[RFC UKL 10/10] Kconfig: Add config option for enabling and sample for testing UKL
From: Ali Raza
Date: Mon Oct 03 2022 - 18:24:07 EST
Add the KConfig file that will enable building UKL. Documentation
introduces the technical details for how UKL works and the motivations
behind why it is useful. Sample provides a simple program that still uses
the standard system call interface, but does not require a modified C
library.
Cc: Jonathan Corbet <corbet@xxxxxxx>
Cc: Masahiro Yamada <masahiroy@xxxxxxxxxx>
Cc: Michal Marek <michal.lkml@xxxxxxxxxxx>
Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Borislav Petkov <bp@xxxxxxxxx>
Cc: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx>
Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
Cc: Andy Lutomirski <luto@xxxxxxxxxx>
Cc: Eric Biederman <ebiederm@xxxxxxxxxxxx>
Cc: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx>
Cc: Arnd Bergmann <arnd@xxxxxxxx>
Cc: Juri Lelli <juri.lelli@xxxxxxxxxx>
Cc: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
Cc: Dietmar Eggemann <dietmar.eggemann@xxxxxxx>
Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
Cc: Ben Segall <bsegall@xxxxxxxxxx>
Cc: Mel Gorman <mgorman@xxxxxxx>
Cc: Daniel Bristot de Oliveira <bristot@xxxxxxxxxx>
Cc: Valentin Schneider <vschneid@xxxxxxxxxx>
Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Cc: Josh Poimboeuf <jpoimboe@xxxxxxxxxx>
Co-developed-by: Eric B Munson <munsoner@xxxxxx>
Signed-off-by: Eric B Munson <munsoner@xxxxxx>
Co-developed-by: Ali Raza <aliraza@xxxxxx>
Signed-off-by: Ali Raza <aliraza@xxxxxx>
---
Documentation/index.rst | 1 +
Documentation/ukl/ukl.rst | 104 ++++++++++++++++++++++++++++++++++++++
Kconfig | 2 +
kernel/Kconfig.ukl | 41 +++++++++++++++
samples/ukl/Makefile | 16 ++++++
samples/ukl/README | 17 +++++++
samples/ukl/syscall.S | 28 ++++++++++
samples/ukl/tcp_server.c | 99 ++++++++++++++++++++++++++++++++++++
8 files changed, 308 insertions(+)
create mode 100644 Documentation/ukl/ukl.rst
create mode 100644 kernel/Kconfig.ukl
create mode 100644 samples/ukl/Makefile
create mode 100644 samples/ukl/README
create mode 100644 samples/ukl/syscall.S
create mode 100644 samples/ukl/tcp_server.c
diff --git a/Documentation/index.rst b/Documentation/index.rst
index 4737c18c97ff..42f8cb7d4cae 100644
--- a/Documentation/index.rst
+++ b/Documentation/index.rst
@@ -167,6 +167,7 @@ to ReStructured Text format, or are simply too old.
tools/index
staging/index
+ ukl/ukl.rst
Translations
diff --git a/Documentation/ukl/ukl.rst b/Documentation/ukl/ukl.rst
new file mode 100644
index 000000000000..a07ebb51169e
--- /dev/null
+++ b/Documentation/ukl/ukl.rst
@@ -0,0 +1,104 @@
+SPDX-License-Identifier: GPL-2.0
+
+Unikernel Linux (UKL)
+=====================
+
+Unikernel Linux (UKL) is a research project aimed at integrating
+application specific optimizations to the Linux kernel. This RFC aims to
+introduce this research to the community. Any feedback regarding the idea,
+goals, implementation and research is highly appreciated.
+
+Unikernels are specialized operating systems where an application is linked
+directly with the kernel and runs in supervisor mode. This allows the
+developers to implement application specific optimizations to the kernel,
+which can be directly invoked by the application (without going through the
+syscall path). An application can control scheduling and resource
+management and directly access the hardware. Application and the kernel can
+be co-optimized, e.g., through LTO, PGO, etc. All of these optimizations,
+and others, provide applications with huge performance benefits over
+general purpose operating systems.
+
+Linux is the de-facto operating system of today. Applications depend on its
+battle tested code base, large developer community, support for legacy
+code, a huge ecosystem of tools and utilities, and a wide range of
+compatible hardware and device drivers. Linux also allows some degree of
+application specific optimizations through build time config options,
+runtime configuration, and recently through eBPF. But still, there is a
+need for even more fine-grained application specific optimizations, and
+some developers resort to kernel bypass techniques.
+
+Unikernel Linux (UKL) aims to get the best of both worlds by bringing
+application specific optimizations to the Linux ecosystem. This way,
+unmodified applications can keep getting the benefits of Linux while taking
+advantage of the unikernel-style optimizations. Optionally, applications
+can be modified to invoke deeper optimizations.
+
+There are two steps to unikernel-izing Linux, i.e., first, equip Linux with
+a unikernel model, and second, actually use that model to implement
+application specific optimizations. This patch focuses on the first part.
+Through this patch, unmodified applications can be built as Linux
+unikernels, albeit with only modest performance advantages. Like
+unikernels, UKL would allow an application to be statically linked into the
+kernel and executed in supervisor mode. However, UKL preserves most of the
+invariants and design of Linux, including a separate page-able application
+portion of the address space and a pinned kernel portion, the ability to
+run multiple processes, and distinct execution modes for application and
+kernel code. Kernel execution mode and application execution mode are
+different, e.g., the application execution mode allows application threads
+to be scheduled, handle signals, etc., which do not apply to kernel
+threads. Application built as a Linux unikernel will have its text and data
+loaded with the kernel at boot time, while the rest of the address space
+would remain unchanged. These applications invoke the system call
+functionality through a function call into the kernel system call entry
+point instead of through the syscall assembly instruction. UKL would
+support a normal userspace so the UKL application can be started, managed,
+profiled, etc., using normal command line utilities.
+
+Once Linux has a unikernel model, different application specific
+optimizations are possible. We have tried a few, e.g., fast system call
+transitions, shared stacks to allow LTO, invoking kernel functions
+directly, etc. We have seen huge performance benefits, details of which are
+not relevant to this patch and can be found in our paper.
+(https://arxiv.org/pdf/2206.00789.pdf)
+
+UKL differs significantly from previous projects, e.g., UML, KML and LKL.
+User Mode Linux (UML) is a virtual machine monitor implemented on syscall
+interface, a very different goal from UKL. Kernel Mode Linux (KML) allows
+applications to run in kernel mode and replaces syscalls with function
+calls. While KML stops there, UKL goes further. UKL links applications and
+kernel together which allows further optimizations e.g., fast system call
+transitions, shared stacks to allow LTO, invoking kernel functions directly
+etc. Details can be found in the paper linked above. Linux Kernel Library
+(LKL) harvests arch independent code from Linux, takes it to userspace as a
+library to be linked with applications. A host needs to provide arch
+dependent functionality. This model is very different from UKL. A detailed
+discussion of related work is present in the paper linked above.
+
+See samples/ukl for a simple TCP echo server example which can be built as
+a normal user space application and also as a UKL application. In the Linux
+config options, a path to the compiled and partially linked application
+binary can be specified. Kernel built with UKL enabled will search this
+location for the binary and link with the kernel. Applications and required
+libraries need to be compiled with -mno-red-zone -mcmodel=kernel flags
+because kernel mode execution can trample on application red zones and in
+order to link with the kernel and be loaded in the high end of the address
+space, application should have the correct memory model. Examples of other
+applications like Redis, Memcached etc along with glibc and libgcc etc.,
+can be found at https://github.com/unikernelLinux/ukl
+
+List of authors and contributors:
+=================================
+
+Ali Raza - aliraza@xxxxxx
+Thomas Unger - tommyu@xxxxxx
+Matthew Boyd - mboydmcse@xxxxxxxxx
+Eric Munson - munsoner@xxxxxx
+Parul Sohal - psohal@xxxxxx
+Ulrich Drepper - drepper@xxxxxxxxxx
+Richard Jones - rjones@xxxxxxxxxx
+Daniel Bristot de Oliveira - bristot@xxxxxxxxxx
+Larry Woodman - lwoodman@xxxxxxxxxx
+Renato Mancuso - rmancuso@xxxxxx
+Jonathan Appavoo - jappavoo@xxxxxx
+Orran Krieger - okrieg@xxxxxx
+
diff --git a/Kconfig b/Kconfig
index 745bc773f567..2a4594ae472c 100644
--- a/Kconfig
+++ b/Kconfig
@@ -29,4 +29,6 @@ source "lib/Kconfig"
source "lib/Kconfig.debug"
+source "kernel/Kconfig.ukl"
+
source "Documentation/Kconfig"
diff --git a/kernel/Kconfig.ukl b/kernel/Kconfig.ukl
new file mode 100644
index 000000000000..c2c5e1003605
--- /dev/null
+++ b/kernel/Kconfig.ukl
@@ -0,0 +1,41 @@
+menuconfig UNIKERNEL_LINUX
+ bool "Unikernel Linux"
+ depends on X86_64 && !RANDOMIZE_BASE && !PAGE_TABLE_ISOLATION
+ help
+ Unikernel Linux allows for a single, privileged process to be
+ linked with the kernel binary and be executed inplace of or
+ along side a more traditional user space.
+
+ If you don't know what this is, say N.
+
+config UKL_TLS
+ bool "Enable TLS for UKL application"
+ depends on UNIKERNEL_LINUX
+ default Y
+ help
+ Not all applications will make use of thread local storage,
+ but we need to account for it in the linker script if used.
+ For the application in samples/ this should be disabled, but
+ if you are working with glibc this should be 'Y'.
+
+ If unsure say 'Y' here
+
+config UKL_NAME
+ string "UKL Exec target"
+ depends on UNIKERNEL_LINUX
+ default "/UKL"
+ help
+ We need a way to trigger the start of the UKL application,
+ either by the kernel inplace of init or userspace when setup
+ is finished. The value given here is compared against the
+ filename passed to exec and if they match UKL is started.
+ For a more 'traditional' unikernel model, the value set here
+ should be given to the init= boot parameter.
+
+config UKL_ARCHIVE_PATH
+ string "Path static application archive"
+ depends on UNIKERNEL_LINUX
+ default "../UKL.a"
+ help
+ Where the linker should look for the statically linked application
+ and dependency archive.
diff --git a/samples/ukl/Makefile b/samples/ukl/Makefile
new file mode 100644
index 000000000000..93beb7750d4b
--- /dev/null
+++ b/samples/ukl/Makefile
@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: GPL-2.0
+
+CFLAGS += -I usr/include -fno-PIC -mno-red-zone -mcmodel=kernel
+
+UKL.a: tcp_server.o syscall.o userspace
+ $(AR) cr UKL.a tcp_server.o syscall.o
+ objcopy --prefix-symbols=ukl_ UKL.a
+
+tcp_server.o: tcp_server.c
+syscall.o: syscall.S
+
+userspace:
+ gcc -o tcp_server tcp_server.c
+
+clean:
+ rm -f UKL.a tcp_server.o syscall.o tcp_server
diff --git a/samples/ukl/README b/samples/ukl/README
new file mode 100644
index 000000000000..fbb771da033a
--- /dev/null
+++ b/samples/ukl/README
@@ -0,0 +1,17 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+UKL test program
+================
+
+tcp_server.c is a epoll based TCP echo server written in C which uses port
+no. 5555 by default. syscall.S translates syscall() function to a call
+instruction in assembly. Normally, C libraries provide syscall() function
+that translate into syscall assembly instruction. Run `make` and it will
+create a UKL.a and a tcp_server. UKL.a can then be copied to where UKL
+Linux build expects it to be present. This can be changed through the Linux
+config options (by running `make menuconfig` etc.) The resulting Linux
+kernel can be run, and once the userspace comes up, the echo server can be
+started by running the UKL exec command, again chosen through the Linux
+config options. tcp_server is a userspace binary of the same echo server
+which can be run normally. This is meant to show that UKL can run code
+which can also be run as a userspace binary without modification.
diff --git a/samples/ukl/syscall.S b/samples/ukl/syscall.S
new file mode 100644
index 000000000000..95d1c177fb05
--- /dev/null
+++ b/samples/ukl/syscall.S
@@ -0,0 +1,28 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+
+ .global _start
+_start:
+ jmp main
+
+ .global syscall
+
+/* Usage: long syscall (syscall_number, arg1, arg2, arg3, arg4, arg5, arg6)
+ We need to do some arg shifting, the syscall_number will be in
+ rax. */
+
+ .text
+syscall:
+ movq %rdi, %rax /* Syscall number -> rax. */
+ movq %rsi, %rdi /* shift arg1 - arg5. */
+ movq %rdx, %rsi
+ movq %rcx, %rdx
+ movq %r8, %r10
+ movq %r9, %r8
+ movq 8(%rsp),%r9 /* arg6 is on the stack. */
+ call entry_SYSCALL_64 /* Do the system call. */
+ cmpq $-4095, %rax /* Check %rax for error. */
+ jae loop /* Jump to error handler if error. */
+ ret /* Return to caller. */
+
+loop:
+ jmp loop
diff --git a/samples/ukl/tcp_server.c b/samples/ukl/tcp_server.c
new file mode 100644
index 000000000000..abf1a0e2bb79
--- /dev/null
+++ b/samples/ukl/tcp_server.c
@@ -0,0 +1,99 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <sys/epoll.h>
+#include <arpa/inet.h>
+#include <netinet/tcp.h>
+
+#define BACKLOG 512
+#define MAX_EVENTS 128
+#define MAX_MESSAGE_LEN 2048
+
+void error(char *msg);
+extern long syscall(long number, ...);
+
+int main(void)
+{
+ // some variables we need
+ struct sockaddr_in server_addr, client_addr;
+ socklen_t client_len = sizeof(client_addr);
+ int bytes_received;
+ char buffer[MAX_MESSAGE_LEN];
+ int on;
+ int result;
+ int sock_listen_fd, newsockfd;
+
+ // setup socket
+ sock_listen_fd = syscall(41, AF_INET, SOCK_STREAM, 0);
+ if (sock_listen_fd < 0)
+ error("Error creating socket..\n");
+
+ server_addr.sin_family = AF_INET;
+ server_addr.sin_port = 45845; //htons(portno);
+ server_addr.sin_addr.s_addr = INADDR_ANY;
+
+ // set TCP NODELAY
+ on = 1;
+ result = syscall(54, sock_listen_fd, IPPROTO_TCP, TCP_NODELAY, &on, sizeof(on));
+ if (result < 0)
+ error("Can't set TCP_NODELAY to on");
+
+ // bind socket and listen for connections
+ if (syscall(49, sock_listen_fd, (struct sockaddr *)&server_addr, sizeof(server_addr)) < 0)
+ error("Error binding socket..\n");
+
+ if (syscall(50, sock_listen_fd, BACKLOG) < 0)
+ error("Error listening..\n");
+
+ struct epoll_event ev, events[MAX_EVENTS];
+ int new_events, sock_conn_fd, epollfd;
+
+ epollfd = syscall(213, MAX_EVENTS);
+ if (epollfd < 0)
+ error("Error creating epoll..\n");
+
+ ev.events = EPOLLIN;
+ ev.data.fd = sock_listen_fd;
+
+ if (syscall(233, epollfd, EPOLL_CTL_ADD, sock_listen_fd, &ev) == -1)
+ error("Error adding new listeding socket to epoll..\n");
+
+ while (1) {
+ new_events = syscall(232, epollfd, events, MAX_EVENTS, -1);
+
+ if (new_events == -1)
+ error("Error in epoll_wait..\n");
+
+ for (int i = 0; i < new_events; ++i) {
+ if (events[i].data.fd == sock_listen_fd) {
+ sock_conn_fd = syscall(288, sock_listen_fd,
+ (struct sockaddr *)&client_addr,
+ &client_len, SOCK_NONBLOCK);
+ if (sock_conn_fd == -1)
+ error("Error accepting new connection..\n");
+
+ ev.events = EPOLLIN | EPOLLET;
+ ev.data.fd = sock_conn_fd;
+ if (syscall(233, epollfd, EPOLL_CTL_ADD, sock_conn_fd, &ev) == -1)
+ error("Error adding new event to epoll..\n");
+ } else {
+ newsockfd = events[i].data.fd;
+ bytes_received = syscall(45, newsockfd, buffer, MAX_MESSAGE_LEN,
+ 0, NULL, NULL);
+ if (bytes_received <= 0) {
+ syscall(233, epollfd, EPOLL_CTL_DEL, newsockfd, NULL);
+ syscall(48, newsockfd, SHUT_RDWR);
+ } else {
+ syscall(44, newsockfd, buffer, bytes_received, 0, NULL, 0);
+ }
+ }
+ }
+ }
+}
+
+void error(char *msg)
+{
+ syscall(1, 1, msg, 15);
+ syscall(60, 1);
+}
--
2.21.3