| Commit message (Collapse) | Author | Age |
|\
| |
| | |
Fix alignment issue in gpr_murmur_hash3
|
|\ \
| | |
| | | |
Bump 1.8.0-dev to 1.9.0-dev
|
|\ \ \
| | | |
| | | | |
Pop lockfree stack off core implementation
|
| |_|/
|/| |
| | |
| | |
| | |
| | | |
This function cast a void* to a uint32_t*. This is invalid, since a uint32_t* must be 32-bit-aligned, while the input key clearly isn't. Even though the function later uses memcpy to access the memory, by that point the compiler is allowed to assume that the pointer is aligned, and so it can output code that does an unaligned memory access.
In practice, this resulted in a crash on some devices when this code is compiled with optimizations for 32-bit ARM with the Android NDK r14.
|
| |/
|/| |
|
| | |
|
|/ |
|
| |
|
|\ |
|
| |\
| | |
| | | |
Add Error String to Recv Status API
|
| | | |
|
| |\ \
| | | |
| | | | |
Avoid calling dtor on grpc_fd before putting it in freelist
|
| | | | |
|
| | | | |
|
| |\ \ \
| | | | |
| | | | | |
Add logic to work around buggy Android NDKs
|
| |\ \ \ \
| | | | | |
| | | | | | |
Eagerly Free Slices in Case of Partial Write
|
| | | | | |
| | | | | |
| | | | | | |
Don't delete the key in the pthread_key_create destructor. The key isn't specific to instances of values.
|
| | | | | | |
|
| | | | | | |
|
| |\ \ \ \ \
| | | |_|/ /
| | |/| | | |
|
| | |\ \ \ \
| | | | | | |
| | | | | | | |
Disable caching start-time for all clock-types except GPR_CLOCK_MONOTONIC
|
| | | | | | | |
|
| | |\ \ \ \ \
| | | | | | | |
| | | | | | | | |
Fix uses of resource quota in UV TCP code
|
| | | | | | | | |
|
| | | |/ / / /
| | |/| | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Caching the start-time for GPR_CLOCK_REALTIME has been causing errors in
cases where the system time is changed (after caching the time). In such
cases, the following functions produce incorrect results (and are off by
how much ever the system time was changed)
grpc_millis_to_timespec() and grpc_timespec_to_millis_round_down()
This can cause problems especially when using the above functions to
get timer deadlines or completion queue timeouts.
(In the worst case scenarios, the timeouts/deadlines will always occur (if the
timeout inverval / deadline was less than the system change delta)
Ideally we should be reverting https://github.com/grpc/grpc/pull/11866
but since that is a large change (which introduced new APIs in
exec_ctx.cc), I am doing this change to effectively revert to the old
behavior (while still keeping the new APIs introduced in exec_ctx)
|
| | | | | | | |
|
| | |\ \ \ \ \
| | | | | | | |
| | | | | | | | |
Use lock when popping requests on server shutdown
|
| | | | | | | | |
|
| | |\ \ \ \ \ \
| | | |_|/ / / /
| | |/| | | | | |
Inline Closure
|
| | | | | | | | |
|
| | | | | | | | |
|
| | | |/ / / /
| | |/| | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Doing this without a lock causes TSAN failures for quic.
There isn't much need to be clever here because this only impacts
shutdown performance, which doesn't really matter.
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | |\
| | | |_|_|_|/
| | |/| | | |
| | | | | | | |
surfacing-error-details
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| | | | |\ \ \
| | | |_|/ / /
| | |/| | | | |
|
|\ \ \ \ \ \ \
| | |/ / / / /
| |/| | | | | |
|
| | | | | | | |
|
| | | | | | | |
|
| |\ \ \ \ \ \
| | | | | | | |
| | | | | | | | |
reland: cpu_linux: Don't spam sched_getcpu failures on qemu
|
| | | | | | | | |
|