From: Keir Fraser <keir.xen@gmail.com>
To: Bei Guan <gbtju85@gmail.com>, Xen Devel <xen-devel@lists.xensource.com>
Cc: edk2-devel@lists.sourceforge.net,
Jordan Justen <jljusten@gmail.com>,
Tim Deegan <Tim.Deegan@citrix.com>,
Andrei Warkentin <andreiw@motorola.com>
Subject: Re: [PATCH 1 of 3] Enable UEFI BIOS(OVMF) support in Xen-unstable HVM
Date: Fri, 22 Jul 2011 17:38:31 +0100 [thread overview]
Message-ID: <CA4F6617.1E6D5%keir.xen@gmail.com> (raw)
In-Reply-To: <CAEQjb-QbaJXHc+O_oJDrxBX1P-fyHjwnUs+KWxDD2dozoMEiUw@mail.gmail.com>
On 22/07/2011 17:23, "Bei Guan" <gbtju85@gmail.com> wrote:
> Hi,
>
> My name is Bei Guan and I am one of this year's GSOS students for Tianocore.
> My project is to enable Xen support in OVMF and the following is about my
> patch for Xen-unstable. Could you give me some comments? Thank you very much.
>
> This patch can enable Xen-unstable hvmloader to load OVMF BIOS as Xen HVM UEFI
> support. It supports OVMF BIOS in IA32 and X86 environment. The loaded OVMF
> BIOS can get Xen SMBIOS and ACPI tables contents inside itself.
>
> In order to be clear, I divide the patch into three parts:
> ovmf_xen_support.patch Enable Xen hvmloader to load OVMF
Looks pretty decent. I wonder why you need to change get_shared_info() --
the existing mapping location is unused at the time hvmloader runs, and you
instead map it over the top of a page of RAM. If you want shared_info mapped
elsewhere, you can map it wherever you like as soon as your BIOS payload
takes over.
I'd also need to see what you use mem_back_ram() for, and maybe give it a
better name, but I'm not against extracting that functionality into a
function for the use of BIOS-specific handlers if the use is sane.
-- Keir
> ovmf_firmware.patch OVMF binary files
> ovmf_xl_xend.patch Add hvmloader/bios xenstore key in libxl and xend
>
> Usage:
> Add an option field in HVM config file.
> # OVMF support. When enabled, hvmloader can load OVMF bios of
> IA32("ovmf-ia32") and X64("ovmf-x64")
> hvmbios = "ovmf-ia32"
> #hvmbios = "ovmf-x64"
>
> Note:
> You should enable the HVM guest ACPI: acpi=1
> You can use the OVMF to boot into a UEFI-aware OS, such as
> ubuntu-10.10-desktop-amd64.
> iso. Just set the "disk" option like this:
> disk = [ 'file:/root/<img_name>.img,ioemu:hda,w',
> 'file:/root/ubuntu-10.10-desktop-amd64.iso,hdc:cdrom,r' ]
>
>
> ovmf_xen_support.patch:
> ------
>
> # HG changeset patch
> # User gbtju85@gmail.com
> #
>
> diff -r e298ce67777e tools/firmware/hvmloader/Makefile
> --- a/tools/firmware/hvmloader/Makefile Mon Jul 18 14:38:31 2011 +0100
> +++ b/tools/firmware/hvmloader/Makefile Fri Jul 22 23:00:20 2011 +0800
> @@ -43,6 +43,19 @@
> CFLAGS += -DENABLE_ROMBIOS
> ROMBIOS_ROM := $(ROMBIOS_DIR)/BIOS-bochs-latest
> endif
> +OVMF_DIR := ../ovmf
> +OVMF32_ROM := $(OVMF_DIR)/ovmf-ia32.bin
> +OVMF64_ROM := $(OVMF_DIR)/ovmf-x64.bin
> +OVMF32_CIRRUS_VGA_ROM := $(OVMF_DIR)/ovmf-ia32-cirrus-vga.bin
> +OVMF64_CIRRUS_VGA_ROM := $(OVMF_DIR)/ovmf-x64-cirrus-vga.bin
> +
> +ifneq ($(OVMF32_ROM),)
> +OBJS += ovmf.o
> +endif
> +
> +ifneq ($(OVMF64_ROM),)
> +OBJS += ovmf.o
> +endif
>
> ifneq ($(SEABIOS_DIR),)
> OBJS += seabios.o
> @@ -69,7 +82,7 @@
> $(OBJCOPY) hvmloader.tmp hvmloader
> rm -f hvmloader.tmp
>
> -roms.inc: $(ROMBIOS_ROM) $(SEABIOS_ROM) $(STDVGA_ROM) $(CIRRUSVGA_ROM)
> ../etherboot/eb-roms.h
> +roms.inc: $(ROMBIOS_ROM) $(SEABIOS_ROM) $(STDVGA_ROM) $(CIRRUSVGA_ROM)
> $(OVMF32_ROM) $(OVMF64_ROM) $(OVMF32_CIRRUS_VGA_ROM) $(OVMF64_CIRRUS_VGA_ROM)
> ../etherboot/eb-roms.h
> echo "/* Autogenerated file. DO NOT EDIT */" > $@.new
>
> ifneq ($(ROMBIOS_ROM),)
> @@ -84,6 +97,30 @@
> echo "#endif" >> $@.new
> endif
>
> +ifneq ($(OVMF32_ROM),)
> + echo "#ifdef ROM_INCLUDE_OVMF32" >> $@.new
> + sh ./mkhex ovmf32 $(OVMF32_ROM) >> $@.new
> + echo "#endif" >> $@.new
> +endif
> +
> +ifneq ($(OVMF64_ROM),)
> + echo "#ifdef ROM_INCLUDE_OVMF64" >> $@.new
> + sh ./mkhex ovmf64 $(OVMF64_ROM) >> $@.new
> + echo "#endif" >> $@.new
> +endif
> +
> +ifneq ($(OVMF32_CIRRUS_VGA_ROM),)
> + echo "#ifdef ROM_INCLUDE_OVMF32_CIRRUS_VGA" >> $@.new
> + sh ./mkhex ovmf32_cirrus_vga $(OVMF32_CIRRUS_VGA_ROM) >> $@.new
> + echo "#endif" >> $@.new
> +endif
> +
> +ifneq ($(OVMF64_CIRRUS_VGA_ROM),)
> + echo "#ifdef ROM_INCLUDE_OVMF64_CIRRUS_VGA" >> $@.new
> + sh ./mkhex ovmf64_cirrus_vga $(OVMF64_CIRRUS_VGA_ROM) >> $@.new
> + echo "#endif" >> $@.new
> +endif
> +
> ifneq ($(STDVGA_ROM),)
> echo "#ifdef ROM_INCLUDE_VGABIOS" >> $@.new
> sh ./mkhex vgabios_stdvga $(STDVGA_ROM) >> $@.new
> diff -r e298ce67777e tools/firmware/hvmloader/config.h
> --- a/tools/firmware/hvmloader/config.h Mon Jul 18 14:38:31 2011 +0100
> +++ b/tools/firmware/hvmloader/config.h Fri Jul 22 23:00:20 2011 +0800
> @@ -3,7 +3,7 @@
>
> #include <stdint.h>
>
> -enum virtual_vga { VGA_none, VGA_std, VGA_cirrus, VGA_pt };
> +enum virtual_vga { VGA_none, VGA_std, VGA_cirrus, VGA_pt, VGA_custom };
> extern enum virtual_vga virtual_vga;
>
> struct bios_config {
> @@ -16,6 +16,9 @@
> /* Physical address to load at */
> unsigned int bios_address;
>
> + /* Custom load function. */
> + void (*load)(const struct bios_config *config);
> + void (*pci_setup)(void);
> /* ROMS */
> int load_roms;
> unsigned int optionrom_start, optionrom_end;
> @@ -36,6 +39,8 @@
>
> extern struct bios_config rombios_config;
> extern struct bios_config seabios_config;
> +extern struct bios_config ovmf32_config;
> +extern struct bios_config ovmf64_config;
>
> #define PAGE_SHIFT 12
> #define PAGE_SIZE (1ul << PAGE_SHIFT)
> diff -r e298ce67777e tools/firmware/hvmloader/hvmloader.c
> --- a/tools/firmware/hvmloader/hvmloader.c Mon Jul 18 14:38:31 2011 +0100
> +++ b/tools/firmware/hvmloader/hvmloader.c Fri Jul 22 23:00:20 2011 +0800
> @@ -360,6 +360,8 @@
> #ifdef ENABLE_SEABIOS
> { "seabios", &seabios_config, },
> #endif
> + { "ovmf-ia32", &ovmf32_config, },
> + { "ovmf-x64", &ovmf64_config, },
> { NULL, NULL }
> };
>
> @@ -416,9 +418,13 @@
> bios->create_smbios_tables();
> }
>
> - printf("Loading %s ...\n", bios->name);
> - memcpy((void *)bios->bios_address, bios->image,
> - bios->image_size);
> + if (bios->load) {
> + bios->load(bios);
> + } else {
> + printf("Loading %s ...\n", bios->name);
> + memcpy((void *)bios->bios_address, bios->image,
> + bios->image_size);
> + }
>
> if (bios->bios_relocate)
> bios->bios_relocate();
> @@ -451,8 +457,10 @@
> vgabios_sz = round_option_rom(
> (*(uint8_t *)(VGABIOS_PHYSICAL_ADDRESS+2)) * 512);
> break;
> + case VGA_custom:
> + break;
> default:
> - printf("No emulated VGA adaptor ...\n");
> + printf("No emulated VGA adaptor ROM...\n");
> break;
> }
>
> diff -r e298ce67777e tools/firmware/hvmloader/ovmf.c
> --- /dev/null Thu Jan 01 00:00:00 1970 +0000
> +++ b/tools/firmware/hvmloader/ovmf.c Fri Jul 22 23:00:20 2011 +0800
> @@ -0,0 +1,189 @@
> +/*
> + * HVM OVMF UEFI support.
> + *
> + * Bei Guan, gbtju85@gmail.com
> + * Andrei Warkentin, andreiw@motorola.com
> + * Leendert van Doorn, leendert@watson.ibm.com
> + * Copyright (c) 2005, International Business Machines Corporation.
> + * Copyright (c) 2006, Keir Fraser, XenSource Inc.
> + * Copyright (c) 2011, Citrix Inc.
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along
> with
> + * this program; if not, write to the Free Software Foundation, Inc., 59
> Temple
> + * Place - Suite 330, Boston, MA 02111-1307 USA.
> + */
> +
> +#include "config.h"
> +#include "smbios_types.h"
> +#include "acpi/acpi2_0.h"
> +#include "apic_regs.h"
> +#include "../rombios/config.h"
> +#include "util.h"
> +#include "pci_regs.h"
> +#include "hypercall.h"
> +
> +#include <xen/hvm/params.h>
> +#include <xen/hvm/ioreq.h>
> +#include <xen/memory.h>
> +
> +#define ROM_INCLUDE_OVMF32
> +#define ROM_INCLUDE_OVMF64
> +#define ROM_INCLUDE_OVMF32_CIRRUS_VGA
> +#define ROM_INCLUDE_OVMF64_CIRRUS_VGA
> +#include "roms.inc"
> +
> +#define OVMF_BEGIN 0xFFF00000ULL
> +#define OVMF_SIZE 0x00100000ULL
> +#define OVMF_MAXOFFSET 0x000FFFFFULL
> +#define OVMF_END (OVMF_BEGIN + OVMF_SIZE)
> +#define LOWCHUNK_BEGIN 0x000F0000
> +#define LOWCHUNK_SIZE 0x00010000
> +#define LOWCHUNK_MAXOFFSET 0x0000FFFF
> +#define LOWCHUNK_END (OVMF_BEGIN + OVMF_SIZE)
> +
> +/*
> + * Set up an empty TSS area for virtual 8086 mode to use.
> + * The only important thing is that it musn't have any bits set
> + * in the interrupt redirection bitmap, so all zeros will do.
> + */
> +static void ovmf_init_vm86_tss(void)
> +{
> + void *tss;
> + struct xen_hvm_param p;
> +
> + tss = mem_alloc(128, 128);
> + memset(tss, 0, 128);
> + p.domid = DOMID_SELF;
> + p.index = HVM_PARAM_VM86_TSS;
> + p.value = virt_to_phys(tss);
> + hypercall_hvm_op(HVMOP_set_param, &p);
> + printf("vm86 TSS at %08lx\n", virt_to_phys(tss));
> +}
> +
> +static void ovmf_load(const struct bios_config *config)
> +{
> + xen_pfn_t mfn;
> + uint64_t addr = OVMF_BEGIN;
> +
> + virtual_vga = VGA_custom;
> +
> + /* Copy video ROM. */
> + if (config == &ovmf32_config) {
> + memcpy((void *)VGABIOS_PHYSICAL_ADDRESS,
> + ovmf32_cirrus_vga, sizeof(ovmf32_cirrus_vga));
> + printf("OVMF32 Cirrus [0x%x-0x%x]\n", VGABIOS_PHYSICAL_ADDRESS,
> + VGABIOS_PHYSICAL_ADDRESS + sizeof(ovmf32_cirrus_vga));
> + } else if (config == &ovmf64_config) {
> + memcpy((void *)VGABIOS_PHYSICAL_ADDRESS,
> + ovmf64_cirrus_vga, sizeof(ovmf64_cirrus_vga));
> + printf("OVMF64 Cirrus [0x%x-0x%x]\n", VGABIOS_PHYSICAL_ADDRESS,
> + VGABIOS_PHYSICAL_ADDRESS + sizeof(ovmf64_cirrus_vga));
> + }
> +
> + /* Copy low-reset vector portion. */
> + memcpy((void *) LOWCHUNK_BEGIN, (uint8_t *) config->image
> + + OVMF_SIZE
> + - LOWCHUNK_SIZE,
> + LOWCHUNK_SIZE);
> +
> + /* Ensure we have backing page prior to moving FD. */
> + while ((addr >> PAGE_SHIFT) != (OVMF_END >> PAGE_SHIFT)) {
> + mfn = (uint32_t) (addr >> PAGE_SHIFT);
> + addr += PAGE_SIZE;
> +
> + BUG_ON(mem_back_ram(mfn));
> + }
> +
> + printf("Initialized FD backing pages...\n");
> +
> + /* Copy FD. */
> + memcpy((void *) OVMF_BEGIN, config->image, OVMF_SIZE);
> + printf("Load complete!\n");
> +}
> +
> +static void ovmf_acpi_build_tables(void)
> +{
> + acpi_build_tables(ACPI_PHYSICAL_ADDRESS);
> +}
> +
> +static void ovmf_create_smbios_tables(void)
> +{
> + hvm_write_smbios_tables(SMBIOS_PHYSICAL_ADDRESS,
> + SMBIOS_PHYSICAL_ADDRESS + sizeof(struct
> smbios_entry_point),
> + SMBIOS_PHYSICAL_END);
> +}
> +
> +struct bios_config ovmf32_config = {
> + .name = "OVMF-IA32",
> +
> + .image = ovmf32,
> + .image_size = sizeof(ovmf32),
> +
> + .bios_address = 0,
> + .load = ovmf_load,
> +
> + .load_roms = 0,
> +
> + .optionrom_start = 0,
> + .optionrom_end = 0,
> +
> + .bios_info_setup = NULL,
> + .bios_info_finish = NULL,
> +
> + .bios_relocate = NULL,
> +
> + .vm86_setup = ovmf_init_vm86_tss,
> + .e820_setup = NULL,
> +
> + .acpi_build_tables = ovmf_acpi_build_tables,
> + .create_mp_tables = NULL,
> + .create_smbios_tables = ovmf_create_smbios_tables,
> + .create_pir_tables = NULL,
> +};
> +
> +struct bios_config ovmf64_config = {
> + .name = "OVMF-X64",
> +
> + .image = ovmf64,
> + .image_size = sizeof(ovmf64),
> +
> + .bios_address = 0,
> + .load = ovmf_load,
> +
> + .load_roms = 0,
> +
> + .optionrom_start = 0,
> + .optionrom_end = 0,
> +
> + .bios_info_setup = NULL,
> + .bios_info_finish = NULL,
> +
> + .bios_relocate = NULL,
> +
> + .vm86_setup = ovmf_init_vm86_tss,
> + .e820_setup = NULL,
> +
> + .acpi_build_tables = ovmf_acpi_build_tables,
> + .create_mp_tables = NULL,
> + .create_smbios_tables = ovmf_create_smbios_tables,
> + .create_pir_tables = NULL,
> +};
> +
> +/*
> + * Local variables:
> + * mode: C
> + * c-set-style: "BSD"
> + * c-basic-offset: 4
> + * tab-width: 4
> + * indent-tabs-mode: nil
> + * End:
> + */
> diff -r e298ce67777e tools/firmware/hvmloader/util.c
> --- a/tools/firmware/hvmloader/util.c Mon Jul 18 14:38:31 2011 +0100
> +++ b/tools/firmware/hvmloader/util.c Fri Jul 22 23:00:20 2011 +0800
> @@ -303,12 +303,54 @@
> *p = '\0';
> }
>
> +/*
> + * Ensures mfn is backed by an accessible RAM page.
> + * Returns 0 on success.
> + */
> +int mem_back_ram(xen_pfn_t mfn)
> +{
> + struct xen_add_to_physmap xatp;
> + struct xen_memory_reservation xmr;
> + static int over_allocated = 0;
> +
> + if ( !over_allocated )
> + {
> + xmr.domid = DOMID_SELF;
> + xmr.mem_flags = 0;
> + xmr.extent_order = 0;
> + xmr.nr_extents = 1;
> + set_xen_guest_handle(xmr.extent_start, &mfn);
> + if ( hypercall_memory_op(XENMEM_populate_physmap, &xmr) == 1 )
> + return 0;
> + over_allocated = 1;
> + }
> +
> + /*
> + * Couldn't allocate more memory for domain,
> + * move an existing physical page from end
> + * of RAM.
> + */
> + if ( hvm_info->high_mem_pgend )
> + {
> + xatp.idx = --hvm_info->high_mem_pgend;
> + if ( xatp.idx == (1ull << (32 - PAGE_SHIFT)) )
> + hvm_info->high_mem_pgend = 0;
> + }
> + else
> + {
> + xatp.idx = --hvm_info->low_mem_pgend;
> + }
> + xatp.domid = DOMID_SELF;
> + xatp.space = XENMAPSPACE_gmfn;
> + xatp.gpfn = mfn;
> + if ( hypercall_memory_op(XENMEM_add_to_physmap, &xatp) != 0 )
> + BUG();
> + return 0;
> +}
> +
> void *mem_alloc(uint32_t size, uint32_t align)
> {
> static uint32_t reserve = RESERVED_MEMBASE - 1;
> - static int over_allocated;
> - struct xen_add_to_physmap xatp;
> - struct xen_memory_reservation xmr;
> xen_pfn_t mfn;
> uint32_t s, e;
>
> @@ -326,35 +368,7 @@
> reserve += PAGE_SIZE;
> mfn = reserve >> PAGE_SHIFT;
>
> - /* Try to allocate a brand new page in the reserved area. */
> - if ( !over_allocated )
> - {
> - xmr.domid = DOMID_SELF;
> - xmr.mem_flags = 0;
> - xmr.extent_order = 0;
> - xmr.nr_extents = 1;
> - set_xen_guest_handle(xmr.extent_start, &mfn);
> - if ( hypercall_memory_op(XENMEM_populate_physmap, &xmr) == 1 )
> - continue;
> - over_allocated = 1;
> - }
> -
> - /* Otherwise, relocate a page from the ordinary RAM map. */
> - if ( hvm_info->high_mem_pgend )
> - {
> - xatp.idx = --hvm_info->high_mem_pgend;
> - if ( xatp.idx == (1ull << (32 - PAGE_SHIFT)) )
> - hvm_info->high_mem_pgend = 0;
> - }
> - else
> - {
> - xatp.idx = --hvm_info->low_mem_pgend;
> - }
> - xatp.domid = DOMID_SELF;
> - xatp.space = XENMAPSPACE_gmfn;
> - xatp.gpfn = mfn;
> - if ( hypercall_memory_op(XENMEM_add_to_physmap, &xatp) != 0 )
> - BUG();
> + BUG_ON(mem_back_ram(mfn));
> }
>
> reserve = e;
> @@ -624,22 +638,24 @@
> return table;
> }
>
> -struct shared_info *get_shared_info(void)
> +static struct shared_info *shared_info = NULL;
> +
> +struct shared_info *get_shared_info(void)
> {
> - static struct shared_info *shared_info = NULL;
> struct xen_add_to_physmap xatp;
>
> if ( shared_info != NULL )
> return shared_info;
>
> + /* Guarantee shinfo lives in a safe (reserved) place */
> + shared_info = mem_alloc(PAGE_SIZE, PAGE_SIZE);
> +
> xatp.domid = DOMID_SELF;
> xatp.space = XENMAPSPACE_shared_info;
> xatp.idx = 0;
> - xatp.gpfn = 0xfffffu;
> - shared_info = (struct shared_info *)(xatp.gpfn << PAGE_SHIFT);
> + xatp.gpfn = ((uint32_t) shared_info) >> PAGE_SHIFT;
> if ( hypercall_memory_op(XENMEM_add_to_physmap, &xatp) != 0 )
> BUG();
> -
> return shared_info;
> }
>
> diff -r e298ce67777e tools/firmware/hvmloader/util.h
> --- a/tools/firmware/hvmloader/util.h Mon Jul 18 14:38:31 2011 +0100
> +++ b/tools/firmware/hvmloader/util.h Fri Jul 22 23:00:20 2011 +0800
> @@ -3,6 +3,7 @@
>
> #include <stdarg.h>
> #include <stdint.h>
> +#include <xen/memory.h>
> #include <xen/hvm/hvm_info_table.h>
>
> #define __STR(...) #__VA_ARGS__
> @@ -164,6 +165,9 @@
> int printf(const char *fmt, ...) __attribute__ ((format (printf, 1, 2)));
> int vprintf(const char *fmt, va_list ap);
>
> +/* Create RAM backing for guest machine frame. */
> +int mem_back_ram(xen_pfn_t mfn);
> +
> /* Allocate memory in a reserved region below 4GB. */
> void *mem_alloc(uint32_t size, uint32_t align);
> #define virt_to_phys(v) ((unsigned long)(v))
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
next prev parent reply other threads:[~2011-07-22 16:38 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-07-22 16:23 [PATCH 1 of 3] Enable UEFI BIOS(OVMF) support in Xen-unstable HVM Bei Guan
2011-07-22 16:38 ` Keir Fraser [this message]
2011-07-22 18:38 ` Andrei Warkentin
2011-07-23 6:53 ` Keir Fraser
2011-07-23 9:06 ` Keir Fraser
2011-07-23 15:18 ` Bei Guan
2011-07-23 16:50 ` Keir Fraser
2011-07-25 14:03 ` Konrad Rzeszutek Wilk
2011-07-25 14:17 ` Keir Fraser
2011-07-25 15:12 ` Bei Guan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CA4F6617.1E6D5%keir.xen@gmail.com \
--to=keir.xen@gmail.com \
--cc=Tim.Deegan@citrix.com \
--cc=andreiw@motorola.com \
--cc=edk2-devel@lists.sourceforge.net \
--cc=gbtju85@gmail.com \
--cc=jljusten@gmail.com \
--cc=xen-devel@lists.xensource.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.