Add fallback LE/BE for architectures with known endianness + SHA256 selftest #799

pull sipa wants to merge 2 commits into bitcoin-core:master from sipa:202008_add_lebe changing 4 files +54 −6
  1. sipa commented at 6:51 pm on August 14, 2020: contributor

    These are all the architecture macros I could find with known endianness. Use those as a fallback when BYTE_ORDER isn’t available.

    See #787 (comment)

    It also adds a SHA256 selftest, so that improperly overriding the endianness detection will be detected at runtime.

  2. sipa cross-referenced this on Aug 14, 2020 from issue Use preprocessor macros instead of autoconf to detect endianness by real-or-random
  3. gmaxwell commented at 6:54 pm on August 14, 2020: contributor
    Great work! testing now.
  4. gmaxwell commented at 8:13 pm on August 14, 2020: contributor
    I can confirm that this works on the systems that were broken for me before (debian + GCC 2.95 and 32-bit haiku). I’m installing a mips (be) host right now to check what macros it has.
  5. real-or-random commented at 10:28 pm on August 14, 2020: contributor
    This could be helpful: https://github.com/rofl0r/endianness.h/blob/master/endianness.h#L52 but we should not overdo it I think .
  6. sipa force-pushed on Aug 15, 2020
  7. sipa commented at 2:46 am on August 15, 2020: contributor
    @real-or-random Might as well use something that has had other eyes on it. I’ve switched to it.
  8. sipa force-pushed on Aug 15, 2020
  9. luke-jr approved
  10. luke-jr commented at 2:55 am on August 15, 2020: member
    utACK
  11. gmaxwell commented at 5:34 am on August 15, 2020: contributor

    ACK 7fde5566a91ffc7fee4606a0a7b6dc6fb5b47444

    looks fine, works on the weird things I’ve tested– hakiu, gcc 2.95 i386, openwatcom, debian mips.

  12. in src/util.h:180 in 7fde5566a9 outdated
    185@@ -186,6 +186,26 @@ static SECP256K1_INLINE void *manual_alloc(void** prealloc_ptr, size_t alloc_siz
    186 #if defined(_MSC_VER) && defined(_WIN32) && !defined(SECP256K1_LITTLE_ENDIAN)
    187 # define SECP256K1_LITTLE_ENDIAN
    188 #endif
    189+#if !defined(SECP256K1_LITTLE_ENDIAN) && !defined(SECP256K1_BIG_ENDIAN)
    


    real-or-random commented at 10:17 am on August 15, 2020:
    Do we need this #if ? It breaks with the idea of the code above, which tries to detect from all available sources instead and then fail below if both are defined, e.g., if a user supplied #define is inconsistent with our detection. Seems a tiny bit safer.

    gmaxwell commented at 3:49 pm on August 15, 2020:
    Hm? if the detection is wrong (e.g. because a system header sets something it shouldn’t) the user should be able to override.

    real-or-random commented at 6:30 pm on August 15, 2020:
    For example if the user wrongly sets SECP256K1_BIG_ENDIAN, then __x86_64__ will not be checked at all due to this line (if I’m not getting it wrong entirely).

    gmaxwell commented at 6:36 pm on August 15, 2020:

    That is my understanding of the behaviour. If some dumb header has x86_64 in it why shouldn’t the user should be able to override?

    I’ve seen stuff set things like that just to make stuff compile when some code was erroneously testing i386 to determine if the processor was 32-bit.


    real-or-random commented at 6:46 pm on August 15, 2020:

    Ok, we can discuss this but it should be consistent at least. The existing code is different, see my commit message in #787.

    The macros are carefully written to err on the side of the caution, e.g., we #error if the user manually configures a different endianness than what we detect.


    gmaxwell commented at 7:28 pm on August 15, 2020:
    What you’re suggesting would override the BYTE_ORDER macro. I don’t think it makes any sense at all to have arch guessing override explicitlyly set BYTE_ORDER or the user. There is a difference between a conflicting setting and a guess conflicting with something explicit.

    sipa commented at 8:50 pm on August 15, 2020:

    So really we have 3 potentially different sources:

    • User manually setting SECP256K1_{LITTLE,BIG}_ENDIAN
    • __BYTE_ORDER__
    • The stuff I’m adding in this PR

    Should all 3 be consistent with each other, if they’re available? That means no mechanism is left for the user to override if the detection is somehow wrong. Should we have a USE_FORCE_{BIG,LITTLE}_ENDIAN for that case?

    One possibility is just using SECP256K1_{BIG,LITTLE}_ENDIAN. If exactly one of those two is set, that’s what is used. If not, all other sources that are found have to agree.

  13. sipa force-pushed on Aug 16, 2020
  14. sipa commented at 2:20 am on August 16, 2020: contributor

    I’ve changed the approach here to only check the system macros in case SECP256K1_{BIG,LITTLE}_ENDIAN aren’t explicitly set, so it’s possible to override a possible incorrect detection. I’ve also merged all the macros into one check for each of LE and BE.

    In GCC/Clang we could always do the detection, and give a warning if it doesn’t match an explicit SECP256K1_{BIG,LITTLE}_ENDIAN, but not all compilers have a way of doing so in a standard way.

  15. Use additional system macros to figure out endianness
    Also permit it being overridden by explicitly passing SECP256K1_{BIG,LITTLE}_ENDIAN
    5e5fb28b4a
  16. sipa force-pushed on Aug 16, 2020
  17. real-or-random commented at 9:41 am on August 16, 2020: contributor

    I’m ok with this but then we should really add a SHA256 self-test or something similar to context creation. For example, one scenario that concerns me is the user provides a custom nonce derivation function for Schnorr signing and for whatever reason uses the wrong endianness.

    Then the endianness affects the challenge hash but possibly not the nonce, so this could mean that two different challenge hashes are signed for the same nonce.

  18. gmaxwell commented at 4:47 pm on August 16, 2020: contributor

    Self tests are good. There could be a context creation argument to disable it if you really didn’t want it.

    How would you imagine the schnorr hash would change but the nonce wouldn’t? maybe that’s something that could be specifically fought against.

  19. real-or-random commented at 5:07 pm on August 16, 2020: contributor

    How would you imagine the schnorr hash would change but the nonce wouldn’t?

    As I wrote above, this could happen with a user-provided nonce function which just doesn’t care about the endianness assumptions in our code. But that’s the only scenario that comes to my mind.

  20. sipa renamed this:
    Add fallback LE/BE for architectures with known endianness
    Add fallback LE/BE for architectures with known endianness + SHA256 selftest
    on Aug 17, 2020
  21. sipa commented at 8:51 pm on August 17, 2020: contributor
    I’ve added a single SHA256 selftest, as that should be enough to detect an incorrectly-configured endianness. Other selftests could be added later.
  22. in src/secp256k1.c:152 in 84d2df1237 outdated
    145@@ -144,8 +146,12 @@ secp256k1_context* secp256k1_context_preallocated_create(void* prealloc, unsigne
    146 }
    147 
    148 secp256k1_context* secp256k1_context_create(unsigned int flags) {
    149-    size_t const prealloc_size = secp256k1_context_preallocated_size(flags);
    150-    secp256k1_context* ctx = (secp256k1_context*)checked_malloc(&default_error_callback, prealloc_size);
    151+    size_t prealloc_size;
    152+    secp256k1_context* ctx;
    153+
    154+    CHECK(secp256k1_selftest());
    


    real-or-random commented at 9:16 am on August 18, 2020:
    CHECK is intended to be used in the tests only. It calls abort() (hardcoded) instead of the error callback, so can’t even override this on compile time.

    sipa commented at 6:20 pm on August 21, 2020:
    Indeed, fixed.
  23. in src/secp256k1.c:155 in 84d2df1237 outdated
    151+    size_t prealloc_size;
    152+    secp256k1_context* ctx;
    153+
    154+    CHECK(secp256k1_selftest());
    155+    prealloc_size = secp256k1_context_preallocated_size(flags);
    156+    ctx = (secp256k1_context*)checked_malloc(&default_error_callback, prealloc_size);
    


    real-or-random commented at 9:19 am on August 18, 2020:

    I think this is what we should do, namely call default_error_callback, which can at least be overridden at compile time. We do this here because we can’t call the error callback unfortunately. It’s just not yet determined at context creation time. I’m going to add a note to #780, maybe we can improve this in the future.

    Another advantage of the (default) error callback is that we can provide a more meaningful message, which may even hint at endianness.


    sipa commented at 6:22 pm on August 21, 2020:
    Replaced it with a call to default_error_callback, and tested that it works by incorrectly forcing a build with SECP256K1_BIG_ENDIAN.
  24. real-or-random commented at 9:21 am on August 18, 2020: contributor

    Very neat!

    I grepped for CHECK and apparently we violate this rule in one more place, namely when checking computation of an inverse by gmp. I guess noone noticed this so far simply because if you don’t have abort() on an embedded system, you’ll probably not want to use gmp anyway. We should fix this one too (in a different PR) and then probably move CHECK out of util.h.

    edit: Hm now I vaguely recall having that discussion before but I can’t find it. I think we didn’t want to pass the callback down to the arithmetic functions. Maybe the problem just disappears if we can get rid of GMP. :shrug:

  25. sipa force-pushed on Aug 21, 2020
  26. sipa commented at 6:22 pm on August 21, 2020: contributor
    @real-or-random My preference is getting rid of libgmp :)
  27. gmaxwell commented at 11:52 pm on August 21, 2020: contributor

    I guess noone noticed this so far simply because if you don’t have abort() on an embedded system

    It was intentional. Search IRC logs. :) I believe our thinking was that a clean crash in an “impossible” state is preferable to giving a wrong result… returning an error isn’t useful because the caller can’t reasonable be expected to handle impossible errors.

    It might still be preferable to keep a sanity check on the inversion after getting rid of libgmp: it’s absurdly cheap and any fast inversion algorithm is fairly complicated. I think if every piece of secp256k1 computation could be checked as relatively cheaply as inversion it might make sense to add additional checks but alas few to nothing else is so cheap. to check. :P

  28. real-or-random commented at 9:01 am on August 22, 2020: contributor

    I guess noone noticed this so far simply because if you don’t have abort() on an embedded system

    It was intentional. Search IRC logs. :) I believe our thinking was that a clean crash in an “impossible” state is preferable to giving a wrong result… returning an error isn’t useful because the caller can’t reasonable be expected to handle impossible errors.

    I don’t understand. Both calling abort() and the error callback are clean crashes. What I’m saying is that your platform may not have abort(), so you can’t even compile our gmp inversion wrapper there.

  29. in src/secp256k1.c:154 in a468d55baa outdated
    147@@ -144,8 +148,14 @@ secp256k1_context* secp256k1_context_preallocated_create(void* prealloc, unsigne
    148 }
    149 
    150 secp256k1_context* secp256k1_context_create(unsigned int flags) {
    151-    size_t const prealloc_size = secp256k1_context_preallocated_size(flags);
    152-    secp256k1_context* ctx = (secp256k1_context*)checked_malloc(&default_error_callback, prealloc_size);
    153+    size_t prealloc_size;
    154+    secp256k1_context* ctx;
    155+
    156+    if (!secp256k1_selftest()) {
    


    real-or-random commented at 9:30 am on August 22, 2020:
    This function later calls secp256k1_context_preallocated_create, which already calls the self-tests, so we don’t need to call it again here. (Unless we’re worried that future additions to the selftest should be run before secp256k1_context_preallocated_size or checked_malloc but assume this is not what you had in mind here.)

    sipa commented at 11:34 pm on August 22, 2020:
    No, indeed. I missed that secp256k1_context_create was called secp256k1_context_preallocated_create. Fixed.
  30. sipa force-pushed on Aug 22, 2020
  31. Add SHA256 selftest 8bc6aeffa9
  32. sipa force-pushed on Aug 22, 2020
  33. real-or-random commented at 9:00 am on August 23, 2020: contributor
    ACK 8bc6aeffa9a191e677cb9e3a22fff130f16990f3 I read the diff, and tested that the self-test passes/fails with/without the correct endianness setting
  34. real-or-random approved
  35. jonasnick commented at 1:03 pm on August 23, 2020: contributor

    It also adds a SHA256 selftest, so that improperly overriding the endianness detection will be detected at runtime.

    I don’t understand the rationale behind the selftest. Improper endianness setting would also be detected by the regular tests. Or is an endianness self test useful because it can be difficult to port the tests to the target platform (I know one example where this was the case)?

  36. gmaxwell commented at 1:14 pm on August 23, 2020: contributor

    ACK 8bc6aeffa9a191e677cb9e3a22fff130f16990f3 looks good and I also ran the tests on MIPS-BE and verified that forcing it to LE makes the runtime test fail. @jonasnick right now the tests are hard to make run on any low memory target. That should be improved but runtime tests have been discussed for a while– the tests may never get run by the particular target– e.g. probably not anywhere where the build system isn’t used, probably not by 99% of bitcoin users that compile their own bitcoin, and they test a different binary (due to the VERIFY_CHECK macros) which may hide miscompilation in the real binary. If sha256’s implementation is made developer-substitutable errors by the library user that could potentially cause silent bad nonce generation… etc.

    What I’d suggested before is to have a context creation argument to bypass them if you really must (e.g. rapid context creation) but otherwise, some simple runtime sanity checks seem prudent.

  37. elichai commented at 8:42 am on August 26, 2020: contributor

    @jonasnick right now the tests are hard to make run on any low memory target.

    Is this because of heap allocation? too much stack allocation? something else? if you could specify the requirements/how can I check it myself I can start working on it if it’s feasible to fix without adding too much complexity

  38. real-or-random cross-referenced this on Aug 26, 2020 from issue Tests on embedded platforms by real-or-random
  39. sipa commented at 2:38 am on September 9, 2020: contributor
    I think this is ready.
  40. jonasnick commented at 1:18 pm on September 9, 2020: contributor
    concept ACK
  41. real-or-random merged this on Sep 9, 2020
  42. real-or-random closed this on Sep 9, 2020

  43. real-or-random cross-referenced this on Sep 9, 2020 from issue Document / assert non-portable assumptions by real-or-random
  44. deadalnix referenced this in commit cb45b773f9 on Oct 22, 2020
  45. deadalnix referenced this in commit 141c8d0dd2 on Oct 23, 2020

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin-core/secp256k1. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2024-11-24 16:15 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me