| Message ID | 20191108213257.3097633-10-arnd@arndb.de |
|---|---|
| State | Accepted |
| Commit | 6408c40c39d8eee5caaf97f5219b7dd4e041cc59 |
| Headers | show |
| Series | [01/16] staging: exfat: use prandom_u32() for i_generation | expand |
On Fri, Nov 08, 2019 at 10:32:47PM +0100, Arnd Bergmann wrote: > On 32-bit architectures, get_seconds() returns an unsigned 32-bit > time value, which also matches the type used in the nft_meta > code. This will not overflow in year 2038 as a time_t would, but > it still suffers from the overflow problem later on in year 2106. I wonder if the assumption that people will still use nft_meta 80 years from now is an optimistic or pessimistic one. :) > Change this instance to use the time64_t type consistently > and avoid the deprecated get_seconds(). > > The nft_meta_weekday() calculation potentially gets a little slower > on 32-bit architectures, but now it has the same behavior as on > 64-bit architectures and does not overflow. > > Fixes: 63d10e12b00d ("netfilter: nft_meta: support for time matching") > Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Phil Sutter <phil@nwl.cc>
On Fri, Nov 08, 2019 at 10:32:47PM +0100, Arnd Bergmann wrote: > On 32-bit architectures, get_seconds() returns an unsigned 32-bit > time value, which also matches the type used in the nft_meta > code. This will not overflow in year 2038 as a time_t would, but > it still suffers from the overflow problem later on in year 2106. > > Change this instance to use the time64_t type consistently > and avoid the deprecated get_seconds(). > > The nft_meta_weekday() calculation potentially gets a little slower > on 32-bit architectures, but now it has the same behavior as on > 64-bit architectures and does not overflow. Applied, thanks Arnd.
diff --git a/net/netfilter/nft_meta.c b/net/netfilter/nft_meta.c index 317e3a9e8c5b..dda1e55d5801 100644 --- a/net/netfilter/nft_meta.c +++ b/net/netfilter/nft_meta.c @@ -33,19 +33,19 @@ static DEFINE_PER_CPU(struct rnd_state, nft_prandom_state); -static u8 nft_meta_weekday(unsigned long secs) +static u8 nft_meta_weekday(time64_t secs) { unsigned int dse; u8 wday; secs -= NFT_META_SECS_PER_MINUTE * sys_tz.tz_minuteswest; - dse = secs / NFT_META_SECS_PER_DAY; + dse = div_u64(secs, NFT_META_SECS_PER_DAY); wday = (4 + dse) % NFT_META_DAYS_PER_WEEK; return wday; } -static u32 nft_meta_hour(unsigned long secs) +static u32 nft_meta_hour(time64_t secs) { struct tm tm; @@ -250,10 +250,10 @@ void nft_meta_get_eval(const struct nft_expr *expr, nft_reg_store64(dest, ktime_get_real_ns()); break; case NFT_META_TIME_DAY: - nft_reg_store8(dest, nft_meta_weekday(get_seconds())); + nft_reg_store8(dest, nft_meta_weekday(ktime_get_real_seconds())); break; case NFT_META_TIME_HOUR: - *dest = nft_meta_hour(get_seconds()); + *dest = nft_meta_hour(ktime_get_real_seconds()); break; default: WARN_ON(1);
On 32-bit architectures, get_seconds() returns an unsigned 32-bit time value, which also matches the type used in the nft_meta code. This will not overflow in year 2038 as a time_t would, but it still suffers from the overflow problem later on in year 2106. Change this instance to use the time64_t type consistently and avoid the deprecated get_seconds(). The nft_meta_weekday() calculation potentially gets a little slower on 32-bit architectures, but now it has the same behavior as on 64-bit architectures and does not overflow. Fixes: 63d10e12b00d ("netfilter: nft_meta: support for time matching") Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- net/netfilter/nft_meta.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) -- 2.20.0