From: Jack Lange Date: Tue, 12 Jan 2010 20:39:21 +0000 (-0600) Subject: Merge branch 'devel' of ssh://palacios@newskysaw.cs.northwestern.edu/home/palacios... X-Git-Url: http://v3vee.org/palacios/gitweb/gitweb.cgi?p=palacios.git;a=commitdiff_plain;h=2b9ea885a5ccfc672b0fd05302a271bbcd61d17f;hp=7021908ce7f63df7c8a9e8e007471d8df96aa1b6 Merge branch 'devel' of ssh://palacios@newskysaw.cs.northwestern.edu/home/palacios/palacios into devel --- diff --git a/manual/guest_build/Palacios_Guest_Build.tex b/manual/guest_build/Palacios_Guest_Build.tex index 8f3e319..d772496 100644 --- a/manual/guest_build/Palacios_Guest_Build.tex +++ b/manual/guest_build/Palacios_Guest_Build.tex @@ -25,44 +25,80 @@ \section{Getting the guest image build tools} -In order to build the bootable guest iso image, we need to build a Linux kernel -image from source and an initial file system containing a set of useful binary -files which will be described in the following text. We will use a new directory -for demonstration; the root directory for the following examples is -\verb+test/+: +In order to build the bootable guest ISO image, we need to build a Linux kernel + from source and an initial ramdisk file system containing a set of useful +tools. We will use a new directory for demonstration; the root directory for the +following examples is ``\verb+test/+": \begin{verbatim} -mkdir test/ -cd test/ +[jdoe@newskysaw ~]$ mkdir test/ \end{verbatim} -\vspace{10pt} \noindent There are a set of tools and sources that are useful for the guest image -building procedure. You can checkout these resources from our git repositories; -to check them out to your local \verb+test/+ directory use the following -commands: +building procedure. You can obtain these resources from our git repositories. +Change to the ``\verb+test/+" directory and clone the resources: \begin{verbatim} -git clone http://hornet.cs.northwestern.edu:9005/busybox -git clone http://hornet.cs.northwestern.edu:9005/initrd -git clone http://hornet.cs.northwestern.edu:9005/linux-2.6.30.y +[jdoe@newskysaw test]$ git clone http://hornet.cs.northwestern.edu:9005/busybox +[jdoe@newskysaw test]$ git clone http://hornet.cs.northwestern.edu:9005/initrd +[jdoe@newskysaw test]$ git clone http://hornet.cs.northwestern.edu:9005/linux-2.6.30.y \end{verbatim} \section{Building the ramdisk filesystem} +The guest requires an initial ramdisk filesystem. Jack has made one that you can +leverage; it is temporarily located in his home directory. You will need sudo +or root access to create the device files when you unpack the archive: -% Introductory text explaining why a ramdisk filesystem is necessary, and a -% small blurb about what it is. Mostly this is necessary because the -% introduction said that the "useful binary files" would be described. +\begin{verbatim} +[jdoe@newskysaw test]$ cp /home/jarusl/initrd/disks/v3vee_initramfs.tar.gz . +[jdoe@newskysaw test]$ sudo tar -C initrd -xzf v3vee_initramfs.tar.gz +\end{verbatim} + +\noindent +If you require a custom initial ramdisk filesystem, change to the +``\verb|initrd/initramfs/|" directory and perform the following steps: -Jack has made an initial ramdisk system that you can leverage. The file is -temporarily in the directory -\verb|/home/jarusl/initrd/disks/v3vee_initramfs.tar.gz| on the -newskysaw machine. If you require a custom initial ramdisk, copy the directories -and files that you require into the \verb+initramfs+/ directory. For minimal -device support, copy theses devices into the \verb+initramfs/dev/+ directory: -console, ram, null, tty (you probably need root privilege to copy and make the -device files). +\begin{verbatim} +[jdoe@newskysaw initramfs]$ mkdir -p proc sys var/log +\end{verbatim} + +\noindent +Edit the ``\verb|init_task|" script and uncomment these lines: + +\begin{verbatim} +#mknod /dev/tty0 c 4 0 +#mknod /dev/tty1 c 4 1 +#mknod /dev/tty2 c 4 2 +\end{verbatim} + +\pagebreak + +\noindent +Create the ``\verb|console|" device. If you have sudo or root access it is +possible to create this device manually: + +\begin{verbatim} +[jdoe@newskysaw initramfs]$ sudo mknod dev/console c 5 1 +[jdoe@newskysaw initramfs]$ sudo chmod 0600 dev/console +\end{verbatim} + +\noindent +If you do not have sudo or root access it is still possible to create the +``\verb|console|" device indirectly through the kernel build. Change to the +``\verb|initrd/|" directory and create a file called ``\verb|root_files|". Add +the following line: + +\begin{verbatim} +nod /dev/console 0600 0 0 c 5 1 +\end{verbatim} + +\noindent +The ``\verb|root_files|" file is used when building the Linux kernel in the +section Configuring and building the Linux kernel. Finally, create any +additional directories and copy any additional files that you need. Your initial +ramdisk filesystem is prepped and ready for installation of the BusyBox tools as +described in the section Configuring and installing BusyBox tools. @@ -75,7 +111,7 @@ device files). \end{center} \end{figure} -\begin{figure}[h] +\begin{figure}[ht] \begin{center} \colfigsize\epsffile{busyboxConf2.eps} \end{center} @@ -87,37 +123,40 @@ device files). BusyBox is a software application released as Free software under the GNU GPL that provides many standard Unix tools. BusyBox combines tiny versions of many -common UNIX utilities into a single small executable. For more details on -BusyBox, visiting \url{http://busybox.net}. To configure BusyBox, in the -\verb+busybox/+ directory, type the following: +common UNIX utilities into a single, small executable. For more details on +BusyBox visit \url{http://busybox.net}. To configure BusyBox, in the +``\verb+busybox/+" directory, type the following: \begin{verbatim} -make menuconfig +[jdoe@newskysaw busybox]$ make menuconfig \end{verbatim} +\noindent or \begin{verbatim} -make xconfig (X version) +[jdoe@newskysaw busybox]$ make xconfig \end{verbatim} -\vspace{10pt} \noindent -You can add the tools you need into the guest image. There are two required -configuration options: in ``\verb|BusyBox settings->Build Options|", check the -``\verb|Build BusyBox as a static binary|" option, and in -``\verb|BusyBox settings->Installation Options|", set the -``\verb|Busybox installation prefix|" to the path of your \verb|initramfs| -directory, as shown in figure \ref{fig:busyboxcf2}. After you finish configuring -BusyBox, save your configuration and quit the window. Then, to make the BusyBox -tools, type the following: +The BusyBox tools will be installed in the guest's initial ramdisk filesystem; +you can add any tools that you need. There are two required configuration +options. In the +``\verb|BusyBox settings->Build Options|" menu check the +``\verb|Build BusyBox as a static binary (no shared libs)|" option, as shown in +figure \ref{fig:busyboxcf1}, and in the +``\verb|BusyBox settings->Installation Options|" menu set the +``\verb|Busybox installation prefix|" to the path of the +``\verb|initrd/initramfs|" directory, as shown in figure \ref{fig:busyboxcf2}. +After you finish configuring BusyBox, save your configuration and quit the +window. Then, to make the BusyBox tools, type the following: \begin{verbatim} -make +[jdoe@newskysaw busybox]$ make \end{verbatim} -Install the tools to your initial ramdisk directory: +Install the tools to the guest's initial ramdisk filesystem directory: \begin{verbatim} -make install +[jdoe@newskysaw busybox]$ make install \end{verbatim} \begin{figure}[ht] @@ -129,74 +168,159 @@ make install \end{figure} -\section{Configuring and compiling the Linux kernel} +\section{Configuring and building the Linux kernel} -Change to the \verb|linux-2.6.30.y/| directory (or whatever your Linux kernel -source directory is named) and type the following: +The following procedure demonstrates how to configure and build a 32-bit Linux +kernel. Change to the ``\verb|linux-2.6.30.y/|" directory. There is a custom +configuration file ``\verb|jrl-default-config|" which is configured with minimal +kernel options (all unnecessary options are removed to keep the guest booting +process fast). If you are using the custom configuration file type the +following: \begin{verbatim} -make menuconfig +[jdoe@newskysaw linux-2.6.30.y]$ cp jrl-default-config .config \end{verbatim} -or + +\noindent +Configure the kernel to meet your requirements. For more on configuring and +building Linux kernels, check online. Type the following: + \begin{verbatim} -make xconfig (X version) +[jdoe@newskysaw linux-2.6.30.y]$ make ARCH=i386 menuconfig \end{verbatim} -\vspace{10pt} \noindent -Configure the kernel to meet your requirements. There is a custom configuration -file \verb|jrl-default-config| which is configured with minimal kernel options -(all unnecessary options are removed to keep the guest booting process fast). -For more on configuring and compiling Linux kernel images, check online. +or + +\begin{verbatim} +[jdoe@newskysaw linux-2.6.30.y]$ make ARCH=i386 xconfig +\end{verbatim} -\vspace{5pt} \noindent The kernel must be configured with the initial ramdisk file system directory -(e.g. \verb|initrd/initramfs|): in the ``\verb|General setup|" menu under +(e.g. ``\verb|initrd/initramfs/|"): in the ``\verb|General setup|" menu under option ``\verb|Initial RAM filesystem and RAM disk support|" set the -``\verb|Initramfs source file(s)|" option to the path of your \verb|initramfs| -directory as shown in figure \ref{fig:linuxcf}. When you are finished -configuring the kernel, save your configuration, and type the following: +``\verb|Initramfs source file(s)|" option to the path of the +``\verb|initrd/initramfs/|" directory, as shown in figure \ref{fig:linuxcf}. +Additionally, if you are using the ``\verb|root_files|" file to create devices +files, add the ``\verb|root_files|" file path, separated by a space, after the +initial ramdisk filesystem directory. When you are finished configuring the +kernel, save your configuration, and type the following: + \begin{verbatim} -make +[jdoe@newskysaw linux-2.6.30.y]$ make ARCH=i386 \end{verbatim} -Some blurb about where the kernel image is... +\noindent +The Linux kernel can be found here: ``\verb|arch/x86/boot/bzImage|". The initial +ramdisk filesystem can be found here: ``\verb|usr/initramfs_data.cpio|". The +Linux kernel and initial ramdisk filesystem are used to build the Linux ISO +image in the section Building the Linux ISO image. -\section{Configuring guest devices} -Checkout the updated Palacios repository to \verb|palacios/|. (You can find -instructions for checking out the Palacios and Kitten repositories at -\url{http://www.v3vee.org/palacios/}). To build the guest VM creator tool, -change to the \verb|palacios/utils/guest_creator| directory and type the -following: +\section{Building the Linux ISO image} + +The Linux ISO image is a bootable image containing the Linux kernel, initial +ramdisk filesystem, a boot loader, and a boot loader configuration file. For +this procedure, we'll use the ``\verb|test/iso/|" directory as the Linux ISO +build directory: + +\begin{verbatim} +[jdoe@newskysaw test]$ mkdir iso +\end{verbatim} + +\noindent +Change to the ``\verb|iso/|" directory and copy the required files: \begin{verbatim} -make +[jdoe@newskysaw iso]$ cp ../linux-2.6.30.y/arch/x86/boot/bzImage vmlinuz +[jdoe@newskysaw iso]$ cp ../linux-2.6.30.y/usr/initramfs_data.cpio initramfs +[jdoe@newskysaw iso]$ cp /usr/lib/syslinux/isolinux.bin . \end{verbatim} -\vspace{10pt} \noindent -You will get the \verb|build_vm| executable. The guest configuration file uses -XML. A sample configuration file is provided: \verb|default.xml|. +Create a file called ``\verb|isolinux.cfg|" and add the following lines: + +\begin{verbatim} +default linux +prompt 0 + +label linux + kernel vmlinuz + append initrd=initrd +\end{verbatim} -\vspace{5pt} \noindent -***Various information about how to configure the VM.*** +Change back to the ``\verb|test/|" directory and build the Linux ISO image: + +\begin{verbatim} +[jdoe@newskysaw test]$ mkisofs -o linux.iso -b isolinux.bin -no-emul-boot \ +-boot-load-size 4 -boot-info-table -iso-level 2 -input-charset UTF-8 iso/ +\end{verbatim} -\vspace{5pt} \noindent -Once you have configured the VM, type the following to build the full guest VM -image: +The ``\verb|linux.iso|'' file is the Linux ISO image and is used to build the +guest image in the section Configuring and building the guest image: + \begin{verbatim} -./build_vm myconfig.xml -o guest.iso +[jdoe@newskysaw test]$ file linux.iso +linux.iso: ISO 9660 CD-ROM filesystem data 'CDROM ' (bootable) \end{verbatim} -where \verb+myconfig.xml+ is your guest configuration file, and \verb+guest.iso+ -is the output image file that will be used to configure kitten in the next -section. +\section{Configuring and building the guest image} + +Checkout the updated Palacios repository to the ``\verb|palacios/|" directory. +(You can find instructions for checking out the Palacios repository at +\url{http://www.v3vee.org/palacios/}). The guest creator utility is required for +building the guest image. Change to the ``\verb|palacios/utils/guest_creator|" +directory and build the guest creator utility: + +\begin{verbatim} +[jdoe@newskysaw guest_creator]$ make +\end{verbatim} + +\noindent +You will get the ``\verb|build_vm|" utility: +\begin{verbatim} +[jdoe@newskysaw guest_creator]$ file build_vm +build_vm: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked +(uses shared libs), for GNU/Linux 2.6.9, not stripped +\end{verbatim} + +\noindent +The guest configuration file is written in XML. A sample configuration file is +provided: ``\verb|default.xml|". Make a copy of the default configuration file +named ``\verb|myconfig.xml|" and edit the configuration elements that you are +interested in (if a device is included in the guest configuration file, it +must be configured in the section Configuring and building Palacios or the guest +will not boot). Of particular importance is the ``\verb|files|" element. Comment +out this attribute: + +\begin{verbatim} + +\end{verbatim} + +\noindent +Add an attribute that specifies the location of the Linux ISO image: + +\begin{verbatim} + +\end{verbatim} + +\noindent +When you are finished editing the guest configuration save the configuration +file. The guest image consists of the guest configuration file and the Linux +ISO image. Build the guest image with the guest creator utility: + +\begin{verbatim} +[jdoe@newskysaw guest_creator]$ ./build_vm myconfig.xml -o guest.iso +\end{verbatim} + +\noindent +The guest image, ``\verb+guest.iso+", is embedded in Kitten's +``\verb|init_task|" in the section Configuring and building Kitten. \pagebreak @@ -208,7 +332,7 @@ section. \label{fig:kittencf} \end{figure} -\begin{figure}[h] +\begin{figure}[ht] \begin{center} \colfigsize\epsffile{kittenConf2.eps} \end{center} @@ -222,78 +346,79 @@ section. You can find the detailed manual of getting and building Palacios and Kitten from scratch in the Palacios website (\url{http://www.v3vee.org/palacios}). Here -we only give the specific requirements related to the booting guest procedure. -To configure Palacios, in the Palacios root directory (i.e. \verb+palacios/+) -type the following: +we only give the specific requirements related to the procedure of booting the +guest. To configure Palacios, change to the ``\verb|test/palacios/|" directory +and type the following: \begin{verbatim} -make config +[jdoe@newskysaw palacios]$ make menuconfig \end{verbatim} + +\noindent or + \begin{verbatim} -make xconfig +[jdoe@newskysaw palacios]$ make xconfig \end{verbatim} -\vspace{10pt} \noindent -When you have configured the components you want to build into Palacios type -the following: +Don't forget to include the devices that your guest image requires. When you +have configured the components you want to build into Palacios, save the +configuration and close the window. To build Palacios type the following: \begin{verbatim} -make +[jdoe@newskysaw palacios]$ make \end{verbatim} or \begin{verbatim} -make all +[jdoe@newskysaw palacios]$ make all \end{verbatim} -\vspace{10pt} \noindent -Once the Palacios static library has been built you can find the library file -\verb+libv3vee.a+ in the Palacios root directory. +Once the Palacios static library has been built you can find the library file, +``\verb+libv3vee.a+", in the Palacios root directory. \subsection*{Configuring and building Kitten} -To build Kitten, first configure it in as you did Palacios. Change to the -\verb+kitten/+ directory and type the following: +Configure Kitten. Change to the ``\verb+test/kitten/+" directory and type the +following: \begin{verbatim} -make config +[jdoe@newskysaw kitten]$ make menuconfig \end{verbatim} + +\noindent or + \begin{verbatim} -make xconfig +[jdoe@newskysaw kitten]$ make xconfig \end{verbatim} -\vspace{10pt} \noindent Under the ``\verb|Virtualization|" menu select the ``\verb|Include Palacios virtual machine monitor|" option. Set the -``\verb|Path to pre-built Palacios tree|" option to the path of your Palacios -build, and set the ``\verb|Path to guest ISO image|" option to the path -containing the guest image that was built in the Configuring guest devices -section of this manual. - -\vspace{10pt} -\noindent -After configuring Kitten, to build Kitten, type the following: +``\verb|Path to pre-built Palacios tree|" option to the Palacios build tree +path, ``\verb|..\palacios|", as shown in figure \ref{fig:kittencf}. Set the +``\verb|Path to guest OS ISO image|" option to the guest image path,\\ +''\verb|../palacios/utils/guest_creator/guest.iso|'', as shown in figure +\ref{fig:kittencf2}. When you have finished configuring Kitten, save the +configuration and close the window. To build Kitten type the following: \begin{verbatim} -make isoimage +[jdoe@newskysaw kitten]$ make isoimage \end{verbatim} -\vspace{10pt} \noindent -This builds the bootable ISO image file with guest OS, Palacios and Kitten. -The ISO file is located in \verb+kitten/arch/x86_64/boot/image.iso+. +This builds the bootable ISO image file with guest image, Palacios, and Kitten. +The ISO file is located in ``\verb+kitten/arch/x86_64/boot/image.iso+". \pagebreak \noindent -You have successfully created a guest CD image file that can be booted on a -machine. You can boot the file on Qemu using the following sample command: +You have successfully created an ISO image file that can be booted on a machine. +You can boot the file on Qemu using the following sample command: \begin{verbatim} -/opt/vmm-tools/qemu/bin/qemu-system-x86_64 \ +[jdoe@newskysaw test]$ /opt/vmm-tools/qemu/bin/qemu-system-x86_64 \ -smp 1 \ -m 2047 \ -serial file:./serial.out \ @@ -301,7 +426,6 @@ machine. You can boot the file on Qemu using the following sample command: < /dev/null \end{verbatim} -\vspace{10pt} \noindent We have finished the entire procedure for building a guest image and booting it on the Palacios VMM. For more updated details, check the Palacios website diff --git a/palacios/include/palacios/vmm_vnet.h b/palacios/include/palacios/vmm_vnet.h index 51d99c4..c94c6c6 100644 --- a/palacios/include/palacios/vmm_vnet.h +++ b/palacios/include/palacios/vmm_vnet.h @@ -56,8 +56,43 @@ struct routing_entry{ }__attribute__((packed)); +struct eth_header { + uchar_t dest[6]; + uchar_t src[6]; + uint16_t type; // indicates layer 3 protocol type +}__attribute__((packed)); + +struct ip_header { + uint8_t version: 4; + uint8_t hdr_len: 4; + uchar_t tos; + uint16_t total_len; + uint16_t id; + uint8_t flags: 3; + uint16_t offset: 13; + uchar_t ttl; + uchar_t proto; + uint16_t cksum; + uint32_t src_addr; + uint32_t dst_addr; +}__attribute__((packed)); + +struct udp_header { + uint16_t src_port; + uint16_t dst_port; + uint16_t len; + uint16_t csum;//set to zero, disable the xsum +}__attribute__((packed)); + +struct udp_link_header { + struct eth_header eth_hdr; + struct ip_header ip_hdr; + struct udp_header udp_hdr; +}__attribute__((packed)); + +#define DEVICE_NAME_LEN 20 struct vnet_if_device { - char name[50]; + char name[DEVICE_NAME_LEN]; uchar_t mac_addr[6]; struct vm_device *dev; @@ -67,14 +102,12 @@ struct vnet_if_device { }__attribute__((packed)); -#define VNET_HEADER_LEN 64 struct vnet_if_link { prot_type_t pro_type; //protocal type of this link unsigned long dest_ip; uint16_t dest_port; - uchar_t vnet_header[VNET_HEADER_LEN]; //header applied to the packet in/out from this link - uint16_t hdr_len; + struct udp_link_header vnet_header; //header applied to the packet in/out from this link int (*input)(uchar_t *data, uint32_t len, void *private_data); @@ -95,7 +128,8 @@ struct link_entry { }__attribute__((packed)); -int v3_vnet_send_pkt(uchar_t *buf, int length); +int v3_vnet_send_rawpkt(uchar_t *buf, int len, void *private_data); +int v3_vnet_send_udppkt(uchar_t *buf, int len, void *private_data); //int vnet_register_device(struct vm_device *vdev, // char *dev_name, // uchar_t mac[6], diff --git a/palacios/src/devices/netdisk.c b/palacios/src/devices/netdisk.c index 552de56..3052a9b 100644 --- a/palacios/src/devices/netdisk.c +++ b/palacios/src/devices/netdisk.c @@ -132,7 +132,7 @@ static int read(uint8_t * buf, uint64_t lba, uint64_t num_bytes, void * private_ } if (ret_len != length) { - PrintError("Read length mismatch (req=%d) (result=%d)\n", length, ret_len); + PrintError("Read length mismatch (req=%llu) (result=%u)\n", length, ret_len); return -1; } diff --git a/palacios/src/devices/pci_passthrough.c b/palacios/src/devices/pci_passthrough.c index 18e4ef4..a799e1f 100644 --- a/palacios/src/devices/pci_passthrough.c +++ b/palacios/src/devices/pci_passthrough.c @@ -331,7 +331,11 @@ static int pci_bar_write(int bar_num, uint32_t * src, void * private_data) { struct pt_bar * pbar = &(state->phys_bars[bar_num]); struct pt_bar * vbar = &(state->virt_bars[bar_num]); - PrintDebug("Bar update src=0x%x\n", *src); + PrintDebug("Bar update: bar_num=%d, src=0x%x\n", bar_num,*src); + PrintDebug("vbar is size=%u, type=%d, addr=0x%x, val=0x%x\n",vbar->size, vbar->type, vbar->addr, vbar->val); + PrintDebug("pbar is size=%u, type=%d, addr=0x%x, val=0x%x\n",pbar->size, pbar->type, pbar->addr, pbar->val); + + if (vbar->type == PT_BAR_NONE) { return 0; diff --git a/palacios/src/palacios/vmm_vnet.c b/palacios/src/palacios/vmm_vnet.c index 72d4049..bf3b2c6 100644 --- a/palacios/src/palacios/vmm_vnet.c +++ b/palacios/src/palacios/vmm_vnet.c @@ -34,8 +34,9 @@ struct ethernet_pkt { - uint32_t size; + uint32_t size; //size of data uint16_t type; + struct udp_link_header ext_hdr; //possible externel header to applied to data before sent char data[ETHERNET_PACKET_LEN]; }; @@ -44,7 +45,7 @@ struct ethernet_pkt { struct in_pkt_header { char ethernetdest[6]; char ethernetsrc[6]; - unsigned char ethernettype[2]; // indicates layer 3 protocol type + unsigned char ethernettype[2]; //layer 3 protocol type char ip[20]; }; @@ -88,7 +89,7 @@ struct route_cache_entry { static struct hashtable * g_route_cache; -static void print_packet(char * pkt, int size) { +static void print_packet(uchar_t *pkt, int size) { PrintDebug("Vnet: print_data_packet: size: %d\n", size); v3_hexdump(pkt, size, NULL, 0); } @@ -108,9 +109,27 @@ static void print_device_addr(char * ethaddr) { } #endif +static uint16_t ip_xsum(struct ip_header *ip_hdr, int hdr_len){ + long sum = 0; -// network connection functions -static inline void ethernet_packet_init(struct ethernet_pkt * pt, const char * data, const size_t size) { + uint16_t *p = (uint16_t*) ip_hdr; + while(hdr_len > 1){ + sum += *(p++); + if(sum & 0x80000000) + sum = (sum & 0xFFFF) + (sum >> 16); + hdr_len -= 2; + } + + if(hdr_len) + sum += (uint16_t) *(uchar_t *)p; + + while(sum>>16) + sum = (sum & 0xFFFF) + (sum >> 16); + + return (uint16_t)~sum; +} + +static inline void ethernet_packet_init(struct ethernet_pkt *pt, uchar_t *data, const size_t size) { pt->size = size; memcpy(pt->data, data, size); } @@ -218,6 +237,7 @@ static int look_into_cache(route_hashkey_t hashkey, int * matches) { return n_matches; } +#ifdef CONFIG_DEBUG_VNET static inline uint8_t hex_nybble_to_nybble(const uint8_t hexnybble) { uint8_t x = toupper(hexnybble); @@ -246,8 +266,9 @@ static inline void mac_to_string(char mac[6], char * buf) { mac[0], mac[1], mac[2], mac[3], mac[4], mac[5]); } +#endif -static int add_link_entry(struct link_entry * link) { +static int __add_link_entry(struct link_entry * link) { int idx; for (idx = 0; idx < MAX_LINKS; idx++) { @@ -263,7 +284,7 @@ static int add_link_entry(struct link_entry * link) { return -1; } -static int add_route_entry(struct routing_entry * route) { +static int __add_route_entry(struct routing_entry * route) { int idx; for (idx = 0; idx < MAX_ROUTES; idx++) { @@ -308,7 +329,7 @@ static int vnet_add_route_entry(char src_mac[6], new_route->src_link_idx = src; new_route->src_type = src_type; - if ((idx = add_route_entry(new_route)) == -1) { + if ((idx = __add_route_entry(new_route)) == -1) { PrintError("Could not add route entry\n"); return -1; } @@ -318,7 +339,7 @@ static int vnet_add_route_entry(char src_mac[6], return idx; } -static void * delete_link_entry(int index) { +static void * __delete_link_entry(int index) { struct link_entry * link = NULL; void * ret = NULL; link_type_t type; @@ -386,11 +407,11 @@ static int find_route_entry(char src_mac[6], return -1; } -static int delete_route_entry(int index) { +static int __delete_route_entry(int index) { struct routing_entry * route; if ((index >= MAX_ROUTES) || (g_routes.routes[index] == NULL)) { - PrintError("Could not find route entry %d\n", index); + PrintDebug("VNET: wrong index in delete route entry %d\n", index); return -1; } @@ -417,11 +438,11 @@ static int vnet_delete_route_entry_by_addr(char src_mac[6], dest_mac_qual, link_idx, type, src, src_type); if (index == -1) { - PrintError("Error deleting route entry by addr\n"); + PrintDebug("VNET: wrong in delete route entry %d\n", index); return -1; } - return delete_route_entry(index); + return __delete_route_entry(index); } static int match_route(uint8_t * src_mac, @@ -559,102 +580,7 @@ static int match_route(uint8_t * src_mac, return no; } -#if 0 -// TODO: To be removed -static int process_udpdata() -{ - struct ethernet_pkt * pt; - - uint32_t dest = 0; - uint16_t remote_port = 0; - SOCK link_sock = g_udp_sockfd; - int length = sizeof(struct ethernet_pkt) - (2 * sizeof(int)); //minus the "size" and "type" - - //run in a loop to get packets from outside network, adding them to the incoming packet queue - while (1) { - pt = (struct ethernet_pkt *)V3_Malloc(sizeof(struct ethernet_pkt)); - - if (pt == NULL){ - PrintError("Vnet: process_udp: Malloc fails\n"); - continue; - } - - PrintDebug("Vnet: route_thread: socket: [%d]. ready to receive from ip [%x], port [%d] or from VMs\n", link_sock, (uint_t)dest, remote_port); - pt->size = V3_RecvFrom_IP( link_sock, dest, remote_port, pt->data, length); - PrintDebug("Vnet: route_thread: socket: [%d] receive from ip [%x], port [%d]\n", link_sock, (uint_t)dest, remote_port); - - if (pt->size <= 0) { - PrintDebug("Vnet: process_udp: receiving packet from UDP fails\n"); - V3_Free(pt); - return -1; - } - - PrintDebug("Vnet: process_udp: get packet\n"); - print_packet(pt->data, pt->size); - } -} - -// TODO: To be removed -static int indata_handler( ) -{ - if (!use_tcp) { - process_udpdata(); - } - - return 0; -} - -// TODO: To be removed -static int start_recv_data() -{ - if (use_tcp){ - - } else { - SOCK udp_data_socket; - - if ((udp_data_socket = V3_Create_UDP_Socket()) < 0){ - PrintError("VNET: Can't setup udp socket\n"); - return -1; - } - PrintDebug("Vnet: vnet_setup_udp: get socket: %d\n", udp_data_socket); - g_udp_sockfd = udp_data_socket; - - store_topologies(udp_data_socket); - - if (V3_Bind_Socket(udp_data_socket, vnet_udp_port) < 0){ - PrintError("VNET: Can't bind socket\n"); - return -1; - } - PrintDebug("VNET: vnet_setup_udp: bind socket successful\n"); - } - - V3_CREATE_THREAD(&indata_handler, NULL, "VNET_DATA_HANDLER"); - return 0; -} - -static void store_test_topologies(SOCK fd) -{ - int i; - int src_mac_qual = MAC_ANY; - int dest_mac_qual = MAC_ANY; - uint_t dest; -#ifndef VNET_SERVER - dest = (0 | 172 << 24 | 23 << 16 | 1 ); - PrintDebug("VNET: store_topologies. NOT VNET_SERVER, dest = %x\n", dest); -#else - dest = (0 | 172 << 24 | 23 << 16 | 2 ); - PrintDebug("VNET: store_topologies. VNET_SERVER, dest = %x\n", dest); -#endif - - int type = UDP_TYPE; - int src = 0; - int src_type= LINK_ANY; //ANY_SRC_TYPE - int data_port = 22; -} - -#endif - -static int handle_one_pkt(struct ethernet_pkt * pkt) { +static int handle_one_pkt(struct ethernet_pkt *pkt) { int src_link_index = -1; //the value of src_link_index of udp always is 0 int i; char src_mac[6]; @@ -687,81 +613,109 @@ static int handle_one_pkt(struct ethernet_pkt * pkt) { num_matched_routes = look_into_cache((route_hashkey_t)hash_key, matches); - if (num_matched_routes == -1) { - // no match - + // no match in the cache num_matched_routes = match_route(src_mac, dst_mac, LINK_ANY, src_link_index, matches); - if (num_matched_routes > 0) { - add_route_to_cache(hash_key, num_matched_routes,matches); - } + if (num_matched_routes > 0) { + add_route_to_cache(hash_key, num_matched_routes,matches); + } } PrintDebug("Vnet: HandleDataOverLink: Matches=%d\n", num_matched_routes); - for (i = 0; i < num_matched_routes; i++) { + for (i = 0; i < num_matched_routes; i++) {//send packet to all destinations int route_index = -1; int link_index = -1; - int pkt_len = 0; - struct link_entry * link = NULL; + int pkt_len = 0; + struct link_entry * link = NULL; route_index = matches[i]; PrintDebug("Vnet: HandleDataOverLink: Forward packet from link according to Route entry %d\n", route_index); - link_index = g_routes.routes[route_index]->link_idx; + link_index = g_routes.routes[route_index]->link_idx; - if ( (link_index < 0) || + if ((link_index < 0) || (link_index > MAX_LINKS) || (g_links.links[link_index] == NULL)) { - continue; - } - - link = g_links.links[link_index]; - + continue; + } - pkt_len = pkt->size; - + link = g_links.links[link_index]; + pkt_len = pkt->size; if (g_routes.routes[route_index]->link_type == LINK_EDGE) { - - // TODO: apply the header in the beginning of the packet to be sent - if ((link->dst_link->input(pkt->data, pkt_len, NULL)) != pkt_len) { - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); - return -1; - } + + //apply the header in the beginning of the packet to be sent + if (link->dst_link->pro_type == UDP_TYPE) { + struct udp_link_header *hdr = &(link->dst_link->vnet_header); + struct ip_header *ip = &hdr->ip_hdr; + struct udp_header *udp = &hdr->udp_hdr; + udp->len = pkt_len + sizeof(struct udp_header); + ip->total_len = pkt_len + sizeof(struct udp_header) + sizeof(struct ip_header); + ip->cksum = ip_xsum(ip, sizeof(struct ip_header)); + + int hdr_size = sizeof(struct udp_link_header); + memcpy(&pkt->ext_hdr, hdr, hdr_size); + + pkt_len += hdr_size; + if ((link->dst_link->input((uchar_t *)&pkt->ext_hdr, pkt_len, link->dst_link->private_data)) != pkt_len) { + PrintDebug("VNET: Packet not sent properly\n"); + return -1; + } + }else { + PrintDebug("VNET: Link protocol type not support\n"); + return -1; + } } else if (g_routes.routes[route_index]->link_type == LINK_INTERFACE) { - if ((link->dst_link->input(pkt->data, pkt_len, NULL)) != pkt_len) { - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); - return -1; - } + if ((link->dst_dev->input(pkt->data, pkt_len, link->dst_dev->private_data)) != pkt_len) { + PrintDebug("VNET: Packet not sent properly\n"); + return -1; + } } else { - PrintError("Vnet: Wrong Edge type\n"); - return -1; + PrintDebug("Vnet: Wrong Edge type\n"); + return -1; } } return 0; } -static int send_ethernet_pkt(char * buf, int length) { - struct ethernet_pkt * pkt; +static int send_ethernet_pkt(uchar_t *data, int len) { + struct ethernet_pkt *pkt; pkt = (struct ethernet_pkt *)V3_Malloc(sizeof(struct ethernet_pkt)); - ethernet_packet_init(pkt, buf, length); //====here we copy sending data once + memset(pkt, 0, sizeof(struct ethernet_pkt)); + + if(pkt == NULL){ + PrintError("VNET: Memory allocate fails\n"); + return -1; + } + + ethernet_packet_init(pkt, data, len); //====here we copy sending data once PrintDebug("VNET: vm_send_pkt: transmitting packet: (size:%d)\n", (int)pkt->size); - print_packet((char *)buf, length); + print_packet((char *)data, len); v3_enqueue(g_inpkt_q, (addr_t)pkt); return 0; } -int v3_vnet_send_pkt(uchar_t * buf, int length) { - PrintDebug("VNET: In V3_Send_pkt: pkt length %d\n", length); +//send raw ethernet packet +int v3_vnet_send_rawpkt(uchar_t * buf, int len, void *private_data) { + PrintDebug("VNET: In V3_Send_pkt: pkt length %d\n", len); - return send_ethernet_pkt((char *)buf, length); + return send_ethernet_pkt(buf, len); +} + +//sending the packet from Dom0, should remove the link header +int v3_vnet_send_udppkt(uchar_t * buf, int len, void *private_data) { + PrintDebug("VNET: In V3_Send_pkt: pkt length %d\n", len); + + uint_t hdr_len = sizeof(struct udp_link_header); + + return send_ethernet_pkt((uchar_t *)(buf+hdr_len), len - hdr_len); } static int search_device(char * device_name) { @@ -774,12 +728,11 @@ static int search_device(char * device_name) { } } } - - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); + return -1; } -static int vnet_register_device(struct vm_device * vdev, +int vnet_register_device(struct vm_device * vdev, char * dev_name, uchar_t mac[6], int (*netif_input)(uchar_t * pkt, uint_t size, void * private_data), @@ -789,11 +742,11 @@ static int vnet_register_device(struct vm_device * vdev, int idx = search_device(dev_name); if (idx != -1) { - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); + PrintDebug("VNET: register device: Already has device with the name %s\n", dev_name); return -1; } - if_dev = (struct vnet_if_device *)V3_Malloc(sizeof(struct vnet_if_device)); + if_dev = (struct vnet_if_device *)(sizeof(struct vnet_if_device)); if (if_dev == NULL) { PrintError("VNET: Malloc fails\n"); @@ -806,29 +759,32 @@ static int vnet_register_device(struct vm_device * vdev, if_dev->input = netif_input; if_dev->private_data = data; - struct link_entry * link = (struct link_entry *)V3_Malloc(sizeof(struct link_entry)); + struct link_entry * link = (struct link_entry *)(sizeof(struct link_entry)); link->type = LINK_INTERFACE; link->dst_dev = if_dev; - idx = add_link_entry(link); + idx = __add_link_entry(link); + + PrintDebug("VNET: Add device %s to link table, idx %d", dev_name, idx); return idx; } +#if 0 static int vnet_unregister_device(char * dev_name) { int idx; idx = search_device(dev_name); if (idx == -1) { - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); + PrintDebug("VNET: No device with name %s found\n", dev_name); return -1; } - struct vnet_if_device * device = (struct vnet_if_device *)delete_link_entry(idx); + struct vnet_if_device * device = (struct vnet_if_device *)__delete_link_entry(idx); if (device == NULL) { - PrintError("LEI -- WHAT IS THE ERROR HERE?\n"); + PrintError("VNET: Device %s not in the link table %d, something may be wrong in link table\n", dev_name, idx); return -1; } @@ -837,6 +793,8 @@ static int vnet_unregister_device(char * dev_name) { return idx; } +#endif + int v3_vnet_pkt_process() { struct ethernet_pkt * pkt; @@ -892,27 +850,118 @@ static void init_pkt_queue() { v3_init_queue(g_inpkt_q); } -#if 0 -// TODO: -static int init_routing_tables(struct routing_entry *route_tab, uint16_t size) -{ - //struct routing_entry *route; +static void free_link_mem(struct link_entry *link){ + V3_Free(link->dst_dev); + V3_Free(link); +} +// TODO: +static int addto_routing_link_tables(struct routing_entry *route_tab, + uint16_t num_routes, + struct link_entry *link_tab, + uint16_t num_links){ + struct routing_entry *route, *new_route; + struct link_entry *link, *new_link; + int new_idx; + int i; + int link_idxs[MAX_LINKS]; - return 0; -} + //add all of the links first, record their new indexs + for (i = 0; i < num_links; i++) { + link_idxs[i] = -1; + link = &link_tab[i]; + new_link = (struct link_entry *)V3_Malloc(sizeof(struct link_entry)); + if (new_link == NULL){ + PrintError("VNET: Memory allocate error\n"); + return -1; + } + + new_link->type = link->type; + + //TODO: how to set the input parameters here + if (link->type == LINK_EDGE){ + struct vnet_if_device *ifdev = (struct vnet_if_device *)V3_Malloc(sizeof(struct vnet_if_device)); + + if (ifdev == NULL){ + PrintError("VNET: Memory allocate fails\n"); + return -1; + } + + memcpy(ifdev->name, link->dst_dev->name, DEVICE_NAME_LEN); + + // TODO: + //ifdev->mac_addr + //ifdev->input + //ifdev->private_data + + new_link->dst_dev = ifdev; + }else if (link->type == LINK_INTERFACE){ + struct vnet_if_link *iflink = (struct vnet_if_link *)V3_Malloc(sizeof(struct vnet_if_link)); + + if (iflink == NULL){ + PrintError("VNET: Memory allocate fails\n"); + return -1; + } + iflink->pro_type = link->dst_link->pro_type; + iflink->dest_ip = link->dst_link->dest_ip; + iflink->dest_port = link->dst_link->dest_port; + memcpy(&iflink->vnet_header, &link->dst_link->vnet_header, sizeof(struct udp_link_header)); + + // TODO: + //iflink->input = + //iflink->private_data = + + new_link->dst_link = iflink; + }else{ + PrintDebug("VNET: invalid link type\n"); + V3_Free(new_link); + continue; + } + + new_idx = __add_link_entry(new_link); + if (new_idx < 0) { + PrintError("VNET: Adding link fails\n"); + free_link_mem(new_link); + continue; + } + link_idxs[i] = new_idx; + } + + //add all of routes, replace with new link indexs + for (i = 0; i < num_routes; i++) { + route = &route_tab[i]; + if (route->link_idx < 0 || route->link_idx >= num_links || + ((route->src_link_idx != -1) && + (route->src_link_idx < 0 || route->src_link_idx >= num_links))){ + PrintError("VNET: There is error in the intial tables data from guest\n"); + continue; + } + + new_route = (struct routing_entry *)V3_Malloc(sizeof(struct routing_entry)); + + if (new_route == NULL){ + PrintError("VNET: Memory allocate fails\n"); + return -1; + } + memcpy(new_route, route, sizeof(struct routing_entry)); + + new_route->link_idx = link_idxs[new_route->link_idx]; + + if (route->src_link_idx != -1) + new_route->src_link_idx = link_idxs[new_route->src_link_idx]; + + if((__add_route_entry(new_route)) == -1){ + PrintDebug("VNET: Adding route fails"); + V3_Free(new_route); + } + new_route = NULL; + } -// TODO: -static int init_link_tables(struct link_entry *link_tab, uint16_t size) -{ - //struct link_entry *link; return 0; } -#endif - struct table_init_info { addr_t routing_table_start; uint16_t routing_table_size; @@ -920,16 +969,34 @@ struct table_init_info { uint16_t link_table_size; }; +//add the guest specified routes and links to the tables static int handle_init_tables_hcall(struct guest_info * info, uint_t hcall_id, void * priv_data) { - uint8_t * buf = NULL; - addr_t info_addr = (addr_t)info->vm_regs.rcx; + addr_t guest_addr = (addr_t)info->vm_regs.rcx; + addr_t info_addr, route_addr, link_addr; + struct table_init_info *init_info; + struct link_entry *link_array; + struct routing_entry *route_array; + + if (guest_va_to_host_va(info, guest_addr, &info_addr) == -1) { + PrintError("VNET: Could not translate guest address\n"); + return -1; + } + + init_info = (struct table_init_info *)info_addr; - if (guest_va_to_host_va(info, info_addr, (addr_t *)&(buf)) == -1) { - PrintError("Could not translate buffer address\n"); + if (guest_va_to_host_va(info, init_info->routing_table_start, &route_addr) == -1) { + PrintError("VNET: Could not translate guest address\n"); return -1; } + route_array = (struct routing_entry *)route_addr; - //struct table_init_info *init_info = (struct table_init_info *)buf; + if (guest_va_to_host_va(info, init_info->link_table_start, &link_addr) == -1) { + PrintError("VNET: Could not translate guest address\n"); + return -1; + } + link_array = (struct link_entry *)link_addr; + + addto_routing_link_tables(route_array, init_info->routing_table_size, link_array, init_info->link_table_size); return 0; } @@ -941,8 +1008,6 @@ void v3_vnet_init(struct guest_info * vm) { v3_register_hypercall(vm, VNET_INITAB_HCALL, handle_init_tables_hcall, NULL); - //store_test_topologies(udp_data_socket); - PrintDebug("VNET Initialized\n"); }