_opt (sbi->s_mount_opt, ERRORS_PANIC);
1031                         set_opt (sbi->s_mount_opt, ERRORS_RO);
1032                         break;
1033                 case Opt_err_cont:
1034                         clear_opt (sbi->s_mount_opt, ERRORS_RO);
1035                         clear_opt (sbi->s_mount_opt, ERRORS_PANIC);
1036                         set_opt (sbi->s_mount_opt, ERRORS_CONT);
1037                         break;

1038 case Opt_nouid32:
1039 set_opt (sbi->s_mount_opt, NO_UID32);
1040 break;
1041 case Opt_nocheck:
1042 clear_opt (sbi->s_mount_opt, CHECK);
1043 break;
1044 case Opt_debug:
1045 set_opt (sbi->s_mount_opt, DEBUG);
1046 break;
1047 case Opt_oldalloc:
1048 ext3_msg(sb, KERN_WARNING,
1049 "Ignoring deprecated oldalloc option");
1050 break;
1051 case Opt_orlov:
1052 ext3_msg(sb, KERN_WARNING,
1053 "Ignoring deprecated orlov option");
1054 break;
1055 #ifdef CONFIG_EXT3_FS_XATTR
1056 case Opt_user_xattr:
1057 set_opt (sbi->s_mount_opt, XATTR_USER);
1058 break;
1059 case Opt_nouser_xattr:
1060 clear_opt (sbi->s_mount_opt, XATTR_USER);
1061 break;
1062 #else
1063 case Opt_user_xattr:
1064 case Opt_nouser_xattr:
1065 ext3_msg(sb, KERN_INFO,
1066 "(no)user_xattr options not supported");
1067 break;
1068 #endif
1069 #ifdef CONFIG_EXT3_FS_POSIX_ACL
1070 case Opt_acl:
1071 set_opt(sbi->s_mount_opt, POSIX_ACL);
1072 break;
1073 case Opt_noacl:
1074 clear_opt(sbi->s_mount_opt, POSIX_ACL);
1075 break;
1076 #else
1077 case Opt_acl:
1078 case Opt_noacl:
1079 ext3_msg(sb, KERN_INFO,
1080 "(no)acl options not supported");
1081 break;
1082 #endif
1083 case Opt_reservation:
1084 set_opt(sbi->s_mount_opt, RESERVATION);
1085 break;
1086 case Opt_noreservation:
1087 clear_opt(sbi->s_mount_opt, RESERVATION);
1088 break;
1089 case Opt_journal_update:
1090 /* @@@ FIXME */
1091 /* Eventually we will want to be able to create

1092                            a journal file here.  For now, only allow the
1093                            user to specify an existing inode to be the
1094                            journal file. */

1095 if (is_remount) {
1096 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1097 "journal on remount");
1098 return 0;

1099                         }
1100                         set_opt (sbi->s_mount_opt, UPDATE_JOURNAL);
1101                         break;

1102 case Opt_journal_inum:
1103 if (is_remount) {
1104 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1105 "journal on remount");
1106 return 0;
1107 }
1108 if (match_int(&args[0], &option))
1109 return 0;
1110 *inum = option;
1111 break;
1112 case Opt_journal_dev:
1113 if (is_remount) {
1114 ext3_msg(sb, KERN_ERR, "error: cannot specify "
1115 "journal on remount");
1116 return 0;
1117 }
1118 if (match_int(&args[0], &option))
1119 return 0;
1120 *journal_devnum = option;
1121 break;
1122 case Opt_noload:
1123 set_opt (sbi->s_mount_opt, NOLOAD);
1124 break;
1125 case Opt_commit:
1126 if (match_int(&args[0], &option))
1127 return 0;
1128 if (option < 0)
1129 return 0;
1130 if (option == 0)
1131 option = JBD_DEFAULT_MAX_COMMIT_AGE;
1132 sbi->s_commit_interval = HZ * option;
1133 break;
1134 case Opt_data_journal:
1135 data_opt = EXT3_MOUNT_JOURNAL_DATA;
1136 goto datacheck;
1137 case Opt_data_ordered:
1138 data_opt = EXT3_MOUNT_ORDERED_DATA;
1139 goto datacheck;
1140 case Opt_data_writeback:
1141 data_opt = EXT3_MOUNT_WRITEBACK_DATA;
1142 datacheck:
1143 if (is_remount) {
1144 if (test_opt(sb, DATA_FLAGS) == data_opt)
1145 break;
1146 ext3_msg(sb, KERN_ERR,

1147                                         "error: cannot change "
1148                                         "data mode on remount. The filesystem "
1149                                         "is mounted in data=%s mode and you "
1150                                         "try to remount it in data=%s mode.",
1151                                         data_mode_string(test_opt(sb,

1152 DATA_FLAGS)),
1153 data_mode_string(data_opt));
1154 return 0;
1155 } else {
1156 clear_opt(sbi->s_mount_opt, DATA_FLAGS);
1157 sbi->s_mount_opt |= data_opt;
1158 }
1159 break;
1160 case Opt_data_err_abort:
1161 set_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1162 break;
1163 case Opt_data_err_ignore:
1164 clear_opt(sbi->s_mount_opt, DATA_ERR_ABORT);
1165 break;
1166 #ifdef CONFIG_QUOTA
1167 case Opt_usrjquota:
1168 if (!set_qf_name(sb, USRQUOTA, &args[0]))
1169 return 0;
1170 break;
1171 case Opt_grpjquota:
1172 if (!set_qf_name(sb, GRPQUOTA, &args[0]))
1173 return 0;
1174 break;
1175 case Opt_offusrjquota:
1176 if (!clear_qf_name(sb, USRQUOTA))
1177 return 0;
1178 break;
1179 case Opt_offgrpjquota:
1180 if (!clear_qf_name(sb, GRPQUOTA))
1181 return 0;
1182 break;

Signature · 7 lines
Chen Gang

Asianux Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
In reply to Chen Gang
Quoted 17 lines · Chen Gang

Hello Theodore Ts'o

in fs/ext3/supper.c
for function set_qf_name:

    sbi->s_qf_names[qtype] may already have owned a memory (line 919..925)
    we set sbi->s_qf_names[qtype] = qname directly without checking (line 926)

for function clear_qf_name:

    we set sbi->s_qf_names[qtype] = NULL (line 942..952)

for function parse_options:
we can call set_qf_name or clear_qf_name with USR or GRP many times.
we find parameters not mind whether they are repeated. (line 975..985) so we may call set_qf_name or clear_qf_name several times.
also may first call set_qf_name, then call clear_qf_name.

Quoted 3 lines

in this situation, we will get memory leak.

please help check this suggestion whether valid (I find it by code review).

Thanks for report. Yes, memory leak seems to be possible. Attached patch
should fix it, I have added it to my tree.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
Quoted 1 line

From 59cefce80eb3e081d0dbef335843340512aff79e Mon Sep 17 00:00:00 2001

From: Jan Kara <[email protected]>
Date: Mon, 31 Dec 2012 12:38:36 +0100
Subject: [PATCH] ext3: Fix memory leak when quota options are specified multiple times

When usrjquota or grpjquota mount options are specified several times with the same file name, we leak memory storing the names. Free the memory correctly.

---
 fs/ext3/super.c |   14 +++++++++-----
 1 files changed, 9 insertions(+), 5 deletions(-)
fs/ext3/super.c +9 -5
index 6e50223..b36cea9 100644
@@ -916,12 +916,16 @@ static int set_qf_name(struct super_block *sb, int qtype, substring_t *args)
 			"Not enough memory for storing quotafile name");
 		return 0;
 	}
-	if (sbi->s_qf_names[qtype] &&
-		strcmp(sbi->s_qf_names[qtype], qname)) {
-		ext3_msg(sb, KERN_ERR,
-			"%s quota file already specified", QTYPE2NAME(qtype));
+	if (sbi->s_qf_names[qtype]) {
+		int same = !strcmp(sbi->s_qf_names[qtype], qname);
+
 		kfree(qname);
-		return 0;
+		if (!same) {
+			ext3_msg(sb, KERN_ERR,
+				 "%s quota file already specified",
+				 QTYPE2NAME(qtype));
+		}
+		return same;
 	}
 	sbi->s_qf_names[qtype] = qname;
 	if (strchr(sbi->s_qf_names[qtype], '/')) {
Signature · 1 line
1.7.1
author
#3 lore raw
Quoted 5 lines · Jan Kara

Thanks for report. Yes, memory leak seems to be possible. Attached patch
should fix it, I have added it to my tree.

Honza

after read your patch, today (although I am not a reviewer). is it a possible situation like this (just my imagination):
1st let it call clear_qf_name successfully.
2nd let it call set_qf_name successfully.
3rd call clear_qf_name again ?
...
if this situation is possible.
I guess the memory still leak, after apply your patch.

if what I guess is true
it seems we have to duplicate memory for old_opts in ext3_remount. and for clear_qf_name, set_qf_name, ext3_remount:
use normal memory allocation and free ways instead of current using.

Regards

Signature · 7 lines
Chen Gang

Asianux Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#4 lore raw
Quoted 4 lines · Chen Gang
Quoted 3 lines · Jan Kara

Thanks for report. Yes, memory leak seems to be possible. Attached patch
should fix it, I have added it to my tree.

Quoted 8 lines
Quoted 2 lines

Honza

after read your patch, today (although I am not a reviewer). is it a possible situation like this (just my imagination):
1st let it call clear_qf_name successfully.
2nd let it call set_qf_name successfully.
3rd call clear_qf_name again ?

Although this is not likely in practice, if you specify mount options
like usrjquota=aquota.user,usrjquota=,usrjquota=aquota.user it can happen. Thanks for spotting this.

Quoted 8 lines
    ...

if this situation is possible.
I guess the memory still leak, after apply your patch.

if what I guess is true
it seems we have to duplicate memory for old_opts in ext3_remount. and for clear_qf_name, set_qf_name, ext3_remount:
use normal memory allocation and free ways instead of current using.

Yes. Updated patch attached.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
Quoted 1 line

From c36504046b93419fba7ec0029a43afd2bbdd6c48 Mon Sep 17 00:00:00 2001

From: Jan Kara <[email protected]>
Date: Mon, 31 Dec 2012 12:38:36 +0100
Subject: [PATCH] ext3: Fix memory leak when quota options are specified multiple times

When usrjquota or grpjquota mount options are specified several times, we leak memory storing the names. Free the memory correctly.

---
 fs/ext3/super.c |   51 ++++++++++++++++++++++++++++++---------------------
 1 files changed, 30 insertions(+), 21 deletions(-)
fs/ext3/super.c +30 -21
index 6e50223..0926fe4 100644
@@ -916,21 +916,24 @@ static int set_qf_name(struct super_block *sb, int qtype, substring_t *args)
 			"Not enough memory for storing quotafile name");
 		return 0;
 	}
-	if (sbi->s_qf_names[qtype] &&
-		strcmp(sbi->s_qf_names[qtype], qname)) {
-		ext3_msg(sb, KERN_ERR,
-			"%s quota file already specified", QTYPE2NAME(qtype));
+	if (sbi->s_qf_names[qtype]) {
+		int same = !strcmp(sbi->s_qf_names[qtype], qname);
+
 		kfree(qname);
-		return 0;
+		if (!same) {
+			ext3_msg(sb, KERN_ERR,
+				 "%s quota file already specified",
+				 QTYPE2NAME(qtype));
+		}
+		return same;
 	}
-	sbi->s_qf_names[qtype] = qname;
-	if (strchr(sbi->s_qf_names[qtype], '/')) {
+	if (strchr(qname, '/')) {
 		ext3_msg(sb, KERN_ERR,
 			"quotafile must be on filesystem root");
-		kfree(sbi->s_qf_names[qtype]);
-		sbi->s_qf_names[qtype] = NULL;
+		kfree(qname);
 		return 0;
 	}
+	sbi->s_qf_names[qtype] = qname;
 	set_opt(sbi->s_mount_opt, QUOTA);
 	return 1;
 }
@@ -945,11 +948,10 @@ static int clear_qf_name(struct super_block *sb, int qtype) {
 			" when quota turned on");
 		return 0;
 	}
-	/*
-	 * The space will be released later when all options are confirmed
-	 * to be correct
-	 */
-	sbi->s_qf_names[qtype] = NULL;
+	if (sbi->s_qf_names[qtype]) {
+		kfree(sbi->s_qf_names[qtype]);
+		sbi->s_qf_names[qtype] = NULL;
+	}
 	return 1;
 }
 #endif
@@ -2605,7 +2607,18 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 #ifdef CONFIG_QUOTA
 	old_opts.s_jquota_fmt = sbi->s_jquota_fmt;
 	for (i = 0; i < MAXQUOTAS; i++)
-		old_opts.s_qf_names[i] = sbi->s_qf_names[i];
+		if (sbi->s_qf_names[i]) {
+			old_opts.s_qf_names[i] = kstrdup(sbi->s_qf_names[i],
+							 GFP_KERNEL);
+			if (!old_opts.s_qf_names[i]) {
+				int j;
+
+				for (j = 0; j < i; j++)
+					kfree(old_opts.s_qf_names[j]);
+				return -ENOMEM;
+			}
+		} else
+			old_opts.s_qf_names[i] = NULL;
 #endif
 
 	/*
@@ -2698,9 +2711,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data)
 #ifdef CONFIG_QUOTA
 	/* Release old quota file names */
 	for (i = 0; i < MAXQUOTAS; i++)
-		if (old_opts.s_qf_names[i] &&
-		    old_opts.s_qf_names[i] != sbi->s_qf_names[i])
-			kfree(old_opts.s_qf_names[i]);
+		kfree(old_opts.s_qf_names[i]);
 #endif
 	if (enable_quota)
 		dquot_resume(sb, -1);
@@ -2714,9 +2725,7 @@ restore_opts:
 #ifdef CONFIG_QUOTA
 	sbi->s_jquota_fmt = old_opts.s_jquota_fmt;
 	for (i = 0; i < MAXQUOTAS; i++) {
-		if (sbi->s_qf_names[i] &&
-		    old_opts.s_qf_names[i] != sbi->s_qf_names[i])
-			kfree(sbi->s_qf_names[i]);
+		kfree(sbi->s_qf_names[i]);
 		sbi->s_qf_names[i] = old_opts.s_qf_names[i];
 	}
 #endif
Signature · 1 line
1.7.1
author
#5 lore raw
Quoted 7 lines · Jan Kara
Quoted 4 lines · Chen Gang

if what I guess is true
it seems we have to duplicate memory for old_opts in ext3_remount. and for clear_qf_name, set_qf_name, ext3_remount:
use normal memory allocation and free ways instead of current using.

Yes. Updated patch attached.

at least for me, Jan Kara's patch is ok. thanks.

welcome another members (especially Theodore Ts'o) to giving confirmation or completion. also welcome to checking this suggestion whether valid for fs/ext4.

Signature · 7 lines
Chen Gang

Asianux Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
author
#6 lore raw
Quoted 6 lines · Chen Gang
Quoted 5 lines · Jan Kara
Quoted 5 lines · Chen Gang
Quoted 4 lines
Quoted 4 lines

if what I guess is true
it seems we have to duplicate memory for old_opts in ext3_remount. and for clear_qf_name, set_qf_name, ext3_remount:
use normal memory allocation and free ways instead of current using.

Quoted 6 lines
Quoted 2 lines
Quoted 2 lines

Yes. Updated patch attached.

at least for me, Jan Kara's patch is ok. thanks.

welcome another members (especially Theodore Ts'o) to giving confirmation or completion. also welcome to checking this suggestion whether valid for fs/ext4.

I will send a relative patch for fs/ext4, just according to Jan Kara's patch. I will send it within 2 weeks (excuse me, I have to do another things this week).

welcome any additional suggestions and completions.

thanks.

Signature · 7 lines
Chen Gang

Asianux Corporation
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html