libbinder: RPC handle builtup refcounts
Generally, in the binder RPC wire protocol, we don't have both the
clients and the servers writing data into sockets. However, in the case
of async transactions, this happens in an unbounded way because a client
may send many oneway transactions, and the server will be sending back
refcounting information related to these transactions (which we process
lazily).
In order to prevent this from building up, when sending a transaction,
if we're unable to write it, instead of waiting, drain that reference
counting information.
Bug: 182940634
Test: binderRpcTest (no longer deadlocks in OnewayStressTest)
Test: manually check 'drainCommands' happens in both raw and tls cases
during this test (checking we are actually getting coverage)
Change-Id: I82039d6188196261b22316e95d8e180c4c33ae73
diff --git a/libs/binder/RpcServer.cpp b/libs/binder/RpcServer.cpp
index 5733993..cf95fda 100644
--- a/libs/binder/RpcServer.cpp
+++ b/libs/binder/RpcServer.cpp
@@ -275,7 +275,7 @@
RpcConnectionHeader header;
if (status == OK) {
status = client->interruptableReadFully(server->mShutdownTrigger.get(), &header,
- sizeof(header));
+ sizeof(header), {});
if (status != OK) {
ALOGE("Failed to read ID for client connecting to RPC server: %s",
statusToString(status).c_str());
@@ -288,7 +288,7 @@
if (header.sessionIdSize > 0) {
sessionId.resize(header.sessionIdSize);
status = client->interruptableReadFully(server->mShutdownTrigger.get(),
- sessionId.data(), sessionId.size());
+ sessionId.data(), sessionId.size(), {});
if (status != OK) {
ALOGE("Failed to read session ID for client connecting to RPC server: %s",
statusToString(status).c_str());
@@ -313,7 +313,7 @@
};
status = client->interruptableWriteFully(server->mShutdownTrigger.get(), &response,
- sizeof(response));
+ sizeof(response), {});
if (status != OK) {
ALOGE("Failed to send new session response: %s", statusToString(status).c_str());
// still need to cleanup before we can return