net: fix pos incrementment in ipv6_route_seq_next
Subsystems (4)
PATCH v2 1/1 net: fix pos incrementment in ipv6_route_seq_next
Commit 4fc427e05158 ("ipv6_route_seq_next should increase position index") tried to fix the issue where seq_file pos is not increased
if a NULL element is returned with seq_ops->next(). See bug
https://bugzilla.kernel.org/show_bug.cgi?id=206283
The commit effectively does:
- increase pos for all seq_ops->start()
- increase pos for all seq_ops->next()
For ipv6_route, increasing pos for all seq_ops->next() is correct. But increasing pos for seq_ops->start() is not correct since pos is used to determine how many items to skip during seq_ops->start():
iter->skip = *pos;seq_ops->start() just fetches the *current* pos item.
The item can be skipped only after seq_ops->show() which essentially is the beginning of seq_ops->next().
For example, I have 7 ipv6 route entries,
root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=4096
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
fe800000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
00000000000000000000000000000001 80 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000003 00000000 80200001 lo
fe800000000000002050e3fffebd3be8 80 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
ff000000000000000000000000000000 08 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000004 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
0+1 records in
0+1 records out
1050 bytes (1.0 kB, 1.0 KiB) copied, 0.00707908 s, 148 kB/s root@arch-fb-vm1:~/net-next
In the above, I specify buffer size 4096, so all records can be returned to user space with a single trip to the kernel.
If I use buffer size 128, since each record size is 149, internally kernel seq_read() will read 149 into its internal buffer and return the data to user space in two read() syscalls. Then user read() syscall will trigger next seq_ops->start(). Since the current implementation increased pos even for seq_ops->start(), it will skip record #2, #4 and #6, assuming the first record is #1.
root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=128
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
fe800000000000002050e3fffebd3be8 80 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
4+1 records in
4+1 records out
600 bytes copied, 0.00127758 s, 470 kB/s
To fix the problem, create a fake pos pointer so seq_ops->start() won't actually increase seq_file pos. With this fix, the above dd command with bs=128 will show correct result.
--- net/ipv6/ip6_fib.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
Changelog:
v1 -> v2:
- instead of push increment of *pos in ipv6_route_seq_next() for seq_ops->next() only. Add a face pos pointer in seq_ops->start() and use it when calling ipv6_route_seq_next().
net/ipv6/ip6_fib.c +3 -1
@@ -2622,8 +2622,10 @@ static void *ipv6_route_seq_start(struct seq_file *seq, loff_t *pos) iter->skip = *pos;if (iter->tbl) { +loff_t p; +ipv6_route_seq_setup_walk (iter, net); -return ipv6_route_seq_next (seq,NULL , pos); +return ipv6_route_seq_next (seq,NULL , &p); }else {return NULL ; }
Signature · 1 line
2.24.1
Quoted 54 lines · Yonghong Song
Commit
4fc427e05158("ipv6_route_seq_nextshould increase position index") tried to fix the issue whereseq_filepos is not increased
if a NULL element is returned withseq_ops->next(). See bug
https://bugzilla.kernel.org/show_bug.cgi?id=206283
The commit effectively does:
- increase pos for allseq_ops->start()
- increase pos for allseq_ops->next()For
ipv6_route, increasing pos for allseq_ops->next()is correct. But increasing pos forseq_ops->start()is not correct since pos is used to determine how many items to skip duringseq_ops->start():
iter->skip= *pos;seq_ops->start()just fetches the *current* pos item.
The item can be skipped only afterseq_ops->show()which essentially is the beginning ofseq_ops->next().For example, I have 7 ipv6 route entries,
root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=4096
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
fe80000000000000000000000000000040 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
00000000000000000000000000000001 80 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000003 00000000 80200001 lo
fe800000000000002050e3fffebd3be880 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
ff00000000000000000000000000000008 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000004 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
0+1 records in
0+1 records out
1050 bytes (1.0 kB, 1.0 KiB) copied, 0.00707908 s, 148 kB/s root@arch-fb-vm1:~/net-nextIn the above, I specify buffer size 4096, so all records can be returned to user space with a single trip to the kernel.
If I use buffer size 128, since each record size is 149, internally kernel
seq_read()will read 149 into its internal buffer and return the data to user space in tworead()syscalls. Then userread()syscall will trigger nextseq_ops->start(). Since the current implementation increased pos even forseq_ops->start(), it will skip record #2, #4 and #6, assuming the first record is #1.root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=128
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
fe800000000000002050e3fffebd3be880 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
4+1 records in
4+1 records out
600 bytes copied, 0.00127758 s, 470 kB/sTo fix the problem, create a fake pos pointer so
seq_ops->start()won't actually increaseseq_filepos. With this fix, the aboveddcommand withbs=128will show correct result.
Quoted 6 lines
Fixes 4fc427e05158 ("ipv6_route_seq_next should increase position index")Cc Vasily Averin <[email protected]>Cc Andrii Nakryiko <[email protected]>Cc Alexei Starovoitov <[email protected]>Suggested-by Vasily Averin <[email protected]>Signed-off-by Yonghong Song <[email protected]>
Quoted 27 lines
--- net/ipv6/ip6_fib.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)Changelog:
v1 -> v2:
- instead of push increment of *pos inipv6_route_seq_next()forseq_ops->next()only. Add a face pos pointer inseq_ops->start()and use it when callingipv6_route_seq_next().net/ipv6/ip6_fib.c +3
-1@@ -2622,8 +2622,10 @@ static void *ipv6_route_seq_start(struct seq_file *seq, loff_t *pos) iter->skip = *pos;if (iter->tbl) { +loff_t p; +ipv6_route_seq_setup_walk (iter, net); -return ipv6_route_seq_next (seq,NULL , pos); +return ipv6_route_seq_next (seq,NULL , &p); }else {return NULL ; }
Quoted 75 lines · Yonghong Song
Commit
4fc427e05158("ipv6_route_seq_nextshould increase position index") tried to fix the issue whereseq_filepos is not increased
if a NULL element is returned withseq_ops->next(). See bug
https://bugzilla.kernel.org/show_bug.cgi?id=206283
The commit effectively does:
- increase pos for allseq_ops->start()
- increase pos for allseq_ops->next()For
ipv6_route, increasing pos for allseq_ops->next()is correct. But increasing pos forseq_ops->start()is not correct since pos is used to determine how many items to skip duringseq_ops->start():
iter->skip= *pos;seq_ops->start()just fetches the *current* pos item.
The item can be skipped only afterseq_ops->show()which essentially is the beginning ofseq_ops->next().For example, I have 7 ipv6 route entries,
root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=4096
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
fe80000000000000000000000000000040 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
00000000000000000000000000000001 80 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000003 00000000 80200001 lo
fe800000000000002050e3fffebd3be880 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
ff00000000000000000000000000000008 00000000000000000000000000000000 00 00000000000000000000000000000000 00000100 00000004 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
0+1 records in
0+1 records out
1050 bytes (1.0 kB, 1.0 KiB) copied, 0.00707908 s, 148 kB/s root@arch-fb-vm1:~/net-nextIn the above, I specify buffer size 4096, so all records can be returned to user space with a single trip to the kernel.
If I use buffer size 128, since each record size is 149, internally kernel
seq_read()will read 149 into its internal buffer and return the data to user space in tworead()syscalls. Then userread()syscall will trigger nextseq_ops->start(). Since the current implementation increased pos even forseq_ops->start(), it will skip record #2, #4 and #6, assuming the first record is #1.root@arch-fb-vm1:~/net-next dd if=/proc/net/ipv6_route bs=128
00000000000000000000000000000000 40 00000000000000000000000000000000 00 00000000000000000000000000000000 00000400 00000001 00000000 00000001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
fe800000000000002050e3fffebd3be880 00000000000000000000000000000000 00 00000000000000000000000000000000 00000000 00000002 00000000 80200001 eth0
00000000000000000000000000000000 00 00000000000000000000000000000000 00 00000000000000000000000000000000 ffffffff 00000001 00000000 00200200 lo
4+1 records in
4+1 records out
600 bytes copied, 0.00127758 s, 470 kB/sTo fix the problem, create a fake pos pointer so
seq_ops->start()won't actually increaseseq_filepos. With this fix, the aboveddcommand withbs=128will show correct result.Fixes 4fc427e05158 ("ipv6_route_seq_next should increase position index")Cc Vasily Averin <[email protected]>Cc Andrii Nakryiko <[email protected]>Cc Alexei Starovoitov <[email protected]>Suggested-by Vasily Averin <[email protected]>Signed-off-by Yonghong Song <[email protected]>--- net/ipv6/ip6_fib.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)Changelog:
v1 -> v2:
- instead of push increment of *pos inipv6_route_seq_next()forseq_ops->next()only. Add a face pos pointer inseq_ops->start()and use it when callingipv6_route_seq_next().net/ipv6/ip6_fib.c +0
-0
net/ipv6/ip6_fib.c static void *ipv6_route_seq_start(struct seq_file *seq, loff_t *pos) · line 2625
Quoted 4 lines
net/ipv6/ip6_fib.c +1
-0@@ -2622,8 +2622,10 @@ static void *ipv6_route_seq_start(struct seq_file *seq, loff_t *pos) iter->skip = *pos;if (iter->tbl) { +loff_t p;
Please init this, otherwise I can guarantee syzbot will be not happy.
p = *pos;
Quoted 8 lines
net/ipv6/ip6_fib.c +2
-1+ipv6_route_seq_setup_walk (iter, net); -return ipv6_route_seq_next (seq,NULL , pos); +return ipv6_route_seq_next (seq,NULL , &p); }else {return NULL ; }