Kernel: kernel doesn't allocate large blocks of data on stack

We used to allocate 1 KiB blocks on multiple places on stack. This
is a problem, since kernel stack shouldn't have to be too big

kernel/kernel/Shell.cpp

@@ -557,7 +557,10 @@ argument_done:
 			int fd = TRY(Process::current()->open(arguments[1], O_RDONLY));
 			BAN::ScopeGuard _([fd] { MUST(Process::current()->close(fd)); });
 
-			char buffer[1024] {};
+			char* buffer = new char[1024];
+			BAN::ScopeGuard buffer_guard([buffer] { delete[] buffer; });
+			ASSERT(buffer);
+
 			while (size_t n_read = TRY(Process::current()->read(fd, buffer, sizeof(buffer))))
 				TTY_PRINT("{}", BAN::StringView(buffer, n_read));
 			TTY_PRINTLN("");
@@ -621,7 +624,10 @@ argument_done:
 				return {};
 			}
 
-			uint8_t buffer[1024];
+			uint8_t* buffer = new uint8_t[1024];
+			BAN::ScopeGuard buffer_guard([buffer] { delete[] buffer; });
+			ASSERT(buffer);
+
 			for (size_t i = 1; i < arguments.size(); i++)
 			{
 				int fd = TRY(Process::current()->open(arguments[i], O_RDONLY));

kernel/kernel/Storage/ATABus.cpp

@@ -58,6 +58,10 @@ namespace Kernel
 	{
 		register_bus_irq_handler(this, irq);
 
+		uint16_t* identify_buffer = new uint16_t[256];
+		ASSERT(identify_buffer);
+		BAN::ScopeGuard _([identify_buffer] { delete[] identify_buffer; });
+
 		for (uint8_t i = 0; i < 2; i++)
 		{
 			m_devices[i] = new ATADevice(this);
@@ -66,7 +70,7 @@ namespace Kernel
 
 			BAN::ScopeGuard guard([this, i] { m_devices[i]->unref(); m_devices[i] = nullptr; });
 
-			uint16_t identify_buffer[256];
+			
 			auto type = identify(device, identify_buffer);
 			if (type == DeviceType::None)
 				continue;

kernel/kernel/Terminal/TTY.cpp

@@ -38,8 +38,9 @@ namespace Kernel
 	{
 		m_width = m_terminal_driver->width();
 		m_height = m_terminal_driver->height();
-		
+
 		m_buffer = new Cell[m_width * m_height];
+		ASSERT(m_buffer);
 
 		if (s_tty == nullptr)
 			s_tty = this;