1435fabc19 Use RdSeed when available, and reduce RdRand load (Pieter Wuille)
Pull request description:
This introduces support for autodetecting and using the RdSeed instruction on x86/x86_64 systems.
In addition:
* In SeedFast, only 64 bits of entropy are generated through RdRand (256 was relatively slow).
* In SeedStartup, 256 bits of entropy are generated, using RdSeed (preferably) or RdRand (otherwise).
Tree-SHA512: fb7d3e22e93e14592f4b07282aa79d7c3cc4e9debdd9978580b8d2562bbad345e289bf3f80de2c50c9b50b8bac2aa9b838f9f272f7f8d43f1efc0913aa8acce3
This introduces support for autodetecting and using the RdSeed instruction.
In addition:
* In SeedFast, only 64 bits of entropy are generated through RdRand (256 was relatively slow).
* In SeedStartup, 256 bits of entropy are generated, using RdSeed (preferably) or RdRand (otherwise).
It includes the following policy changes:
* All GetRand* functions seed the stack pointer and rdrand result
(in addition to the performance counter)
* The periodic entropy added by the idle scheduler now seeds stack pointer,
rdrand and perfmon data (once every 10 minutes) in addition to
just a sleep timing.
* The entropy added when calling GetStrongRandBytes no longer includes
the once-per-10-minutes perfmon data on windows (it is moved to the
idle scheduler instead, where latency matters less).
Other changes:
* OpenSSL is no longer seeded directly anywhere. Instead, any generated
randomness through our own RNG is fed back to OpenSSL (after an
additional hashing step to prevent leaking our RNG state).
* Seeding that was previously done directly in RandAddSeedSleep is now
moved to SeedSleep(), which is indirectly invoked through ProcRand
from RandAddSeedSleep.
* Seeding that was previously done directly in GetStrongRandBytes()
is now moved to SeedSlow(), which is indirectly invoked through
ProcRand from GetStrongRandBytes().
This guarantees that OpenSSL is initialized properly whenever randomness
is used, even when that randomness is invoked from global constructors.
Note that this patch uses Mutex directly, rather than CCriticalSection.
This is because the lock-detection code is not necessarily initialized
during global constructors.
Call sync.h primitives "locks" and "mutexes" instead of "blocks" and "waitable
critical sections" to match current coding conventions and c++11 standard
names.
This PR does not rename the "CCriticalSection" class (though this could be done
as a followup) because it is used everywhere and would swamp the other changes
in this PR. Plain mutexes should mostly be preferred instead of recursive
mutexes in new code anyway.
-BEGIN VERIFY SCRIPT-
set -x
set -e
ren() { git grep -l $1 | xargs sed -i s/$1/$2/; }
ren CCriticalBlock UniqueLock
ren CWaitableCriticalSection Mutex
ren CConditionVariable std::condition_variable
ren cs_GenesisWait g_genesis_wait_mutex
ren condvar_GenesisWait g_genesis_wait_cv
perl -0777 -pi -e 's/.*typedef.*condition_variable.*\n\n?//g' src/sync.h
-END VERIFY SCRIPT-
9c4dc597dd Use LOCK macros for non-recursive locks (Russell Yanofsky)
1382913e61 Make LOCK, LOCK2, TRY_LOCK work with CWaitableCriticalSection (Russell Yanofsky)
ba1f095aad MOVEONLY Move AnnotatedMixin declaration (Russell Yanofsky)
41b88e9337 Add unit test for DEBUG_LOCKORDER code (Russell Yanofsky)
Pull request description:
Make LOCK macros work with non-recursive mutexes, and use wherever possible for better deadlock detection.
Also add unit test for DEBUG_LOCKORDER code.
Tree-SHA512: 64ef209307f28ecd0813a283f15c6406138c6ffe7f6cbbd084161044db60e2c099a7d0d2edcd1c5e7770a115e9b931b486e86c9a777bdc96d2e8a9f4dc192942
b82c55a Add attribute [[noreturn]] (C++11) to functions that will not return (practicalswift)
Pull request description:
Add attribute `[[noreturn]]` (C++11) to functions that will not return.
Rationale:
* Reduce the number of false positives/false negatives from static analyzers with regards to things such as unused or unreachable code
* Potentially enable additional compiler optimizations
Tree-SHA512: 899683fe8b2fcf19bd334352271d368b46b805be9d426aac1808335fd95732d6d7078d3296951b9879196f3f6e3ec0fdb7695d0afdc3fbe4dd78a2ca70e91ff7
90d4d89 scripted-diff: Use the C++11 keyword nullptr to denote the pointer literal instead of the macro NULL (practicalswift)
Pull request description:
Since C++11 the macro `NULL` may be:
* an integer literal with value zero, or
* a prvalue of type `std::nullptr_t`
By using the C++11 keyword `nullptr` we are guaranteed a prvalue of type `std::nullptr_t`.
For a more thorough discussion, see "A name for the null pointer: nullptr" (Sutter &
Stroustrup), http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2431.pdf
With this patch applied there are no `NULL` macro usages left in the repo:
```
$ git grep NULL -- "*.cpp" "*.h" | egrep -v '(/univalue/|/secp256k1/|/leveldb/|_NULL|NULLDUMMY|torcontrol.*NULL|NULL cert)' | wc -l
0
```
The road towards `nullptr` (C++11) is split into two PRs:
* `NULL` → `nullptr` is handled in PR #10483 (scripted, this PR)
* `0` → `nullptr` is handled in PR #10645 (manual)
Tree-SHA512: 3c395d66f2ad724a8e6fed74b93634de8bfc0c0eafac94e64e5194c939499fefd6e68f047de3083ad0b4eff37df9a8a3a76349aa17d55eabbd8e0412f140a297
rbx needs to be stashed in a 64bit register on 64bit platforms. With this crash
in particular, it was holding a stack canary which was not properly restored
after the cpuid.
Split out the x86+PIC case so that x86_64 doesn't have to worry about it.
cb24c85 Use rdrand as entropy source on supported platforms (Pieter Wuille)
Tree-SHA512: c42eaa01a14e6bc097c70b6bf8540d61854c2f76cb32be69c2a3c411a126f7b4bf4a4486e4493c4cc367cc689319abde0d4adb799d29a54fd3e81767ce0766fc
97477c5 Maintain state across GetStrongRandBytes calls (Pieter Wuille)
Tree-SHA512: 77e9b1f3c6eeb0c2a3e0c64358150767222ff0b7120ccd5f4ae0276cea0e4fa275c1b757e3f20be07dc0b4ef07f70ab0b70112080c8d3d0cb6ed703db8a59168
