Larger xtables-save review
Phil Sutter
reviewercover Larger xtables-save review
This series started as a fix to program names mentioned in *-save outputs and ended in merging ebtables-save and arptables-save code into xtables_save_main used by ip{6,}tables-nft-save.
The first patch is actually unrelated but was discovered when testing counter output - depending on environment, ebtables-nft might segfault.
The second patch fixes option '-c' of ebtables-nft-save which enables counter prefixes in dumped rules but failed to disable the classical ebtables-style counters.
Patch three sorts program names quoted in output of any of the *-save programs, patch four unifies the header/footer comments in the same. The latter also drops the extra newline printed in ebtables- and arptables-save output, so test scripts need adjustments beyond dropping the new comment lines from output.
Patch five fixes the table compatibility check in ip{6,}tables-nft-save.
Patches six and eight to ten prepare for integrating arptables- and ebtables-save into the xtables-save code.
Patch seven merely fixes a minor coding-style issue.
Patches eleven and twelve finally perform the actual merge.
Phil Sutter (12):
ebtables: Fix error message for invalid parameters
ebtables-save: Fix counter formatting
xtables-save: Use argv[0] as program name
xtables-save: Unify *-save header/footer comments
xtables-save: Fix table compatibility check
nft: Make nft_for_each_table() more versatile
xtables-save: Avoid mixed code and declarations
xtables-save: Pass optstring/longopts to xtables_save_main()
xtables-save: Make COMMIT line optional
xtables-save: Pass format flags to do_output()
arptables-save: Merge into xtables_save_main()
ebtables-save: Merge into xtables_save_main()
iptables/nft-bridge.c | 39 +-- iptables/nft.c | 6 +- iptables/nft.h | 2 +- .../arptables/0001-arptables-save-restore_0 | 7 +- .../0002-arptables-restore-defaults_0 | 6 +- .../arptables/0003-arptables-verbose-output_0 | 5 +- .../ebtables/0002-ebtables-save-restore_0 | 4 +- .../ebtables/0003-ebtables-restore-defaults_0 | 6 +- .../testcases/ebtables/0004-save-counters_0 | 64 +++++ iptables/xtables-eb.c | 4 +- iptables/xtables-save.c | 242 ++++-------------- 11 files changed, 146 insertions(+), 239 deletions(-) create mode 100755 iptables/tests/shell/testcases/ebtables/0004-save-counters_0
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 1/12 ebtables: Fix error message for invalid parameters
With empty ruleset, ebtables-nft would report the wrong argv:
| % sudo ./install/sbin/ebtables-nft -vnL | ebtables v1.8.3 (nf_tables): Unknown argument: './install/sbin/ebtables-nft' | Try `ebtables -h' or 'ebtables --help' for more information.
After a (successful) call to 'ebtables-nft -L', this would even segfault:
| % sudo ./install/sbin/ebtables-nft -vnL
| zsh: segmentation fault sudo ./install/sbin/ebtables-nft -vnL
--- iptables/xtables-eb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
iptables/xtables-eb.c +1 -1
@@ -1180,7 +1180,7 @@ print_zero:if (ebt_command_default (&cs))xtables_error (PARAMETER_PROBLEM ,"Unknown argument: '%s'" , - argv[optind -1 ]); + argv[optind]);if (command !='A' && command !='I' && command !='D' && command !='C' )
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 2/12 ebtables-save: Fix counter formatting
In reply to Phil SutterThe initial problem was 'ebtables-save -c' printing iptables-style counters but at the same time not disabling ebtables-style counter output (which was even printed in wrong format for ebtables-save).
The code around counter output was complicated enough to motivate a larger rework:
* Make FMT_C_COUNTS indicate the appended counter style for ebtables.
* Use FMT_EBT_SAVE to distinguish between '-c' style counters and the legacy pcnt/bcnt ones.
Consequently, ebtables-save sets format to:
FMT_NOCOUNTS - for no countersFMT_EBT_SAVE - for iptables-style countersFMT_EBT_SAVE | FMT_C_COUNTS - for '-c' style counters
For regular ebtables, list_rules() always sets FMT_C_COUNTS (iptables-style counters are never used there) and FMT_NOCOUNTS if no counters are requested.
The big plus is if neither FMT_NOCOUNTS nor FMT_C_COUNTS is set, iptables-style counters are to be printed - both in iptables and ebtables. This allows to drop the ebtables-specific 'save_counters' callback.
--- iptables/nft-bridge.c | 39 +++-------- .../testcases/ebtables/0004-save-counters_0 | 67 +++++++++++++++++++ iptables/xtables-eb.c | 2 +- iptables/xtables-save.c | 3 +- 4 files changed, 81 insertions(+), 30 deletions(-) create mode 100755 iptables/tests/shell/testcases/ebtables/0004-save-counters_0
iptables/nft-bridge.c +11 -28
@@ -21,8 +21,6 @@#include "nft-bridge.h" #include "nft.h" -static bool ebt_legacy_counter_fmt; -void ebt_cs_clean (struct iptables_command_state *cs) {struct ebt_match *m, *nm;
@@ -422,22 +420,6 @@ static void print_protocol(uint16_t ethproto, bool invert, unsigned int bitmask)printf ("%s " , ent->e_name); } -static void nft_bridge_save_counters (const void *data) -{ -const char *ctr; - -if (ebt_legacy_counter_fmt) -return ; - - ctr =getenv ("EBTABLES_SAVE_COUNTER" ); -if (ctr) { - ebt_legacy_counter_fmt = true; -return ; - } - -save_counters (data); -} -static void nft_bridge_save_rule (const void *data,unsigned int format) {const struct iptables_command_state *cs = data;
@@ -474,15 +456,16 @@ static void nft_bridge_save_rule(const void *data, unsigned int format) cs->target->FMT_NUMERIC ); } -if (format &FMT_EBT_SAVE ) -printf (" -c %" PRIu64" %" PRIu64"" , - (uint64_t )cs->counters.pcnt, - (uint64_t )cs->counters.bcnt); - -if (!(format &FMT_NOCOUNTS )) -printf (" , pcnt = %" PRIu64" -- bcnt = %" PRIu64"" , - (uint64_t )cs->counters.pcnt, - (uint64_t )cs->counters.bcnt); +if ((format & (FMT_NOCOUNTS |FMT_C_COUNTS )) ==FMT_C_COUNTS ) { +if (format &FMT_EBT_SAVE ) +printf (" -c %" PRIu64" %" PRIu64"" , + (uint64_t )cs->counters.pcnt, + (uint64_t )cs->counters.bcnt); +else +printf (" , pcnt = %" PRIu64" -- bcnt = %" PRIu64"" , + (uint64_t )cs->counters.pcnt, + (uint64_t )cs->counters.bcnt); + }if (!(format &FMT_NONEWLINE ))fputc ('\n' , stdout);
@@ -763,7 +746,7 @@ struct nft_family_ops nft_family_ops_bridge = { .print_header = nft_bridge_print_header, .print_rule = nft_bridge_print_rule, .save_rule = nft_bridge_save_rule, - .save_counters = nft_bridge_save_counters, + .save_counters = save_counters, .save_chain = nft_bridge_save_chain, .post_parse = NULL, .rule_to_cs = nft_rule_to_ebtables_command_state,
iptables/tests/shell/testcases/ebtables/0004-save-counters_0 new file +67 -0
67 lines not on this page · raw message
iptables/xtables-eb.c +1 -1
@@ -410,7 +410,7 @@ static int list_rules(struct nft_handle *h, const char *chain, const char *table {unsigned int format; - format =FMT_OPTIONS ; + format =FMT_OPTIONS |FMT_C_COUNTS ;if (verbose) format |=FMT_VIA ;
iptables/xtables-save.c +2 -1
@@ -273,7 +273,8 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool countersprintf ("*%s\n" , tablename);if (counters) - format = ebt_legacy_counter_format ?FMT_EBT_SAVE :0 ; + format =FMT_EBT_SAVE | + (ebt_legacy_counter_format ?FMT_C_COUNTS :0 );/* Dump out chain names first, * thereby preventing dependency conflicts */
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 3/12 xtables-save: Use argv[0] as program name
In reply to Phil SutterDon't hard-code program names. This also fixes for bogus 'xtables-save' name which is no longer used.
--- iptables/xtables-save.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-)
iptables/xtables-save.c +11 -10
@@ -9,6 +9,7 @@#include "config.h" #include <getopt.h> #include <errno.h> +#include <libgen.h> #include <stdio.h> #include <fcntl.h> #include <stdlib.h>
@@ -80,8 +81,8 @@ __do_output(struct nft_handle *h, const char *tablename, bool counters)time_t now =time (NULL ); -printf ("# Generated by xtables-save v%s on %s" , -PACKAGE_VERSION ,ctime (&now)); +printf ("# Generated by %s v%s on %s" , prog_name, + prog_vers,ctime (&now));printf ("*%s\n" , tablename);/* Dump out chain names first,
@@ -122,7 +123,7 @@ do_output(struct nft_handle *h, const char *tablename, bool counters)* rule */ static int -xtables_save_main (int family,const char *progname,int argc,char *argv[]) +xtables_save_main (int family,int argc,char *argv[]) {const struct builtin_table *tables;const char *tablename =NULL ;
@@ -133,7 +134,7 @@ xtables_save_main(int family, const char *progname, int argc, char *argv[])FILE *file =NULL ;int ret, c; - xtables_globals.program_name = progname; + xtables_globals.program_name =basename (*argv);; c =xtables_init_all (&xtables_globals, family);if (c <0 ) { fprintf(stderr, "%s/%s Failed to initialize xtables\n",
@@ -237,12 +238,12 @@ xtables_save_main(int family, const char *progname, int argc, char *argv[])int xtables_ip4_save_main (int argc,char *argv[]) { -return xtables_save_main (NFPROTO_IPV4 ,"iptables-save" , argc, argv); +return xtables_save_main (NFPROTO_IPV4 , argc, argv); }int xtables_ip6_save_main (int argc,char *argv[]) { -return xtables_save_main (NFPROTO_IPV6 ,"ip6tables-save" , argc, argv); +return xtables_save_main (NFPROTO_IPV6 , argc, argv); }static int __ebt_save (struct nft_handle *h,const char *tablename,bool counters)
@@ -266,8 +267,8 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool countersif (first) { now =time (NULL ); -printf ("# Generated by ebtables-save v%s on %s" , -PACKAGE_VERSION ,ctime (&now)); +printf ("# Generated by %s v%s on %s" , prog_name, + prog_vers,ctime (&now)); first = false; }printf ("*%s\n" , tablename);
@@ -308,7 +309,7 @@ int xtables_eb_save_main(int argc_, char *argv_[]) } } - xtables_globals.program_name ="ebtables-save" ; + xtables_globals.program_name =basename (*argv_); c =xtables_init_all (&xtables_globals, h.family);if (c <0 ) { fprintf(stderr, "%s/%s Failed to initialize xtables\n",
@@ -362,7 +363,7 @@ int xtables_arp_save_main(int argc, char **argv) };int c; - xtables_globals.program_name ="arptables-save" ; + xtables_globals.program_name =basename (*argv);; c =xtables_init_all (&xtables_globals, h.family);if (c <0 ) { fprintf(stderr, "%s/%s Failed to initialize xtables\n",
Signature · 1 line
2.22.0
Phil Sutter
reviewerRe: [iptables PATCH 03/12] xtables-save: Use argv[0] as program name
Quoted 4 lines · Phil Sutter
Don't hard-code program names. This also fixes for bogus 'xtables-save' name which is no longer used.
Signed-off-by Phil Sutter <[email protected]>
Argh, I should have pulled upstream first, this one was already accepted. My series rebases cleanly, should I respin?
Sorry for the mess. :(
Pablo Neira Ayuso
maintainerRe: [iptables PATCH 03/12] xtables-save: Use argv[0] as program name
Quoted 8 lines · Phil Sutter
Quoted 4 lines · Phil Sutter
Don't hard-code program names. This also fixes for bogus 'xtables-save' name which is no longer used.
Signed-off-by Phil Sutter <[email protected]>Argh, I should have pulled upstream first, this one was already accepted. My series rebases cleanly, should I respin?
Please, if not matching asking, please do so. Also addressing Florian's feedback updating the comments, thanks.
I'll make a quick review tomorrow, sorry but I cannot get on these any sooner.
Phil Sutter
reviewerPATCH 4/12 xtables-save: Unify *-save header/footer comments
In reply to Phil SutterMake eb- and arptables-save print both header and footer comments, too. Also print them for each table separately - the timing information is worth the extra lines in output.
--- .../arptables/0001-arptables-save-restore_0 | 7 +++---- .../0002-arptables-restore-defaults_0 | 6 ++---- .../arptables/0003-arptables-verbose-output_0 | 5 ++--- .../ebtables/0002-ebtables-save-restore_0 | 4 +--- .../ebtables/0003-ebtables-restore-defaults_0 | 6 ++---- .../testcases/ebtables/0004-save-counters_0 | 9 +++------ iptables/xtables-save.c | 19 ++++++++++--------- 7 files changed, 23 insertions(+), 33 deletions(-)
iptables/tests/shell/testcases/arptables/0001-arptables-save-restore_0 +3 -4
@@ -50,13 +50,12 @@ DUMP='*filter -A foo -j MARK --set-mark 12345 -A foo -j ACCEPT --opcode 1 -A foo -j ACCEPT --proto-type 0x800 --A foo -j ACCEPT -i lo --opcode 1 --proto-type 0x800 -' +-A foo -j ACCEPT -i lo --opcode 1 --proto-type 0x800' -diff -u <(echo -e "$DUMP") <($XT_MULTI arptables-save) +diff -u <(echo -e "$DUMP") <($XT_MULTI arptables-save | grep -v "^#") # make sure dump can be restored and check it didn't change $XT_MULTI arptables -F $XT_MULTI arptables-restore <<<$DUMP -diff -u <(echo -e "$DUMP") <($XT_MULTI arptables-save) +diff -u <(echo -e "$DUMP") <($XT_MULTI arptables-save | grep -v "^#")
iptables/tests/shell/testcases/arptables/0002-arptables-restore-defaults_0 +2 -4
@@ -11,8 +11,7 @@ set -e DUMP='*filter :OUTPUT ACCEPT -A OUTPUT -j mangle --mangle-ip-s 10.0.0.1 --A OUTPUT -j mangle --mangle-ip-d 10.0.0.2 -' +-A OUTPUT -j mangle --mangle-ip-d 10.0.0.2' # note how mangle-ip-s is unset in second rule
@@ -20,8 +19,7 @@ EXPECT='*filter :INPUT ACCEPT :OUTPUT ACCEPT -A OUTPUT -j mangle --mangle-ip-s 10.0.0.1 --A OUTPUT -j mangle --mangle-ip-d 10.0.0.2 -' +-A OUTPUT -j mangle --mangle-ip-d 10.0.0.2' $XT_MULTI arptables -F $XT_MULTI arptables-restore <<<$DUMP
iptables/tests/shell/testcases/arptables/0003-arptables-verbose-output_0 +2 -3
@@ -58,7 +58,6 @@ EXPECT='*filter -A INPUT -j MARK -i eth23 --set-mark 42 -A OUTPUT -j CLASSIFY -o eth23 --set-class 23:42 -A OUTPUT -j foo -o eth23 --A foo -j mangle -o eth23 --mangle-ip-s 10.0.0.1 -' +-A foo -j mangle -o eth23 --mangle-ip-s 10.0.0.1' -diff -u -Z <(echo -e "$EXPECT") <($XT_MULTI arptables-save) +diff -u -Z <(echo -e "$EXPECT") <($XT_MULTI arptables-save | grep -v '^#')
iptables/tests/shell/testcases/ebtables/0002-ebtables-save-restore_0 +1 -3
@@ -99,7 +99,6 @@ DUMP='*filter -A foo --802_3-sap 0x23 --limit 100/sec --limit-burst 5 -j ACCEPT -A foo --pkttype-type multicast --log-level notice --log-prefix "" -j CONTINUE -A foo --pkttype-type multicast --limit 100/sec --limit-burst 5 -j ACCEPT - *nat :PREROUTING ACCEPT :OUTPUT DROP
@@ -107,8 +106,7 @@ DUMP='*filter :nat_foo DROP -A PREROUTING -j redirect -A OUTPUT -j ACCEPT --A POSTROUTING -j ACCEPT -' +-A POSTROUTING -j ACCEPT' diff -u <(echo -e "$DUMP") <($XT_MULTI ebtables-save | grep -v '^#')
iptables/tests/shell/testcases/ebtables/0003-ebtables-restore-defaults_0 +2 -4
@@ -13,8 +13,7 @@ DUMP='*filter -A FORWARD --limit 100 --limit-burst 42 -j ACCEPT -A FORWARD --limit 1000 -j ACCEPT -A FORWARD --log --log-prefix "foobar" --A FORWARD --log -' +-A FORWARD --log' # note how limit-burst is 5 in second rule and log-prefix empty in fourth one
@@ -25,8 +24,7 @@ EXPECT='*filter -A FORWARD --limit 100/sec --limit-burst 42 -j ACCEPT -A FORWARD --limit 1000/sec --limit-burst 5 -j ACCEPT -A FORWARD --log-level notice --log-prefix "foobar" -j CONTINUE --A FORWARD --log-level notice --log-prefix "" -j CONTINUE -' +-A FORWARD --log-level notice --log-prefix "" -j CONTINUE' $XT_MULTI ebtables --init-table $XT_MULTI ebtables-restore <<<$DUMP
iptables/tests/shell/testcases/ebtables/0004-save-counters_0 +3 -6
@@ -32,8 +32,7 @@ EXPECT='*filter :FORWARD ACCEPT :OUTPUT ACCEPT -A FORWARD -i nodev123 -o nodev432 -j ACCEPT --A FORWARD -i nodev432 -o nodev123 -j ACCEPT -' +-A FORWARD -i nodev432 -o nodev123 -j ACCEPT' echo "ebtables-save" diff -u <(echo -e "$EXPECT") <($XT_MULTI ebtables-save | grep -v '^#')
@@ -43,8 +42,7 @@ EXPECT='*filter :FORWARD ACCEPT :OUTPUT ACCEPT [0:0] -A FORWARD -i nodev123 -o nodev432 -j ACCEPT -[0:0] -A FORWARD -i nodev432 -o nodev123 -j ACCEPT -' +[0:0] -A FORWARD -i nodev432 -o nodev123 -j ACCEPT' echo "ebtables-save -c" diff -u <(echo -e "$EXPECT") <($XT_MULTI ebtables-save -c | grep -v '^#')
@@ -60,8 +58,7 @@ EXPECT='*filter :FORWARD ACCEPT :OUTPUT ACCEPT -A FORWARD -i nodev123 -o nodev432 -j ACCEPT -c 0 0 --A FORWARD -i nodev432 -o nodev123 -j ACCEPT -c 0 0 -' +-A FORWARD -i nodev432 -o nodev123 -j ACCEPT -c 0 0' echo "EBTABLES_SAVE_COUNTER=yes ebtables-save" diff -u <(echo -e "$EXPECT") <($XT_MULTI ebtables-save | grep -v '^#')
iptables/xtables-save.c +10 -9
@@ -250,7 +250,6 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool counters {struct nftnl_chain_list *chain_list;unsigned int format =FMT_NOCOUNTS ; -static bool first = true;time_t now;if (!nft_table_find (h, tablename)) {
@@ -265,12 +264,9 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool counters chain_list =nft_chain_list_get (h, tablename); -if (first) { - now =time (NULL ); -printf ("# Generated by %s v%s on %s" , prog_name, - prog_vers,ctime (&now)); - first = false; - } + now =time (NULL ); +printf ("# Generated by %s v%s on %s" , prog_name, + prog_vers,ctime (&now));printf ("*%s\n" , tablename);if (counters)
@@ -281,7 +277,8 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool counters* thereby preventing dependency conflicts */ nft_chain_save (h, chain_list);nft_rule_save (h, tablename, format); -printf ("\n" ); + now =time (NULL ); +printf ("# Completed on %s" ,ctime (&now));return 0 ; }
@@ -361,6 +358,7 @@ int xtables_arp_save_main(int argc, char **argv)struct nft_handle h = { .family =NFPROTO_ARP , }; +time_t now;int c; xtables_globals.program_name =basename (*argv);;
@@ -407,10 +405,13 @@ int xtables_arp_save_main(int argc, char **argv)return 0 ; } +printf ("# Generated by %s v%s on %s" , prog_name, + prog_vers,ctime (&now));printf ("*filter\n" );nft_chain_save (&h,nft_chain_list_get (&h,"filter" ));nft_rule_save (&h,"filter" , show_counters ?0 :FMT_NOCOUNTS ); -printf ("\n" ); + now =time (NULL ); +printf ("# Completed on %s" ,ctime (&now));nft_fini (&h);return 0 ; }
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 5/12 xtables-save: Fix table compatibility check
In reply to Phil SutterThe builtin table check guarding the 'is incompatible' warning was
wrong: The idea was to print the warning only for incompatible tables which are builtin, not for others. Yet the code would print the warning only for non-builtin ones.
Also reorder the checks: nft_table_builtin_find() is fast and therefore a quick way to bail for uninteresting tables. The compatibility check is needed for the remaining tables, only.
--- iptables/xtables-save.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
iptables/xtables-save.c +4 -3
@@ -67,11 +67,12 @@ __do_output(struct nft_handle *h, const char *tablename, bool counters) {struct nftnl_chain_list *chain_list; +if (!nft_table_builtin_find (h, tablename)) +return 0 ;if (!nft_is_table_compatible (h, tablename)) { -if (!nft_table_builtin_find (h, tablename)) -printf ("# Table `%s' is incompatible, use 'nft' tool.\n" , - tablename); +printf ("# Table `%s' is incompatible, use 'nft' tool.\n" , + tablename);return 0 ; }
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 6/12 nft: Make nft_for_each_table() more versatile
In reply to Phil SutterSupport passing arbitrary data (via void pointer) to the callback.
--- iptables/nft.c | 6 +++--- iptables/nft.h | 2 +- iptables/xtables-save.c | 27 +++++++++++++++++---------- 3 files changed, 21 insertions(+), 14 deletions(-)
iptables/nft.c +3 -3
@@ -2092,8 +2092,8 @@ err: }int nft_for_each_table (struct nft_handle *h, -int (*func)(struct nft_handle *h,const char *tablename,bool counters), -bool counters) +int (*func)(struct nft_handle *h,const char *tablename,void *data), +void *data) {struct nftnl_table_list *list;struct nftnl_table_list_iter *iter;
@@ -2112,7 +2112,7 @@ int nft_for_each_table(struct nft_handle *h,const char *tablename =nftnl_table_get (t,NFTNL_TABLE_NAME ); -func (h, tablename, counters); +func (h, tablename, data); t =nftnl_table_list_iter_next (iter); }
iptables/nft.h +1 -1
@@ -81,7 +81,7 @@ void nft_build_cache(struct nft_handle *h);struct nftnl_table ;struct nftnl_chain_list ; -int nft_for_each_table (struct nft_handle *h,int (*func)(struct nft_handle *h,const char *tablename,bool counters),bool counters); +int nft_for_each_table (struct nft_handle *h,int (*func)(struct nft_handle *h,const char *tablename,void *data),void *data);bool nft_table_find (struct nft_handle *h,const char *tablename);int nft_table_purge_chains (struct nft_handle *h,const char *table,struct nftnl_chain_list *list);int nft_table_flush (struct nft_handle *h,const char *table);
iptables/xtables-save.c +17 -10
@@ -62,10 +62,15 @@ static const struct option ebt_save_options[] = {static bool ebt_legacy_counter_format; +struct do_output_data { +bool counters; +}; +static int -__do_output (struct nft_handle *h,const char *tablename,bool counters) +__do_output (struct nft_handle *h,const char *tablename,void *data) {struct nftnl_chain_list *chain_list; +struct do_output_data *d = data;if (!nft_table_builtin_find (h, tablename))return 0 ;
@@ -89,7 +94,7 @@ __do_output(struct nft_handle *h, const char *tablename, bool counters)/* Dump out chain names first, * thereby preventing dependency conflicts */ nft_chain_save (h, chain_list); -nft_rule_save (h, tablename, counters ?0 :FMT_NOCOUNTS ); +nft_rule_save (h, tablename, d->counters ?0 :FMT_NOCOUNTS ); now =time (NULL );printf ("COMMIT\n" );
@@ -98,12 +103,12 @@ __do_output(struct nft_handle *h, const char *tablename, bool counters) }static int -do_output (struct nft_handle *h,const char *tablename,bool counters) +do_output (struct nft_handle *h,const char *tablename,struct do_output_data *d) {int ret;if (!tablename) { - ret =nft_for_each_table (h, __do_output, counters); + ret =nft_for_each_table (h, __do_output, d);nft_check_xt_legacy (h->family, true);return !!ret; }
@@ -114,7 +119,7 @@ do_output(struct nft_handle *h, const char *tablename, bool counters)return 1 ; } - ret =__do_output (h, tablename, counters); + ret =__do_output (h, tablename, d);nft_check_xt_legacy (h->family, true);return ret; }
@@ -128,6 +133,7 @@ xtables_save_main(int family, int argc, char *argv[]) {const struct builtin_table *tables;const char *tablename =NULL ; +struct do_output_data d = {};bool dump = false;struct nft_handle h = { .family = family,
@@ -150,7 +156,7 @@ xtables_save_main(int family, int argc, char *argv[])fprintf (stderr,"-b/--binary option is not implemented\n" );break ;case 'c' : - show_counters = true; + d.counters = true;break ;case 't' :
@@ -229,7 +235,7 @@ xtables_save_main(int family, int argc, char *argv[])exit (EXIT_FAILURE ); } - ret =do_output (&h, tablename, show_counters); + ret =do_output (&h, tablename, &d);nft_fini (&h);if (dump)exit (0 );
@@ -247,10 +253,11 @@ int xtables_ip6_save_main(int argc, char *argv[])return xtables_save_main (NFPROTO_IPV6 , argc, argv); } -static int __ebt_save (struct nft_handle *h,const char *tablename,bool counters) +static int __ebt_save (struct nft_handle *h,const char *tablename,void *data) {struct nftnl_chain_list *chain_list;unsigned int format =FMT_NOCOUNTS ; +bool *counters = data;time_t now;if (!nft_table_find (h, tablename)) {
@@ -286,9 +293,9 @@ static int __ebt_save(struct nft_handle *h, const char *tablename, bool countersstatic int ebt_save (struct nft_handle *h,const char *tablename,bool counters) {if (!tablename) -return nft_for_each_table (h, __ebt_save, counters); +return nft_for_each_table (h, __ebt_save, &counters); -return __ebt_save (h, tablename, counters); +return __ebt_save (h, tablename, &counters); }int xtables_eb_save_main (int argc_,char *argv_[])
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 7/12 xtables-save: Avoid mixed code and declarations
In reply to Phil SutterAlso move time() calls to where they are used.
--- iptables/xtables-save.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
iptables/xtables-save.c +4 -4
@@ -71,6 +71,7 @@ __do_output(struct nft_handle *h, const char *tablename, void *data) {struct nftnl_chain_list *chain_list;struct do_output_data *d = data; +time_t now;if (!nft_table_builtin_find (h, tablename))return 0 ;
@@ -85,19 +86,18 @@ __do_output(struct nft_handle *h, const char *tablename, void *data)if (!chain_list)return 0 ; -time_t now =time (NULL ); - + now =time (NULL );printf ("# Generated by %s v%s on %s" , prog_name, prog_vers,ctime (&now)); -printf ("*%s\n" , tablename); +printf ("*%s\n" , tablename);/* Dump out chain names first, * thereby preventing dependency conflicts */ nft_chain_save (h, chain_list);nft_rule_save (h, tablename, d->counters ?0 :FMT_NOCOUNTS ); +printf ("COMMIT\n" ); now =time (NULL ); -printf ("COMMIT\n" );printf ("# Completed on %s" ,ctime (&now));return 0 ; }
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 8/12 xtables-save: Pass optstring/longopts to xtables_save_main()
In reply to Phil SutterIntroduce variables for the different optstrings so short and long options live side-by-side.
In order to make xtables_save_main() more versatile, pass optstring and longopts via parameter.
--- iptables/xtables-save.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-)
iptables/xtables-save.c +13 -7
@@ -33,7 +33,8 @@static bool show_counters = false; -static const struct option options[] = { +static const char *ipt_save_optstring ="bcdt:M:f:46V" ; +static const struct option ipt_save_options[] = { {.name ="counters" , .has_arg = false, .val ='c' }, {.name ="version" , .has_arg = false, .val ='V' }, {.name ="dump" , .has_arg = false, .val ='d' },
@@ -45,6 +46,7 @@ static const struct option options[] = { {NULL}, }; +static const char *arp_save_optstring ="cM:V" ;static const struct option arp_save_options[] = { {.name ="counters" , .has_arg = false, .val ='c' }, {.name = "version", .has_arg = false, .val = 'V'},
@@ -52,6 +54,7 @@ static const struct option arp_save_options[] = { {NULL}, }; +static const char *ebt_save_optstring ="ct:M:V" ;static const struct option ebt_save_options[] = { {.name ="counters" , .has_arg = false, .val ='c' }, {.name = "version", .has_arg = false, .val = 'V'},
@@ -129,7 +132,8 @@ do_output(struct nft_handle *h, const char *tablename, struct do_output_data *d)* rule */ static int -xtables_save_main (int family,int argc,char *argv[]) +xtables_save_main (int family,int argc,char *argv[], +const char *optstring,const struct option *longopts) {const struct builtin_table *tables;const char *tablename =NULL ;
@@ -150,7 +154,7 @@ xtables_save_main(int family, int argc, char *argv[])exit (1 ); } -while ((c =getopt_long (argc, argv,"bcdt:M:f:46V" , options,NULL )) !=-1 ) { +while ((c =getopt_long (argc, argv, optstring, longopts,NULL )) !=-1 ) {switch (c) {case 'b' :fprintf (stderr,"-b/--binary option is not implemented\n" );
@@ -245,12 +249,14 @@ xtables_save_main(int family, int argc, char *argv[])int xtables_ip4_save_main (int argc,char *argv[]) { -return xtables_save_main (NFPROTO_IPV4 , argc, argv); +return xtables_save_main (NFPROTO_IPV4 , argc, argv, + ipt_save_optstring, ipt_save_options); }int xtables_ip6_save_main (int argc,char *argv[]) { -return xtables_save_main (NFPROTO_IPV6 , argc, argv); +return xtables_save_main (NFPROTO_IPV6 , argc, argv, + ipt_save_optstring, ipt_save_options); }static int __ebt_save (struct nft_handle *h,const char *tablename,void *data)
@@ -323,7 +329,7 @@ int xtables_eb_save_main(int argc_, char *argv_[])exit (1 ); } -while ((c =getopt_long (argc_, argv_,"ct:M:V" , ebt_save_options,NULL )) !=-1 ) { +while ((c =getopt_long (argc_, argv_, ebt_save_optstring, ebt_save_options,NULL )) !=-1 ) {switch (c) {case 'c' :unsetenv ("EBTABLES_SAVE_COUNTER" );
@@ -378,7 +384,7 @@ int xtables_arp_save_main(int argc, char **argv)exit (1 ); } -while ((c =getopt_long (argc, argv,"cM:V" , arp_save_options,NULL )) !=-1 ) { +while ((c =getopt_long (argc, argv, arp_save_optstring, arp_save_options,NULL )) !=-1 ) {switch (c) {case 'c' : show_counters = true;
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 9/12 xtables-save: Make COMMIT line optional
In reply to Phil SutterChange xtables_save_main to support optional printing of COMMIT line as it is not used in arp- or ebtables.
--- iptables/xtables-save.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
iptables/xtables-save.c +4 -1
@@ -67,6 +67,7 @@ static bool ebt_legacy_counter_format;struct do_output_data {bool counters; +bool commit; }; static int
@@ -98,7 +99,8 @@ __do_output(struct nft_handle *h, const char *tablename, void *data)* thereby preventing dependency conflicts */ nft_chain_save (h, chain_list);nft_rule_save (h, tablename, d->counters ?0 :FMT_NOCOUNTS ); -printf ("COMMIT\n" ); +if (d->commit) +printf ("COMMIT\n" ); now =time (NULL );printf ("# Completed on %s" ,ctime (&now));
@@ -219,6 +221,7 @@ xtables_save_main(int family, int argc, char *argv[],init_extensions4 ();#endif tables = xtables_ipv4; + d.commit = true;break ;case NFPROTO_ARP : tables = xtables_arp;
Signature · 1 line
2.22.0
Florian Westphal
maintainerRe: [iptables PATCH 09/12] xtables-save: Make COMMIT line optional
Quoted 2 lines · Phil Sutter
Change
xtables_save_mainto support optional printing of COMMIT line as it is not used in arp- or ebtables.
Why? Is this so ebt-save dumps are compatible with the old ebt-restore?
Phil Sutter
reviewerRe: [iptables PATCH 09/12] xtables-save: Make COMMIT line optional
Quoted 6 lines · Florian Westphal
Quoted 2 lines · Phil Sutter
Change
xtables_save_mainto support optional printing of COMMIT line as it is not used in arp- or ebtables.Why? Is this so ebt-save dumps are compatible with the old ebt-restore?
Hmm. I haven't considered deviating in ebtables-nft-save format from legacy one yet, but it may actually be no problem given that people shouldn't switch between both. The only bummer might be if someone has a custom script which then trips over the unknown non-comment line.
I'm undecided, but not against "breaking" with legacy save dumps per se. If you think it's not a problem, we could just simplify things. OTOH this is not a big patch, either so maybe not even worth the risk. :)
Thanks, Phil
Florian Westphal
maintainerRe: [iptables PATCH 09/12] xtables-save: Make COMMIT line optional
Quoted 8 lines · Phil Sutter
Quoted 6 lines · Florian Westphal
Quoted 2 lines · Phil Sutter
Change
xtables_save_mainto support optional printing of COMMIT line as it is not used in arp- or ebtables.Why? Is this so ebt-save dumps are compatible with the old ebt-restore?
Quoted 8 lines
Hmm. I haven't considered deviating in ebtables-nft-save format from legacy one yet, but it may actually be no problem given that people shouldn't switch between both. The only bummer might be if someone has a custom script which then trips over the unknown non-comment line.
I'm undecided, but not against "breaking" with legacy save dumps per se. If you think it's not a problem, we could just simplify things. OTOH this is not a big patch, either so maybe not even worth the risk. :)
I'm fine with this, I was just wondering because the changelog explains whats happening, not why.
Phil Sutter
reviewerRe: [iptables PATCH 09/12] xtables-save: Make COMMIT line optional
Quoted 14 lines · Florian Westphal
Quoted 13 lines · Phil Sutter
Quoted 6 lines · Florian Westphal
Quoted 2 lines · Phil Sutter
Change
xtables_save_mainto support optional printing of COMMIT line as it is not used in arp- or ebtables.Why? Is this so ebt-save dumps are compatible with the old ebt-restore?
Hmm. I haven't considered deviating in ebtables-nft-save format from legacy one yet, but it may actually be no problem given that people shouldn't switch between both. The only bummer might be if someone has a custom script which then trips over the unknown non-comment line.
Quoted 6 lines
Quoted 3 lines
I'm undecided, but not against "breaking" with legacy save dumps per se. If you think it's not a problem, we could just simplify things. OTOH this is not a big patch, either so maybe not even worth the risk. :)
I'm fine with this, I was just wondering because the changelog explains whats happening, not why.
Yes, typical pitfall for me. Maybe I should put a sticky note somewhere. :)
Thanks, Phil
Phil Sutter
reviewerPATCH 10/12 xtables-save: Pass format flags to do_output()
In reply to Phil SutterLet callers define the flags to pass to nft_rule_save() instead of just setting the counters boolean.
--- iptables/xtables-save.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
iptables/xtables-save.c +6 -4
@@ -66,7 +66,7 @@ static const struct option ebt_save_options[] = {static bool ebt_legacy_counter_format;struct do_output_data { -bool counters; +unsigned int format;bool commit; };
@@ -98,7 +98,7 @@ __do_output(struct nft_handle *h, const char *tablename, void *data)/* Dump out chain names first, * thereby preventing dependency conflicts */ nft_chain_save (h, chain_list); -nft_rule_save (h, tablename, d->counters ?0 :FMT_NOCOUNTS ); +nft_rule_save (h, tablename, d->format);if (d->commit)printf ("COMMIT\n" );
@@ -139,7 +139,9 @@ xtables_save_main(int family, int argc, char *argv[], {const struct builtin_table *tables;const char *tablename =NULL ; -struct do_output_data d = {}; +struct do_output_data d = { + .format =FMT_NOCOUNTS , + };bool dump = false;struct nft_handle h = { .family = family,
@@ -162,7 +164,7 @@ xtables_save_main(int family, int argc, char *argv[],fprintf (stderr,"-b/--binary option is not implemented\n" );break ;case 'c' : - d.counters = true; + d.format &= ~FMT_NOCOUNTS ;break ;case 't' :
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 11/12 arptables-save: Merge into xtables_save_main()
In reply to Phil SutterWith all preparations in place, xtables_save_main() can replace it with not further changes.
--- iptables/xtables-save.c | 63 ++--------------------------------------- 1 file changed, 3 insertions(+), 60 deletions(-)
iptables/xtables-save.c +3 -60
@@ -372,65 +372,8 @@ int xtables_eb_save_main(int argc_, char *argv_[])return 0 ; } -int xtables_arp_save_main (int argc,char **argv) +int xtables_arp_save_main (int argc,char *argv[]) { -struct nft_handle h = { - .family =NFPROTO_ARP , - }; -time_t now; -int c; - - xtables_globals.program_name =basename (*argv);; - c =xtables_init_all (&xtables_globals, h.family); -if (c <0 ) { -fprintf (stderr,"%s/%s Failed to initialize xtables\n" , - xtables_globals.program_name, - xtables_globals.program_version); -exit (1 ); - } - -while ((c =getopt_long (argc, argv, arp_save_optstring, arp_save_options,NULL )) !=-1 ) { -switch (c) { -case 'c' : - show_counters = true; -break ; -case 'M' : - xtables_modprobe_program = optarg; -break ; -case 'V' : -printf ("%s v%s (nf_tables)\n" , prog_name, prog_vers); -exit (0 ); -default : -fprintf (stderr, -"Look at manual page `%s.8' for more information.\n" , - prog_name); -exit (1 ); - } - } - -if (nft_init (&h, xtables_arp) <0 ) { -fprintf (stderr,"%s/%s Failed to initialize nft: %s\n" , - xtables_globals.program_name, - xtables_globals.program_version, -strerror (errno)); -exit (EXIT_FAILURE ); - } - -if (!nft_table_find (&h,"filter" )) -return 0 ; - -if (!nft_is_table_compatible (&h,"filter" )) { -printf ("# Table `filter' is incompatible, use 'nft' tool.\n" ); -return 0 ; - } - -printf ("# Generated by %s v%s on %s" , prog_name, - prog_vers,ctime (&now)); -printf ("*filter\n" ); -nft_chain_save (&h,nft_chain_list_get (&h,"filter" )); -nft_rule_save (&h,"filter" , show_counters ?0 :FMT_NOCOUNTS ); - now =time (NULL ); -printf ("# Completed on %s" ,ctime (&now)); -nft_fini (&h); -return 0 ; +return xtables_save_main (NFPROTO_ARP , argc, argv, + arp_save_optstring, arp_save_options); }
Signature · 1 line
2.22.0
Phil Sutter
reviewerPATCH 12/12 ebtables-save: Merge into xtables_save_main()
In reply to Phil SutterThe only thing missing was handling of EBTABLES_SAVE_COUNTER env var, but that can be done after parsing parameters in bridge-specific code.
--- iptables/xtables-save.c | 123 +++++----------------------------------- 1 file changed, 13 insertions(+), 110 deletions(-)
iptables/xtables-save.c +13 -110
@@ -31,8 +31,6 @@#define prog_name xtables_globals.program_name#define prog_vers xtables_globals.program_version -static bool show_counters = false; -static const char *ipt_save_optstring ="bcdt:M:f:46V" ;static const struct option ipt_save_options[] = { {.name ="counters" , .has_arg = false, .val ='c' },
@@ -63,8 +61,6 @@ static const struct option ebt_save_options[] = { {NULL}, }; -static bool ebt_legacy_counter_format; - struct do_output_data { unsigned int format; bool commit;
@@ -228,9 +224,18 @@ xtables_save_main(int family, int argc, char *argv[],case NFPROTO_ARP : tables = xtables_arp;break ; -case NFPROTO_BRIDGE : +case NFPROTO_BRIDGE : { +const char *ctr =getenv ("EBTABLES_SAVE_COUNTER" ); + +if (!(d.format &FMT_NOCOUNTS )) { + d.format |=FMT_EBT_SAVE ; + }else if (ctr && !strcmp (ctr,"yes" )) { + d.format &= ~FMT_NOCOUNTS ; + d.format |=FMT_C_COUNTS |FMT_EBT_SAVE ; + } tables = xtables_bridge;break ; + }default :fprintf (stderr,"Unknown family %d\n" , family);return 1 ;
@@ -264,112 +269,10 @@ int xtables_ip6_save_main(int argc, char *argv[]) ipt_save_optstring, ipt_save_options); } -static int __ebt_save (struct nft_handle *h,const char *tablename,void *data) +int xtables_eb_save_main (int argc,char *argv[]) { -struct nftnl_chain_list *chain_list; -unsigned int format =FMT_NOCOUNTS ; -bool *counters = data; -time_t now; - -if (!nft_table_find (h, tablename)) { -printf ("Table `%s' does not exist\n" , tablename); -return 1 ; - } - -if (!nft_is_table_compatible (h, tablename)) { -printf ("# Table `%s' is incompatible, use 'nft' tool.\n" , tablename); -return 0 ; - } - - chain_list =nft_chain_list_get (h, tablename); - - now =time (NULL ); -printf ("# Generated by %s v%s on %s" , prog_name, - prog_vers,ctime (&now)); -printf ("*%s\n" , tablename); - -if (counters) - format =FMT_EBT_SAVE | - (ebt_legacy_counter_format ?FMT_C_COUNTS :0 ); - -/* Dump out chain names first, -* thereby preventing dependency conflicts */ -nft_chain_save (h, chain_list); -nft_rule_save (h, tablename, format); - now =time (NULL ); -printf ("# Completed on %s" ,ctime (&now)); -return 0 ; -} - -static int ebt_save (struct nft_handle *h,const char *tablename,bool counters) -{ -if (!tablename) -return nft_for_each_table (h, __ebt_save, &counters); - -return __ebt_save (h, tablename, &counters); -} - -int xtables_eb_save_main (int argc_,char *argv_[]) -{ -const char *ctr =getenv ("EBTABLES_SAVE_COUNTER" ); -const char *tablename =NULL ; -struct nft_handle h = { - .family =NFPROTO_BRIDGE , - }; -int c; - -if (ctr) { -if (strcmp (ctr,"yes" ) ==0 ) { - ebt_legacy_counter_format = true; - show_counters = true; - } - } - - xtables_globals.program_name =basename (*argv_); - c =xtables_init_all (&xtables_globals, h.family); -if (c <0 ) { -fprintf (stderr,"%s/%s Failed to initialize xtables\n" , - xtables_globals.program_name, - xtables_globals.program_version); -exit (1 ); - } - -while ((c =getopt_long (argc_, argv_, ebt_save_optstring, ebt_save_options,NULL )) !=-1 ) { -switch (c) { -case 'c' : -unsetenv ("EBTABLES_SAVE_COUNTER" ); - show_counters = true; - ebt_legacy_counter_format = false; -break ; -case 't' : -/* Select specific table. */ - tablename = optarg; -break ; -case 'M' : - xtables_modprobe_program = optarg; -break ; -case 'V' : -printf ("%s v%s (nf_tables)\n" , prog_name, prog_vers); -exit (0 ); -default : -fprintf (stderr, -"Look at manual page `%s.8' for more information.\n" , - prog_name); -exit (1 ); - } - } - -if (nft_init (&h, xtables_bridge) <0 ) { -fprintf (stderr,"%s/%s Failed to initialize nft: %s\n" , - xtables_globals.program_name, - xtables_globals.program_version, -strerror (errno)); -exit (EXIT_FAILURE ); - } - -ebt_save (&h, tablename, show_counters); -nft_fini (&h); -return 0 ; +return xtables_save_main (NFPROTO_BRIDGE , argc, argv, + ebt_save_optstring, ebt_save_options); }int xtables_arp_save_main (int argc,char *argv[])
Signature · 1 line
2.22.0
Florian Westphal
maintainerQuoted 17 lines · Phil Sutter
This series started as a fix to program names mentioned in *-save outputs and ended in merging ebtables-save and arptables-save code into
xtables_save_mainused by ip{6,}tables-nft-save.The first patch is actually unrelated but was discovered when testing counter output - depending on environment, ebtables-nft might segfault.
The second patch fixes option '-c' of ebtables-nft-save which enables counter prefixes in dumped rules but failed to disable the classical ebtables-style counters.
Patch three sorts program names quoted in output of any of the *-save programs, patch four unifies the header/footer comments in the same. The latter also drops the extra newline printed in ebtables- and arptables-save output, so test scripts need adjustments beyond dropping the new comment lines from output.
Quoted 8 lines
Patch five fixes the table compatibility check in ip{6,}tables-nft-save.
Patches six and eight to ten prepare for integrating arptables- and ebtables-save into the xtables-save code.
Patch seven merely fixes a minor coding-style issue.
Patches eleven and twelve finally perform the actual merge.
Looks good, feel free to rebase this on top of master and then you can push this out.
In case my comment wrt. 'COMMIT line optional' is right, consider ammending the commit message so that this reasoning is recorded in the changelog.
Thanks!
Pablo Neira Ayuso
maintainerQuoted 17 lines · Phil Sutter
This series started as a fix to program names mentioned in *-save outputs and ended in merging ebtables-save and arptables-save code into
xtables_save_mainused by ip{6,}tables-nft-save.The first patch is actually unrelated but was discovered when testing counter output - depending on environment, ebtables-nft might segfault.
The second patch fixes option '-c' of ebtables-nft-save which enables counter prefixes in dumped rules but failed to disable the classical ebtables-style counters.
Patch three sorts program names quoted in output of any of the *-save programs, patch four unifies the header/footer comments in the same. The latter also drops the extra newline printed in ebtables- and arptables-save output, so test scripts need adjustments beyond dropping the new comment lines from output.
Quoted 8 lines
Patch five fixes the table compatibility check in ip{6,}tables-nft-save.
Patches six and eight to ten prepare for integrating arptables- and ebtables-save into the xtables-save code.
Patch seven merely fixes a minor coding-style issue.
Patches eleven and twelve finally perform the actual merge.
Series look good after quick review.