On Mon, Apr 25, 2022 at 08:07:48AM -0400, Mikulas Patocka wrote:
> From: Mikulas Patocka <[email protected]>
>
> The function hex2bin is used to load cryptographic keys into device mapper
> targets dm-crypt and dm-integrity. It should take constant time
> independent on the processed data, so that concurrently running
> unprivileged code can't infer any information about the keys via
> microarchitectural convert channels.
>
> This patch changes the function hex_to_bin so that it contains no branches
> and no memory accesses.
>
> Note that this shouldn't cause performance degradation because the size of
> the new function is the same as the size of the old function (on x86-64) -
> and the new function causes no branch misprediction penalties.
>
> I compile-tested this function with gcc on aarch64 alpha arm hppa hppa64
> i386 ia64 m68k mips32 mips64 powerpc powerpc64 riscv sh4 s390x sparc32
> sparc64 x86_64 and with clang on aarch64 arm hexagon i386 mips32 mips64
> powerpc powerpc64 s390x sparc32 sparc64 x86_64 to verify that there are no
> branches in the generated code.
>
> Signed-off-by: Mikulas Patocka <[email protected]>
> Cc: [email protected]
>
> ---
> include/linux/kernel.h | 2 +-
> lib/hexdump.c | 32 +++++++++++++++++++++++++-------
> 2 files changed, 26 insertions(+), 8 deletions(-)
>
> Index: linux-2.6/lib/hexdump.c
> ===================================================================
> --- linux-2.6.orig/lib/hexdump.c 2022-04-24 18:51:20.000000000 +0200
> +++ linux-2.6/lib/hexdump.c 2022-04-25 13:15:26.000000000 +0200
> @@ -22,15 +22,33 @@ EXPORT_SYMBOL(hex_asc_upper);
> *
> * hex_to_bin() converts one hex digit to its actual value or -1 in case of bad
> * input.
> + *
> + * This function is used to load cryptographic keys, so it is coded in such a
> + * way that there are no conditions or memory accesses that depend on data.
> + *
> + * Explanation of the logic:
> + * (ch - '9' - 1) is negative if ch <= '9'
> + * ('0' - 1 - ch) is negative if ch >= '0'
> + * we "and" these two values, so the result is negative if ch is in the range
> + * '0' ... '9'
> + * we are only interested in the sign, so we do a shift ">> 8"; note that right
> + * shift of a negative value is implementation-defined, so we cast the
> + * value to (unsigned) before the shift --- we have 0xffffff if ch is in
> + * the range '0' ... '9', 0 otherwise
> + * we "and" this value with (ch - '0' + 1) --- we have a value 1 ... 10 if ch is
> + * in the range '0' ... '9', 0 otherwise
> + * we add this value to -1 --- we have a value 0 ... 9 if ch is in the range '0'
> + * ... '9', -1 otherwise
> + * the next line is similar to the previous one, but we need to decode both
> + * uppercase and lowercase letters, so we use (ch & 0xdf), which converts
> + * lowercase to uppercase
> */
> -int hex_to_bin(char ch)
> +int hex_to_bin(unsigned char ch)
> {
> - if ((ch >= '0') && (ch <= '9'))
> - return ch - '0';
> - ch = tolower(ch);
> - if ((ch >= 'a') && (ch <= 'f'))
> - return ch - 'a' + 10;
> - return -1;
> + unsigned char cu = ch & 0xdf;
> + return -1 +
> + ((ch - '0' + 1) & (unsigned)((ch - '9' - 1) & ('0' - 1 - ch)) >> 8) +
> + ((cu - 'A' + 11) & (unsigned)((cu - 'F' - 1) & ('A' - 1 - cu)) >> 8);
> }
> EXPORT_SYMBOL(hex_to_bin);
Hello,
Just a heads up it seems this patch is causing some instability with crypto self
tests on OpenRISC when using a PREEMPT kernel (no SMP).
This was reported by Jason A. Donenfeld as it came up in wireguard testing.
I am trying to figure out if this is an OpenRISC PREEMPT issue or something
else.
The warning I am seeing is a bit random but looks something like the
following:
[ 0.164000] Freeing initrd memory: 1696K
[ 0.188000] xor: measuring software checksum speed
[ 0.196000] 8regs : 1343 MB/sec
[ 0.204000] 8regs_prefetch : 1347 MB/sec
[ 0.212000] 32regs : 1335 MB/sec
[ 0.220000] 32regs_prefetch : 1277 MB/sec
[ 0.220000] xor: using function: 8regs_prefetch (1347 MB/sec)
[ 0.252000] SARU running 25519 tests
[ 0.424000] curve25519 self-test 5: FAIL
[ 0.496000] curve25519 self-test 7: FAIL
[ 1.744000] curve25519 self-test 45: FAIL
[ 3.448000] ------------[ cut here ]------------
[ 3.448000] WARNING: CPU: 0 PID: 1 at lib/crypto/curve25519.c:19 curve25519_init+0x38/0x50
[ 3.448000] CPU: 0 PID: 1 Comm: swapper Not tainted 5.18.0-rc4+ #758
[ 3.448000] Call trace:
[ 3.448000] [<(ptrval)>] ? __warn+0xe0/0x114
[ 3.448000] [<(ptrval)>] ? curve25519_init+0x38/0x50
[ 3.448000] [<(ptrval)>] ? warn_slowpath_fmt+0x5c/0x94
[ 3.448000] [<(ptrval)>] ? curve25519_init+0x0/0x50
[ 3.452000] [<(ptrval)>] ? curve25519_init+0x38/0x50
[ 3.452000] [<(ptrval)>] ? do_one_initcall+0x98/0x328
[ 3.452000] [<(ptrval)>] ? proc_register+0x4c/0x284
[ 3.452000] [<(ptrval)>] ? ignore_unknown_bootoption+0x0/0x8
[ 3.452000] [<(ptrval)>] ? kernel_init_freeable+0x1fc/0x2a8
[ 3.452000] [<(ptrval)>] ? ignore_unknown_bootoption+0x0/0x8
[ 3.452000] [<(ptrval)>] ? kernel_init+0x0/0x164
[ 3.452000] [<(ptrval)>] ? kernel_init+0x28/0x164
[ 3.452000] [<(ptrval)>] ? schedule_tail+0x18/0xac
[ 3.452000] [<(ptrval)>] ? ret_from_fork+0x1c/0x9c
[ 3.452000] ---[ end trace 0000000000000000 ]---
[ 3.452000] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
[ 3.464000] printk: console [ttyS0] disabled
[ 3.464000] 90000000.serial: ttyS0 at MMIO 0x90000000 (irq = 2, base_baud = 1250000) is a 16550A
Example config: https://xn--4db.cc/cCRlQ1AE
The self-test iteration number that fails is always a bit different. I am
still in progress of investigating this and will not have a lot of time new the
next few days. If anything ring's a bell let me know.
-Stafford
On Wed, 4 May 2022, Stafford Horne wrote:
> On Mon, Apr 25, 2022 at 08:07:48AM -0400, Mikulas Patocka wrote:
>
> > From: Mikulas Patocka <[email protected]>
> >
> > -int hex_to_bin(char ch)
> > +int hex_to_bin(unsigned char ch)
> > {
> > - if ((ch >= '0') && (ch <= '9'))
> > - return ch - '0';
> > - ch = tolower(ch);
> > - if ((ch >= 'a') && (ch <= 'f'))
> > - return ch - 'a' + 10;
> > - return -1;
> > + unsigned char cu = ch & 0xdf;
> > + return -1 +
> > + ((ch - '0' + 1) & (unsigned)((ch - '9' - 1) & ('0' - 1 - ch)) >> 8) +
> > + ((cu - 'A' + 11) & (unsigned)((cu - 'F' - 1) & ('A' - 1 - cu)) >> 8);
> > }
> > EXPORT_SYMBOL(hex_to_bin);
>
> Hello,
>
> Just a heads up it seems this patch is causing some instability with crypto self
> tests on OpenRISC when using a PREEMPT kernel (no SMP).
>
> This was reported by Jason A. Donenfeld as it came up in wireguard testing.
>
> I am trying to figure out if this is an OpenRISC PREEMPT issue or something
> else.
Hi
That patch is so simple that I can't imagine how could it break the
curve25519 test. Are you sure that you bisected it correctly?
Mikulas
On Wed, May 04, 2022 at 04:57:35AM -0400, Mikulas Patocka wrote:
> On Wed, 4 May 2022, Stafford Horne wrote:
> > On Mon, Apr 25, 2022 at 08:07:48AM -0400, Mikulas Patocka wrote:
...
> > Just a heads up it seems this patch is causing some instability with crypto self
> > tests on OpenRISC when using a PREEMPT kernel (no SMP).
> >
> > This was reported by Jason A. Donenfeld as it came up in wireguard testing.
> >
> > I am trying to figure out if this is an OpenRISC PREEMPT issue or something
> > else.
> That patch is so simple that I can't imagine how could it break the
> curve25519 test. Are you sure that you bisected it correctly?
Can you provide a test cases for hex_to_bin()?
--
With Best Regards,
Andy Shevchenko