Sinipelto/banan-os
1
0
Fork 0
forked from Bananymous/banan-os
Code Pull Requests Activity

Kernel: Fix most of mutex + block race conditions

Browse Source 
All block functions now take an optional mutex parameter that is
atomically unlocked instead of having the user unlock it before hand.
This prevents a ton of race conditions everywhere in the code!
This commit is contained in:
Bananymous
2025-06-06 03:59:22 +03:00
parent 96d5ed9cc7
commit eecdad50a6
36 changed files with 374 additions and 322 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
kernel/kernel/Storage/DiskCache.cpp
View File

@@ -1,5 +1,4 @@
#include <kernel/BootInfo.h>
#include <kernel/Lock/LockGuard.h>
#include <kernel/Memory/Heap.h>
#include <kernel/Memory/PageTable.h>
#include <kernel/Storage/DiskCache.h>

13
kernel/kernel/Storage/NVMe/Queue.cpp
View File

@@ -1,4 +1,4 @@
#include <kernel/Lock/LockGuard.h>
#include <kernel/Lock/SpinLockAsMutex.h>
#include <kernel/Storage/NVMe/Queue.h>
#include <kernel/Thread.h>
#include <kernel/Timer/Timer.h>
@@ -72,7 +72,7 @@ namespace Kernel
// scheduler has put the current thread blocking.
// EINTR should also be handled here.
while (!(m_done_mask & cid_mask) && SystemTimer::get().ms_since_boot() < start_time_ms + s_nvme_command_timeout_ms)
m_thread_blocker.block_with_wake_time_ms(start_time_ms + s_nvme_command_timeout_ms);
m_thread_blocker.block_with_wake_time_ms(start_time_ms + s_nvme_command_timeout_ms, nullptr);
if (m_done_mask & cid_mask)
{
@@ -87,12 +87,12 @@ namespace Kernel
uint16_t NVMeQueue::reserve_cid()
{
auto state = m_lock.lock();
SpinLockGuard guard(m_lock);
while (~m_used_mask == 0)
{
m_lock.unlock(state);
m_thread_blocker.block_with_timeout_ms(s_nvme_command_timeout_ms);
state = m_lock.lock();
SpinLockGuardAsMutex smutex(guard);
m_thread_blocker.block_with_timeout_ms(s_nvme_command_timeout_ms, &smutex);
}
uint16_t cid = 0;
@@ -104,7 +104,6 @@ namespace Kernel
m_used_mask |= (size_t)1 << cid;
m_lock.unlock(state);
return cid;
}