RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-25 06:47:35 +00:00
Code Issues Activity Actions

Kernel: Migrate TCP socket tables locking to ProtectedValue

Browse source 
Note: TCPSocket::create_client() has a dubious locking process where
the sockets by tuple table is first shared lock to check if the socket
exists and bail out if it does, then unlocks, then exclusively locks to
add the tuple. There could be a race condition where two client
creation requests for the same tuple happen at the same time and both
cleared the shared lock check. When in doubt, lock exclusively the
whole time.
This commit is contained in:
Jean-Baptiste Boric 2021-07-18 12:14:43 +02:00 • committed by Andreas Kling
parent 583abc27d8
commit 9216c72bfe
3 changed files with 90 additions and 85 deletions
Download patch file Download diff file Expand all files Collapse all files

9
Kernel/Net/NetworkTask.cpp
View file

@ -6,6 +6,7 @@
#include <Kernel/Debug.h>
#include <Kernel/Locking/Mutex.h>
#include <Kernel/Locking/ProtectedValue.h>
#include <Kernel/Net/ARP.h>
#include <Kernel/Net/EtherType.h>
#include <Kernel/Net/EthernetFrameHeader.h>
@ -639,11 +640,9 @@ void retransmit_tcp_packets()
// We must keep the sockets alive until after we've unlocked the hash table
// in case retransmit_packets() realizes that it wants to close the socket.
NonnullRefPtrVector<TCPSocket, 16> sockets;
{
MutexLocker locker(TCPSocket::sockets_for_retransmit().lock(), LockMode::Shared);
for (auto& socket : TCPSocket::sockets_for_retransmit().resource())
sockets.append(*socket);
}
TCPSocket::sockets_for_retransmit().for_each_shared([&](const auto& socket) {
sockets.append(*socket);
});
for (auto& socket : sockets) {
MutexLocker socket_locker(socket.lock());