-
Notifications
You must be signed in to change notification settings - Fork 87
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Handle synchronization on struct page #211
Comments
xairy
referenced
this issue
in google/ktsan
Jan 31, 2019
Fixed function sk_msg_clone() to prevent overflow of 'dst' while adding pages in scatterlist entries. The overflow of 'dst' causes crash in kernel tls module while doing record encryption. Crash fixed by this patch. [ 78.796119] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000008 [ 78.804900] Mem abort info: [ 78.807683] ESR = 0x96000004 [ 78.810744] Exception class = DABT (current EL), IL = 32 bits [ 78.816677] SET = 0, FnV = 0 [ 78.819727] EA = 0, S1PTW = 0 [ 78.822873] Data abort info: [ 78.825759] ISV = 0, ISS = 0x00000004 [ 78.829600] CM = 0, WnR = 0 [ 78.832576] user pgtable: 4k pages, 48-bit VAs, pgdp = 00000000bf8ee311 [ 78.839195] [0000000000000008] pgd=0000000000000000 [ 78.844081] Internal error: Oops: 96000004 [#1] PREEMPT SMP [ 78.849642] Modules linked in: tls xt_conntrack ipt_REJECT nf_reject_ipv4 ip6table_filter ip6_tables xt_CHECKSUM cpve cpufreq_conservative lm90 ina2xx crct10dif_ce [ 78.865377] CPU: 0 PID: 6007 Comm: openssl Not tainted 4.20.0-rc6-01647-g754d5da63145-dirty #107 [ 78.874149] Hardware name: LS1043A RDB Board (DT) [ 78.878844] pstate: 60000005 (nZCv daif -PAN -UAO) [ 78.883632] pc : scatterwalk_copychunks+0x164/0x1c8 [ 78.888500] lr : scatterwalk_copychunks+0x160/0x1c8 [ 78.893366] sp : ffff00001d04b600 [ 78.896668] x29: ffff00001d04b600 x28: ffff80006814c680 [ 78.901970] x27: 0000000000000000 x26: ffff80006c8de786 [ 78.907272] x25: ffff00001d04b760 x24: 000000000000001a [ 78.912573] x23: 0000000000000006 x22: ffff80006814e440 [ 78.917874] x21: 0000000000000100 x20: 0000000000000000 [ 78.923175] x19: 000081ffffffffff x18: 0000000000000400 [ 78.928476] x17: 0000000000000008 x16: 0000000000000000 [ 78.933778] x15: 0000000000000100 x14: 0000000000000001 [ 78.939079] x13: 0000000000001080 x12: 0000000000000020 [ 78.944381] x11: 0000000000001080 x10: 00000000ffff0002 [ 78.949683] x9 : ffff80006814c248 x8 : 00000000ffff0000 [ 78.954985] x7 : ffff80006814c318 x6 : ffff80006c8de786 [ 78.960286] x5 : 0000000000000f80 x4 : ffff80006c8de000 [ 78.965588] x3 : 0000000000000000 x2 : 0000000000001086 [ 78.970889] x1 : ffff7e0001b74e02 x0 : 0000000000000000 [ 78.976192] Process openssl (pid: 6007, stack limit = 0x00000000291367f9) [ 78.982968] Call trace: [ 78.985406] scatterwalk_copychunks+0x164/0x1c8 [ 78.989927] skcipher_walk_next+0x28c/0x448 [ 78.994099] skcipher_walk_done+0xfc/0x258 [ 78.998187] gcm_encrypt+0x434/0x4c0 [ 79.001758] tls_push_record+0x354/0xa58 [tls] [ 79.006194] bpf_exec_tx_verdict+0x1e4/0x3e8 [tls] [ 79.010978] tls_sw_sendmsg+0x650/0x780 [tls] [ 79.015326] inet_sendmsg+0x2c/0xf8 [ 79.018806] sock_sendmsg+0x18/0x30 [ 79.022284] __sys_sendto+0x104/0x138 [ 79.025935] __arm64_sys_sendto+0x24/0x30 [ 79.029936] el0_svc_common+0x60/0xe8 [ 79.033588] el0_svc_handler+0x2c/0x80 [ 79.037327] el0_svc+0x8/0xc [ 79.040200] Code: 6b01005f 54fff788 940169b1 f9000320 (b9400801) [ 79.046283] ---[ end trace 74db007d069c1cf7 ]--- Fixes: d829e9c ("tls: convert to generic sk_msg interface") Signed-off-by: Vakul Garg <[email protected]> Acked-by: John Fastabend <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We've removed all synchronization on page struct because it lead to enormous memory consumption for vector clocks (there were millions of sync objects). Initially we've not noticed false positives.
But here is one:
https://groups.google.com/forum/#!topic/ktsan/m1lJug5oAIg
(see the explanation by Theodore that the code is actually synchronized on struct page->flags).
There are hundreds of uses of lock_page and trylock_page throughout kernel code, so we need to restore synchronization on struct page. Otherwise we will be constantly hitting false positives.
The text was updated successfully, but these errors were encountered: