Re: [patches] [PATCH] [21/22] x86_64: Extend bzImage protocol forrelocatable bzImage
From: Rusty Russell
Date: Mon Apr 30 2007 - 23:39:27 EST
On Mon, 2007-04-30 at 09:34 -0600, Eric W. Biederman wrote:
> Reading this it occurs to me what I object to wasn't that clear.
>
> I have no problem with the testing of %cs to see if we are not in ring0.
> That part while a little odd is fine, and we will certainly need a test
> to skip the protected instructions in head.S
>
> What I object to in particular is having (struct lguest_info?) instead
> of using the standard format for kernel parameters pointed to in %esi.
Here's a rough patch to see what it looks like from an lguest POV. It's
an improvement in many ways: I chose to hardcode the search for matching
backend rather than use paravirt_probe-style magic.
It'd be nicer if there were a "struct boot_params" declaration, but we
can't have everything.
Cheers,
Rusty.
diff -r 9a673a220ad6 Documentation/lguest/lguest.c
--- a/Documentation/lguest/lguest.c Mon Apr 30 20:10:26 2007 +1000
+++ b/Documentation/lguest/lguest.c Tue May 01 13:19:02 2007 +1000
@@ -30,10 +30,12 @@
#include <termios.h>
#include <getopt.h>
#include <zlib.h>
+typedef unsigned long long u64;
typedef uint32_t u32;
typedef uint16_t u16;
typedef uint8_t u8;
#include "../../include/linux/lguest_launcher.h"
+#include "../../include/asm-i386/e820.h"
#define PAGE_PRESENT 0x7 /* Present, RW, Execute */
#define NET_PEERNUM 1
@@ -915,10 +917,10 @@ static void usage(void)
int main(int argc, char *argv[])
{
- unsigned long mem, pgdir, start, page_offset;
+ unsigned long mem, pgdir, start, page_offset, initrd_size = 0;
int c, lguest_fd, waker_fd;
struct device_list device_list;
- struct lguest_boot_info *boot = (void *)0;
+ void *boot = (void *)0;
const char *initrd_name = NULL;
device_list.max_infd = -1;
@@ -966,15 +968,24 @@ int main(int argc, char *argv[])
map_device_descriptors(&device_list, mem);
/* Map the initrd image if requested */
- if (initrd_name)
- boot->initrd_size = load_initrd(initrd_name, mem);
+ if (initrd_name) {
+ initrd_size = load_initrd(initrd_name, mem);
+ *(unsigned long *)(boot+0x218) = mem - initrd_size;
+ *(unsigned long *)(boot+0x21c) = initrd_size;
+ }
/* Set up the initial linar pagetables. */
- pgdir = setup_pagetables(mem, boot->initrd_size, page_offset);
-
- /* Give the guest the boot information it needs. */
- concat(boot->cmdline, argv+optind+2);
- boot->max_pfn = mem/getpagesize();
+ pgdir = setup_pagetables(mem, initrd_size, page_offset);
+
+ /* E820 memory map: ours is a simple, single region. */
+ *(char*)(boot+E820NR) = 1;
+ *((struct e820entry *)(boot+E820MAP))
+ = ((struct e820entry) { 0, mem, E820_RAM });
+ /* Command line pointer and command line (at 4096) */
+ *(void **)(boot + 0x228) = boot + 4096;
+ concat(boot + 4096, argv+optind+2);
+ /* Paravirt type: 1 == lguest */
+ *(int *)(boot + 0x23c) = 1;
lguest_fd = tell_kernel(pgdir, start, page_offset);
waker_fd = setup_waker(&device_list);
diff -r 9a673a220ad6 arch/i386/kernel/head.S
--- a/arch/i386/kernel/head.S Mon Apr 30 20:10:26 2007 +1000
+++ b/arch/i386/kernel/head.S Tue May 01 12:29:55 2007 +1000
@@ -504,34 +504,17 @@ ignore_int:
#ifdef CONFIG_PARAVIRT
startup_paravirt:
cld
+ movl %esi, %eax
+ addl $__PAGE_OFFSET, %eax
movl $(init_thread_union+THREAD_SIZE),%esp
- /* We take pains to preserve all the regs. */
- pushl %edx
- pushl %ecx
- pushl %eax
-
- pushl $__start_paravirtprobe
-1:
- movl 0(%esp), %eax
- cmpl $__stop_paravirtprobe, %eax
- je unhandled_paravirt
- pushl (%eax)
- movl 8(%esp), %eax
- call *(%esp)
- popl %eax
-
- movl 4(%esp), %eax
- movl 8(%esp), %ecx
- movl 12(%esp), %edx
-
- addl $4, (%esp)
- jmp 1b
-
-unhandled_paravirt:
+#ifdef CONFIG_LGUEST_GUEST
+ cmpl $1, 0x23c(%eax)
+ je lguest_init
+#endif
/* Nothing wanted us: we're screwed. */
ud2
-#endif
+#endif /* CONFIG_PARAVIRT */
/*
* Real beginning of normal "text" segment
diff -r 9a673a220ad6 drivers/lguest/lguest.c
--- a/drivers/lguest/lguest.c Mon Apr 30 20:10:26 2007 +1000
+++ b/drivers/lguest/lguest.c Tue May 01 13:28:06 2007 +1000
@@ -53,7 +53,6 @@ struct lguest_data lguest_data = {
.blocked_interrupts = { 1 }, /* Block timer interrupts */
};
struct lguest_device_desc *lguest_devices;
-static __initdata const struct lguest_boot_info *boot = __va(0);
static enum paravirt_lazy_mode lazy_mode;
static void lguest_lazy_mode(enum paravirt_lazy_mode mode)
@@ -378,8 +377,7 @@ static __init char *lguest_memory_setup(
/* We do this here because lockcheck barfs if before start_kernel */
atomic_notifier_chain_register(&panic_notifier_list, &paniced);
- e820.nr_map = 0;
- add_memory_region(0, PFN_PHYS(boot->max_pfn), E820_RAM);
+ add_memory_region(E820_MAP->addr, E820_MAP->size, E820_MAP->type);
return "LGUEST";
}
@@ -410,8 +408,14 @@ static unsigned lguest_patch(u8 type, u1
return insn_len;
}
-__init void lguest_init(void)
-{
+__init void lguest_init(void *boot)
+{
+ /* Copy boot parameters first. */
+ memcpy(boot_params, boot, PARAM_SIZE);
+ memcpy(boot_command_line,
+ __va(*(unsigned long *)(boot_params + NEW_CL_POINTER)),
+ COMMAND_LINE_SIZE);
+
paravirt_ops.name = "lguest";
paravirt_ops.paravirt_enabled = 1;
paravirt_ops.kernel_rpl = 1;
@@ -460,7 +464,6 @@ __init void lguest_init(void)
paravirt_ops.wbinvd = lguest_wbinvd;
hcall(LHCALL_LGUEST_INIT, __pa(&lguest_data), 0, 0);
- strncpy(boot_command_line, boot->cmdline, COMMAND_LINE_SIZE);
/* We use top of mem for initial pagetables. */
init_pg_tables_end = __pa(pg0);
@@ -487,14 +490,6 @@ __init void lguest_init(void)
add_preferred_console("hvc", 0, NULL);
- if (boot->initrd_size) {
- /* We stash this at top of memory. */
- INITRD_START = boot->max_pfn*PAGE_SIZE - boot->initrd_size;
- INITRD_SIZE = boot->initrd_size;
- LOADER_TYPE = 0xFF;
- }
-
pm_power_off = lguest_power_off;
start_kernel();
}
-paravirt_probe(lguest_maybe_init);
diff -r 9a673a220ad6 drivers/lguest/lguest_asm.S
--- a/drivers/lguest/lguest_asm.S Mon Apr 30 20:10:26 2007 +1000
+++ b/drivers/lguest/lguest_asm.S Tue May 01 12:37:25 2007 +1000
@@ -5,31 +5,12 @@
/* FIXME: Once asm/processor-flags.h goes in, include that */
#define X86_EFLAGS_IF 0x00000200
-/*
- * This is where we begin: head.S notes that paging is already enabled (which
- * doesn't happen in native boot) and calls the registered paravirt_probe
- * functions one at a time. Ours is a simple assembler test for magic
- * registers. If they're correct we jump to lguest_init.
- *
- * We put it in .init.text will be discarded after boot.
- */
-.section .init.text, "ax", @progbits
-ENTRY(lguest_maybe_init)
- cmpl $LGUEST_MAGIC_EBP, %ebp
- jne out
- cmpl $LGUEST_MAGIC_EDI, %edi
- jne out
- cmpl $LGUEST_MAGIC_ESI, %esi
- jne out
- je lguest_init
-out:
- ret
-
/* The templates for inline patching. */
#define LGUEST_PATCH(name, insns...) \
lgstart_##name: insns; lgend_##name:; \
.globl lgstart_##name; .globl lgend_##name
+.section .init.text, "ax", @progbits
LGUEST_PATCH(cli, movl $0, lguest_data+LGUEST_DATA_irq_enabled)
LGUEST_PATCH(sti, movl $X86_EFLAGS_IF, lguest_data+LGUEST_DATA_irq_enabled)
LGUEST_PATCH(popf, movl %eax, lguest_data+LGUEST_DATA_irq_enabled)
diff -r 9a673a220ad6 drivers/lguest/lguest_user.c
--- a/drivers/lguest/lguest_user.c Mon Apr 30 20:10:26 2007 +1000
+++ b/drivers/lguest/lguest_user.c Tue May 01 12:11:50 2007 +1000
@@ -7,13 +7,11 @@ static void setup_regs(struct lguest_reg
static void setup_regs(struct lguest_regs *regs, unsigned long start)
{
/* Write out stack in format lguest expects, so we can switch to it. */
- regs->edi = LGUEST_MAGIC_EDI;
- regs->ebp = LGUEST_MAGIC_EBP;
- regs->esi = LGUEST_MAGIC_ESI;
regs->ds = regs->es = regs->ss = __KERNEL_DS|GUEST_PL;
regs->cs = __KERNEL_CS|GUEST_PL;
regs->eflags = 0x202; /* Interrupts enabled. */
regs->eip = start;
+ /* esi points to our boot information (physical address 0) */
}
/* + addr */
diff -r 9a673a220ad6 include/asm-i386/paravirt.h
--- a/include/asm-i386/paravirt.h Mon Apr 30 20:10:26 2007 +1000
+++ b/include/asm-i386/paravirt.h Tue May 01 11:35:17 2007 +1000
@@ -221,11 +221,6 @@ struct paravirt_ops
void (*irq_enable_sysexit)(void);
void (*iret)(void);
};
-
-/* Mark a paravirt probe function. */
-#define paravirt_probe(fn) \
- static asmlinkage void (*__paravirtprobe_##fn)(void) __attribute_used__ \
- __attribute__((__section__(".paravirtprobe"))) = fn
extern struct paravirt_ops paravirt_ops;
diff -r 9a673a220ad6 include/linux/lguest.h
--- a/include/linux/lguest.h Mon Apr 30 20:10:26 2007 +1000
+++ b/include/linux/lguest.h Tue May 01 12:11:07 2007 +1000
@@ -2,11 +2,6 @@
* this is subject to wild and random change between versions. */
#ifndef _ASM_LGUEST_H
#define _ASM_LGUEST_H
-
-/* These are randomly chosen numbers which indicate we're an lguest at boot */
-#define LGUEST_MAGIC_EBP 0x4C687970
-#define LGUEST_MAGIC_EDI 0x652D4D65
-#define LGUEST_MAGIC_ESI 0xFFFFFFFF
#ifndef __ASSEMBLY__
#include <asm/irq.h>
diff -r 9a673a220ad6 include/linux/lguest_launcher.h
--- a/include/linux/lguest_launcher.h Mon Apr 30 20:10:26 2007 +1000
+++ b/include/linux/lguest_launcher.h Tue May 01 11:33:59 2007 +1000
@@ -15,14 +15,6 @@ struct lguest_dma
u32 used_len;
u32 addr[LGUEST_MAX_DMA_SECTIONS];
u16 len[LGUEST_MAX_DMA_SECTIONS];
-};
-
-/* This is found at address 0. */
-struct lguest_boot_info
-{
- u32 max_pfn;
- u32 initrd_size;
- char cmdline[256];
};
struct lguest_block_page
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/