Re: [PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Nov 7, 2013, at 4:26 PM, David Turner wrote: > On Fri, 2013-11-08 at 00:14 +0100, Jan Kara wrote: >> Still unnecessary type cast here (but that's a cosmetic issue). > ... >> Otherwise the patch looks good. You can add: >> Reviewed-by: Jan Kara > > Thanks. A version with this correction and the reviewed-by follows. Thanks for working on this. It was (seriously) in my list of things to get done, but I figured I still had a few years to work on it... My (unfinished) version of this patch had a nice comment at ext4_encode_time() that explained the encoding that is being used very clearly: /* * We need is an encoding that preserves the times for extra epoch "00": * * extra msb of adjust for signed * epoch 32-bit 32-bit tv_sec to * bits timedecoded 64-bit tv_sec 64-bit tv_sec valid time range * 0 01-0x8000..-0x0001 0x0 1901-12-13..1969-12-31 * 0 000x0..0x07fff 0x0 1970-01-01..2038-01-19 * 0 110x08000..0x0 0x1 2038-01-19..2106-02-07 * 0 100x1..0x17fff 0x1 2106-02-07..2174-02-25 * 1 010x18000..0x1 0x2 2174-02-25..2242-03-16 * 1 000x2..0x27fff 0x2 2242-03-16..2310-04-04 * 1 110x28000..0x2 0x3 2310-04-04..2378-04-22 * 1 100x3..0x37fff 0x3 2378-04-22..2446-05-10 */ It seems that your version of the patch seems to use a different encoding. Not that this is a problem, per-se, since my patch wasn’t in use anywhere, but it would be nice to have a similarly clear explanation of what the mapping is so that it can be clearly understood. My ext4_{encode,decode}_extra_time() used add/subtract instead of mask/OR ops, which changed the on-disk format for times beyond 2038, but those were already broken, and presumably not in use by anyone yet. However, it seemed to me this was easier to produce the correct results. Have you tested your patch with a variety of timestamps to verify its correctness? It looks to me like you have mapped the 1901-1969 range onto 0x3 for the epoch bits, instead of the (IMHO) natural 0x0 bits. The critical time ranges are listed above. Cheers, Andreas > -- > In ext4, the bottom two bits of {a,c,m}time_extra are used to extend > the {a,c,m}time fields, deferring the year 2038 problem to the year > 2446. The representation (which this patch does not alter) is a bit > hackish, in that the most-significant bit is no longer (alone) > sufficient to indicate the sign. That's because we're representing an > asymmetric range, with seven times as many positive values as > negative. > > When decoding these extended fields, for times whose bottom 32 bits > would represent a negative number, sign extension causes the 64-bit > extended timestamp to be negative as well, which is not what's > intended. This patch corrects that issue, so that the only negative > {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed > timestamps). > > Signed-off-by: David Turner > Reported-by: Mark Harris > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 > Reviewed-by: Jan Kara > --- > fs/ext4/ext4.h | 13 + > 1 file changed, 9 insertions(+), 4 deletions(-) > > diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h > index af815ea..3c2d0b3 100644 > --- a/fs/ext4/ext4.h > +++ b/fs/ext4/ext4.h > @@ -722,10 +722,15 @@ static inline __le32 ext4_encode_extra_time(struct > timespec *time) > > static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra) > { > - if (sizeof(time->tv_sec) > 4) > -time->tv_sec |= (__u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK) > -<< 32; > - time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> > EXT4_EPOCH_BITS; > + if (sizeof(time->tv_sec) > 4) { > + u64 extra_bits = le32_to_cpu(extra) & EXT4_EPOCH_MASK; > + if (time->tv_sec > 0 || extra_bits != EXT4_EPOCH_MASK) { > + time->tv_sec &= 0x; > + time->tv_sec |= extra_bits << 32; > + } > + } > + time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> > + EXT4_EPOCH_BITS; > } > > #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode)\ > -- > 1.8.1.2 > > > Cheers, Andreas -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Nov 7, 2013, at 4:26 PM, David Turner nova...@novalis.org wrote: On Fri, 2013-11-08 at 00:14 +0100, Jan Kara wrote: Still unnecessary type cast here (but that's a cosmetic issue). ... Otherwise the patch looks good. You can add: Reviewed-by: Jan Kara j...@suse.cz Thanks. A version with this correction and the reviewed-by follows. Thanks for working on this. It was (seriously) in my list of things to get done, but I figured I still had a few years to work on it... My (unfinished) version of this patch had a nice comment at ext4_encode_time() that explained the encoding that is being used very clearly: /* * We need is an encoding that preserves the times for extra epoch 00: * * extra msb of adjust for signed * epoch 32-bit 32-bit tv_sec to * bits timedecoded 64-bit tv_sec 64-bit tv_sec valid time range * 0 01-0x8000..-0x0001 0x0 1901-12-13..1969-12-31 * 0 000x0..0x07fff 0x0 1970-01-01..2038-01-19 * 0 110x08000..0x0 0x1 2038-01-19..2106-02-07 * 0 100x1..0x17fff 0x1 2106-02-07..2174-02-25 * 1 010x18000..0x1 0x2 2174-02-25..2242-03-16 * 1 000x2..0x27fff 0x2 2242-03-16..2310-04-04 * 1 110x28000..0x2 0x3 2310-04-04..2378-04-22 * 1 100x3..0x37fff 0x3 2378-04-22..2446-05-10 */ It seems that your version of the patch seems to use a different encoding. Not that this is a problem, per-se, since my patch wasn’t in use anywhere, but it would be nice to have a similarly clear explanation of what the mapping is so that it can be clearly understood. My ext4_{encode,decode}_extra_time() used add/subtract instead of mask/OR ops, which changed the on-disk format for times beyond 2038, but those were already broken, and presumably not in use by anyone yet. However, it seemed to me this was easier to produce the correct results. Have you tested your patch with a variety of timestamps to verify its correctness? It looks to me like you have mapped the 1901-1969 range onto 0x3 for the epoch bits, instead of the (IMHO) natural 0x0 bits. The critical time ranges are listed above. Cheers, Andreas -- In ext4, the bottom two bits of {a,c,m}time_extra are used to extend the {a,c,m}time fields, deferring the year 2038 problem to the year 2446. The representation (which this patch does not alter) is a bit hackish, in that the most-significant bit is no longer (alone) sufficient to indicate the sign. That's because we're representing an asymmetric range, with seven times as many positive values as negative. When decoding these extended fields, for times whose bottom 32 bits would represent a negative number, sign extension causes the 64-bit extended timestamp to be negative as well, which is not what's intended. This patch corrects that issue, so that the only negative {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed timestamps). Signed-off-by: David Turner nova...@novalis.org Reported-by: Mark Harris mh8...@yahoo.com Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 Reviewed-by: Jan Kara j...@suse.cz --- fs/ext4/ext4.h | 13 + 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index af815ea..3c2d0b3 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -722,10 +722,15 @@ static inline __le32 ext4_encode_extra_time(struct timespec *time) static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra) { - if (sizeof(time-tv_sec) 4) -time-tv_sec |= (__u64)(le32_to_cpu(extra) EXT4_EPOCH_MASK) - 32; - time-tv_nsec = (le32_to_cpu(extra) EXT4_NSEC_MASK) EXT4_EPOCH_BITS; + if (sizeof(time-tv_sec) 4) { + u64 extra_bits = le32_to_cpu(extra) EXT4_EPOCH_MASK; + if (time-tv_sec 0 || extra_bits != EXT4_EPOCH_MASK) { + time-tv_sec = 0x; + time-tv_sec |= extra_bits 32; + } + } + time-tv_nsec = (le32_to_cpu(extra) EXT4_NSEC_MASK) + EXT4_EPOCH_BITS; } #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode)\ -- 1.8.1.2 Cheers, Andreas -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Thu, Nov 07, 2013 at 06:26:47PM -0500, David Turner wrote: > In ext4, the bottom two bits of {a,c,m}time_extra are used to extend > the {a,c,m}time fields, deferring the year 2038 problem to the year > 2446. The representation (which this patch does not alter) is a bit > hackish, in that the most-significant bit is no longer (alone) > sufficient to indicate the sign. That's because we're representing an > asymmetric range, with seven times as many positive values as > negative. > > When decoding these extended fields, for times whose bottom 32 bits > would represent a negative number, sign extension causes the 64-bit > extended timestamp to be negative as well, which is not what's > intended. This patch corrects that issue, so that the only negative > {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed > timestamps). > > Signed-off-by: David Turner > Reported-by: Mark Harris > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 > Reviewed-by: Jan Kara Thanks, applied. - Ted -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Fri, 2013-11-08 at 00:14 +0100, Jan Kara wrote: > Still unnecessary type cast here (but that's a cosmetic issue). ... > Otherwise the patch looks good. You can add: > Reviewed-by: Jan Kara Thanks. A version with this correction and the reviewed-by follows. -- In ext4, the bottom two bits of {a,c,m}time_extra are used to extend the {a,c,m}time fields, deferring the year 2038 problem to the year 2446. The representation (which this patch does not alter) is a bit hackish, in that the most-significant bit is no longer (alone) sufficient to indicate the sign. That's because we're representing an asymmetric range, with seven times as many positive values as negative. When decoding these extended fields, for times whose bottom 32 bits would represent a negative number, sign extension causes the 64-bit extended timestamp to be negative as well, which is not what's intended. This patch corrects that issue, so that the only negative {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed timestamps). Signed-off-by: David Turner Reported-by: Mark Harris Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 Reviewed-by: Jan Kara --- fs/ext4/ext4.h | 13 + 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index af815ea..3c2d0b3 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -722,10 +722,15 @@ static inline __le32 ext4_encode_extra_time(struct timespec *time) static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra) { - if (sizeof(time->tv_sec) > 4) - time->tv_sec |= (__u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK) - << 32; - time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> EXT4_EPOCH_BITS; + if (sizeof(time->tv_sec) > 4) { + u64 extra_bits = le32_to_cpu(extra) & EXT4_EPOCH_MASK; + if (time->tv_sec > 0 || extra_bits != EXT4_EPOCH_MASK) { + time->tv_sec &= 0x; + time->tv_sec |= extra_bits << 32; + } + } + time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> + EXT4_EPOCH_BITS; } #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \ -- 1.8.1.2 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Fri, 2013-11-08 at 00:14 +0100, Jan Kara wrote: Still unnecessary type cast here (but that's a cosmetic issue). ... Otherwise the patch looks good. You can add: Reviewed-by: Jan Kara j...@suse.cz Thanks. A version with this correction and the reviewed-by follows. -- In ext4, the bottom two bits of {a,c,m}time_extra are used to extend the {a,c,m}time fields, deferring the year 2038 problem to the year 2446. The representation (which this patch does not alter) is a bit hackish, in that the most-significant bit is no longer (alone) sufficient to indicate the sign. That's because we're representing an asymmetric range, with seven times as many positive values as negative. When decoding these extended fields, for times whose bottom 32 bits would represent a negative number, sign extension causes the 64-bit extended timestamp to be negative as well, which is not what's intended. This patch corrects that issue, so that the only negative {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed timestamps). Signed-off-by: David Turner nova...@novalis.org Reported-by: Mark Harris mh8...@yahoo.com Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 Reviewed-by: Jan Kara j...@suse.cz --- fs/ext4/ext4.h | 13 + 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index af815ea..3c2d0b3 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -722,10 +722,15 @@ static inline __le32 ext4_encode_extra_time(struct timespec *time) static inline void ext4_decode_extra_time(struct timespec *time, __le32 extra) { - if (sizeof(time-tv_sec) 4) - time-tv_sec |= (__u64)(le32_to_cpu(extra) EXT4_EPOCH_MASK) - 32; - time-tv_nsec = (le32_to_cpu(extra) EXT4_NSEC_MASK) EXT4_EPOCH_BITS; + if (sizeof(time-tv_sec) 4) { + u64 extra_bits = le32_to_cpu(extra) EXT4_EPOCH_MASK; + if (time-tv_sec 0 || extra_bits != EXT4_EPOCH_MASK) { + time-tv_sec = 0x; + time-tv_sec |= extra_bits 32; + } + } + time-tv_nsec = (le32_to_cpu(extra) EXT4_NSEC_MASK) + EXT4_EPOCH_BITS; } #define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \ -- 1.8.1.2 -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3] ext4: Fix reading of extended tv_sec (bug 23732)
On Thu, Nov 07, 2013 at 06:26:47PM -0500, David Turner wrote: In ext4, the bottom two bits of {a,c,m}time_extra are used to extend the {a,c,m}time fields, deferring the year 2038 problem to the year 2446. The representation (which this patch does not alter) is a bit hackish, in that the most-significant bit is no longer (alone) sufficient to indicate the sign. That's because we're representing an asymmetric range, with seven times as many positive values as negative. When decoding these extended fields, for times whose bottom 32 bits would represent a negative number, sign extension causes the 64-bit extended timestamp to be negative as well, which is not what's intended. This patch corrects that issue, so that the only negative {a,c,m}times are those between 1901 and 1970 (as per 32-bit signed timestamps). Signed-off-by: David Turner nova...@novalis.org Reported-by: Mark Harris mh8...@yahoo.com Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=23732 Reviewed-by: Jan Kara j...@suse.cz Thanks, applied. - Ted -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/