| Message ID | 20200922225341.8976-4-phil@nwl.cc |
|---|---|
| State | Accepted |
| Delegated to: | Pablo Neira |
| Headers | show |
| Series | libxtables: Fix for pointless socket() calls | expand |
On Wed, Sep 23, 2020 at 12:53:41AM +0200, Phil Sutter wrote: > The newly introduced ordered insert algorithm in > xtables_register_{match,target}() works best if extensions of same name > are passed in ascending revisions. Since this is the case in about all > extensions' arrays, iterate over them from beginning to end. This patch should come first in the series, my understanding is that 1/3 assumes that extensions are registered from lower to higher revision number. > Signed-off-by: Phil Sutter <phil@nwl.cc> > --- > libxtables/xtables.c | 14 ++++++++------ > 1 file changed, 8 insertions(+), 6 deletions(-) > > diff --git a/libxtables/xtables.c b/libxtables/xtables.c > index de74d361a53af..90b1195c45a58 100644 > --- a/libxtables/xtables.c > +++ b/libxtables/xtables.c > @@ -1139,9 +1139,10 @@ static bool xtables_fully_register_pending_match(struct xtables_match *me, > > void xtables_register_matches(struct xtables_match *match, unsigned int n) > { > - do { > - xtables_register_match(&match[--n]); > - } while (n > 0); > + int i; > + > + for (i = 0; i < n; i++) > + xtables_register_match(&match[i]); > } > > void xtables_register_target(struct xtables_target *me) > @@ -1264,9 +1265,10 @@ static bool xtables_fully_register_pending_target(struct xtables_target *me, > > void xtables_register_targets(struct xtables_target *target, unsigned int n) > { > - do { > - xtables_register_target(&target[--n]); > - } while (n > 0); > + int i; > + > + for (i = 0; i < n; i++) > + xtables_register_target(&target[i]); > } > > /* receives a list of xtables_rule_match, release them */ > -- > 2.28.0 >
On Tue, Oct 06, 2020 at 01:41:21AM +0200, Pablo Neira Ayuso wrote: > On Wed, Sep 23, 2020 at 12:53:41AM +0200, Phil Sutter wrote: > > The newly introduced ordered insert algorithm in > > xtables_register_{match,target}() works best if extensions of same name > > are passed in ascending revisions. Since this is the case in about all > > extensions' arrays, iterate over them from beginning to end. > > This patch should come first in the series, my understanding is that > 1/3 assumes that extensions are registered from lower to higher > revision number. No, the algorithm is supposed to work with arbitrary input. This is merely an optimization given how extension arrays are typically ordered. Cheers, Phil
diff --git a/libxtables/xtables.c b/libxtables/xtables.c index de74d361a53af..90b1195c45a58 100644 --- a/libxtables/xtables.c +++ b/libxtables/xtables.c @@ -1139,9 +1139,10 @@ static bool xtables_fully_register_pending_match(struct xtables_match *me, void xtables_register_matches(struct xtables_match *match, unsigned int n) { - do { - xtables_register_match(&match[--n]); - } while (n > 0); + int i; + + for (i = 0; i < n; i++) + xtables_register_match(&match[i]); } void xtables_register_target(struct xtables_target *me) @@ -1264,9 +1265,10 @@ static bool xtables_fully_register_pending_target(struct xtables_target *me, void xtables_register_targets(struct xtables_target *target, unsigned int n) { - do { - xtables_register_target(&target[--n]); - } while (n > 0); + int i; + + for (i = 0; i < n; i++) + xtables_register_target(&target[i]); } /* receives a list of xtables_rule_match, release them */
The newly introduced ordered insert algorithm in xtables_register_{match,target}() works best if extensions of same name are passed in ascending revisions. Since this is the case in about all extensions' arrays, iterate over them from beginning to end. Signed-off-by: Phil Sutter <phil@nwl.cc> --- libxtables/xtables.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)