All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Millan <rmh@aybabtu.com>
To: The development of GRUB 2 <grub-devel@gnu.org>
Cc: Mikhail Vorozhtsov <mikhail.vorozhtsov@gmail.com>
Subject: Re: Self multiboot patch
Date: Thu, 24 Jan 2008 14:37:35 +0100	[thread overview]
Message-ID: <20080124133735.GA7036@thorin> (raw)
In-Reply-To: <467527E0.1020008@gmail.com>


Anyone could have a look at this?

I can't verify that the code is right, but it doesn't modify or remove
existing code, so if it's unlikely that it will introduce regressions.

Btw, Mikhail, your code doesn't seem to do anything about "self-loading"
(i.e. preventing code from overwriting itself, causing havoc), but just
implements aout-kludge handling, is that right?

On Sun, Jun 17, 2007 at 07:24:00PM +0700, Mikhail Vorozhtsov wrote:
> Hi.
> 
> Here is a patch for booting GRUB via itself with multiboot module. 
> Tested on [http://grub.enbug.org/TestingOnX86]-alike floppy image and QEMU.
> 
> BTW. When I'm trying to add some code/data before/in multiboot_entry in 
> startup.S, GRUB hangs with "Loading kernel" message (even if it booted 
> not via multiboot). How can I avoid it? It would be nice to print some 
> message about invalid magic value, for example. But I just cannot add 
> the code.

> diff -ru grub2.orig/include/grub/i386/pc/loader.h grub2.multiboot/include/grub/i386/pc/loader.h
> --- grub2.orig/include/grub/i386/pc/loader.h	2004-09-12 19:20:52.000000000 +0700
> +++ grub2.multiboot/include/grub/i386/pc/loader.h	2007-06-17 18:31:53.000000000 +0700
> @@ -28,11 +28,17 @@
>  extern char *EXPORT_VAR(grub_linux_tmp_addr);
>  extern char *EXPORT_VAR(grub_linux_real_addr);
>  
> +/* Multiboot loader needs to know boot device. */
> +extern grub_uint32_t EXPORT_VAR(grub_boot_drive);
> +extern grub_int32_t EXPORT_VAR(grub_install_dos_part);
> +extern grub_int32_t EXPORT_VAR(grub_install_bsd_part);
> +
>  void EXPORT_FUNC(grub_linux_boot_zimage) (void) __attribute__ ((noreturn));
>  void EXPORT_FUNC(grub_linux_boot_bzimage) (void) __attribute__ ((noreturn));
>  
>  /* This is an asm part of the chainloader.  */
> -void EXPORT_FUNC(grub_chainloader_real_boot) (int drive, void *part_addr) __attribute__ ((noreturn));
> +void EXPORT_FUNC(grub_chainloader_real_boot) (int drive, void *part_addr)
> +     __attribute__ ((noreturn));
>  
>  /* The asm part of the multiboot loader.  */
>  void EXPORT_FUNC(grub_multiboot_real_boot) (grub_addr_t entry, 
> diff -ru grub2.orig/loader/i386/pc/multiboot.c grub2.multiboot/loader/i386/pc/multiboot.c
> --- grub2.orig/loader/i386/pc/multiboot.c	2006-06-04 22:56:54.000000000 +0700
> +++ grub2.multiboot/loader/i386/pc/multiboot.c	2007-06-17 19:02:10.000000000 +0700
> @@ -22,8 +22,6 @@
>   *  FIXME: The following features from the Multiboot specification still
>   *         need to be implemented:
>   *  - VBE support
> - *  - a.out support
> - *  - boot device
>   *  - symbol table
>   *  - memory map
>   *  - drives table
> @@ -236,6 +234,75 @@
>    return grub_error (GRUB_ERR_UNKNOWN_OS, "unknown ELF class");
>  }
>  
> +/* Load with a.out kludge. */
> +static grub_err_t
> +grub_multiboot_load_raw (grub_file_t file, grub_off_t header_offset,
> +			const struct grub_multiboot_header *header)
> +{
> +  grub_off_t file_size = grub_file_size (file);
> +  grub_off_t load_offset = header_offset
> +                           - (header->header_addr - header->load_addr);
> +  grub_uint32_t load_end_addr = header->load_end_addr;
> +  grub_uint32_t bss_end_addr = header->bss_end_addr;
> +  grub_uint32_t load_size, total_size;
> +
> +  if (header->header_addr < header->load_addr)
> +    return grub_error (GRUB_ERR_BAD_OS, "Header precedes code");
> +
> +  if (header_offset < (header->header_addr - header->load_addr)
> +      || load_offset > file_size)
> +    return grub_error (GRUB_ERR_BAD_OS, "Code and data go out of file");
> +
> +  if (load_end_addr == 0)
> +    {
> +      if (file_size - load_offset > 0xFFFFFFFF)
> +        return grub_error (GRUB_ERR_BAD_OS, "Code and data size is too big");
> +      load_size = file_size - load_offset;
> +      load_end_addr = header->load_addr + load_size;
> +    }
> +  else
> +    load_size = load_end_addr - header->load_addr;
> +
> +  if (load_end_addr < header->load_addr)
> +    return grub_error (GRUB_ERR_BAD_OS, "Code and data size is negative");
> +
> +  if (-(grub_uint64_t) load_size <= load_offset
> +      || load_offset + load_size > file_size)
> +    return grub_error (GRUB_ERR_BAD_OS, "Code and data go out of file");
> +
> +  if (header->entry_addr < header->load_addr
> +      || header->entry_addr >= load_end_addr)
> +    return grub_error (GRUB_ERR_BAD_OS,
> +                       "Entry point is outside of code and data area");
> +
> +  if (bss_end_addr == 0)
> +    bss_end_addr = load_end_addr;
> +
> +  if (bss_end_addr < load_end_addr)
> +    return grub_error (GRUB_ERR_BAD_OS, "BSS size is negative");
> +
> +  total_size = load_size + (bss_end_addr - load_end_addr);
> +
> +  if (-load_size <= (bss_end_addr - load_end_addr)
> +      || header->load_addr < grub_os_area_addr
> +      || -total_size <= header->load_addr
> +      || (header->load_addr + total_size)
> +          > (grub_os_area_addr + grub_os_area_size))
> +    return grub_error (GRUB_ERR_BAD_OS,
> +             "Kernel doesn't fit in memory reserved for the OS");
> +
> +  if (grub_file_seek (file, load_offset) == (grub_off_t) -1
> +      || grub_file_read (file, (char *) header->load_addr, load_size)
> +         != (grub_ssize_t) load_size)
> +    return grub_error (GRUB_ERR_READ_ERROR, "Cannot read code and data");
> +
> +  grub_memset ((void *) load_end_addr, 0, total_size - load_size);
> +
> +  entry = header->entry_addr;
> +
> +  return GRUB_ERR_NONE;
> +}
> +
>  void
>  grub_rescue_cmd_multiboot (int argc, char *argv[])
>  {
> @@ -293,7 +360,13 @@
>        goto fail;
>      }
>  
> -  if (grub_multiboot_load_elf (file, buffer) != GRUB_ERR_NONE)
> +  if (header->flags & GRUB_MB_AOUT_KLUDGE)
> +    {
> +      if (grub_multiboot_load_raw (file, (char *) header - buffer, header)
> +          != GRUB_ERR_NONE)
> +        goto fail;
> +    }
> +  else if (grub_multiboot_load_elf (file, buffer) != GRUB_ERR_NONE)
>      goto fail;
>    
>    mbi = grub_malloc (sizeof (struct grub_multiboot_info));
> @@ -325,6 +398,15 @@
>    mbi->flags |= GRUB_MB_INFO_CMDLINE;
>    mbi->cmdline = (grub_uint32_t) cmdline;
>  
> +  mbi->flags |= GRUB_MB_INFO_BOOTDEV;
> +  mbi->boot_device = (grub_boot_drive << 24);
> +  if (grub_install_dos_part >= 0)
> +    mbi->boot_device |= 0x00FFFF | (grub_install_dos_part << 16);
> +  else if (grub_install_bsd_part >= 0)
> +    mbi->boot_device |= 0xFF00FF | (grub_install_bsd_part << 8);
> +  else
> +    mbi->boot_device |= 0xFFFFFF;
> +
>    mbi->flags |= GRUB_MB_INFO_BOOT_LOADER_NAME;
>    mbi->boot_loader_name = (grub_uint32_t) grub_strdup (PACKAGE_STRING);
>  

> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/grub-devel


-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call… if you are unable to speak?
(as seen on /.)



      parent reply	other threads:[~2008-01-24 13:40 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-06-17 12:24 Self multiboot patch Mikhail Vorozhtsov
2007-07-22 13:11 ` Marco Gerards
2008-01-24 13:37 ` Robert Millan [this message]

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=20080124133735.GA7036@thorin \
    --to=rmh@aybabtu.com \
    --cc=grub-devel@gnu.org \
    --cc=mikhail.vorozhtsov@gmail.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.