Hello all,
I believe the x86 setup tree is now finished. I will turn it into a
"clean patchset" later this week, but I wanted to get flamed^W feedback
on it first.
The git tree is at:
http://git.kernel.org/?p=linux/kernel/git/hpa/linux-2.6-newsetup.git;a=summary
git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-newsetup.git
...
... and a flat patch at ...
http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
-hpa
On Tue, May 08, 2007 at 10:15:21PM -0700, H. Peter Anvin wrote:
> Hello all,
>
> I believe the x86 setup tree is now finished. I will turn it into a
> "clean patchset" later this week, but I wanted to get flamed^W feedback
> on it first.
>
> The git tree is at:
>
> http://git.kernel.org/?p=linux/kernel/git/hpa/linux-2.6-newsetup.git;a=summary
> git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-newsetup.git
> ...
>
> ... and a flat patch at ...
>
> http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
Are you planning to rebase to -linus and then let
arch/x86_64/boot/Makefile be a reference to i386/boot/Makefile?
The patch for kbuild to enable this is in -linus now.
As for the important part (the code) I have not looked.
Sam
Sam Ravnborg wrote:
>> http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
>
> Are you planning to rebase to -linus and then let
> arch/x86_64/boot/Makefile be a reference to i386/boot/Makefile?
>
> The patch for kbuild to enable this is in -linus now.
>
> As for the important part (the code) I have not looked.
I have rebased to -linus, but I'm a bit scared about making that
particular change at this time, mostly because it requires some
additional Makefile work which I would be happier to apply later, in
order to make the "create clean patchset" task easier.
All the source code is shared, but for right now the Makefiles are
slightly different, mostly w.r.t. CFLAGS.
-hpa
On Wed, May 09, 2007 at 12:02:45AM -0700, H. Peter Anvin wrote:
> Sam Ravnborg wrote:
> >> http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
> >
> > Are you planning to rebase to -linus and then let
> > arch/x86_64/boot/Makefile be a reference to i386/boot/Makefile?
> >
> > The patch for kbuild to enable this is in -linus now.
> >
> > As for the important part (the code) I have not looked.
>
> I have rebased to -linus, but I'm a bit scared about making that
> particular change at this time, mostly because it requires some
> additional Makefile work which I would be happier to apply later, in
> order to make the "create clean patchset" task easier.
>
> All the source code is shared, but for right now the Makefiles are
> slightly different, mostly w.r.t. CFLAGS.
OK, let me know if you want me to take a look at something specific.
Sam
On Tue, 08 May 2007 22:15:21 -0700, "H. Peter Anvin" <[email protected]>
said:
> I believe the x86 setup tree is now finished. I will turn it into a
> "clean patchset" later this week, but I wanted to get flamed^W feedback
> on it first.
Hi!
Only tiny sparks^Wremarks, I'm afraid. ;)
I've briefly looked at the new code in arch/i386/boot and as far
as I can determine in a short amount of time all functionality is
basically there, with the exception of the 'scan' functionality for
bios modes. Maybe it is not needed anymore? I have used this
functionality to find non-standard modes recently (for a kernel
crash during boot).
Random remarks about config-differences:
The new code does not use VIDEO_SELECT and X86_SPEEDSTEP_SMI
(i386-only): both parts are always compiled in. The new code uses FB:
the old code allowed setting a graphics mode, even without FB support.
Shared config options that affect the code generation: EDD, RELOCATABLE,
FIRMWARE_EDID, and APM (i386-only). Some code also depends on X86_ELAN
and X86_VOYAGER.
Allnoconfig realmode kernel part increases by about 3.5 kb:
old x86_64: Boot sector 512 bytes. / Setup is 4737 bytes.
new x86_64: Setup is 8808 bytes (padded to 9216 bytes).
old i386: Boot sector 512 bytes. / Setup is 4842 bytes.
new i386: Setup is 8776 bytes (padded to 9216 bytes).
When all relevant options are enabled about 1.5 kb:
old x86_64: Boot sector 512 bytes. / Setup is 7394 bytes.
new x86_64: Setup is 9672 bytes (padded to 9728 bytes).
old i386: Boot sector 512 bytes. / Setup is 7649 bytes.
new i386: Setup is 9896 bytes (padded to 10240 bytes).
So the code explosion is not too bad. Although... it is helps quite a
bit that the old setup contains a chunk of slightly more than 3 kb of
zeros which has disappeared with the new code. On the other hand, the
new code provides a valuable debugging tool in the form of a printf
function.
$ size arch/i386/boot/setup.elf
text data bss dec hex filename
9768 96 6056 15920 3e30 setup.elf
How much memory should we expect to be available? Is the assert for
32 kb for text+data+bss in setup.ld reasonable? Is it guaranteed
that the bootloader stores the commandline after _end?
The amount of code that depends on config options is relatively small.
Would it make sense to just compile everything independent of config
options and determine what should be done at run-time? And/or separate
the bootcode related config-options and put them under EMBEDDED/TINY?
If space is not that much of an issue, could you imagine putting
the decompression/relocation/misc.c code in setup too? This would
make it easier to start thinking of a new kernel image format without
breaking bootloaders that use a bzImage in the intended way.
Good luck,
Alexander
--
Alexander van Heukelum
[email protected]
--
http://www.fastmail.fm - A fast, anti-spam email service.
Alexander van Heukelum wrote:
> Hi!
>
> Only tiny sparks^Wremarks, I'm afraid. ;)
>
> I've briefly looked at the new code in arch/i386/boot and as far
> as I can determine in a short amount of time all functionality is
> basically there, with the exception of the 'scan' functionality for
> bios modes. Maybe it is not needed anymore? I have used this
> functionality to find non-standard modes recently (for a kernel
> crash during boot).
As far as I could tell, "scan" simply caused the nonstandard video
driver scan modules (unsafe probes) to be invoked. Since those modules
are no longer present, there appeared to be no need for them. The VGA
and VESA probes are safe.
> Random remarks about config-differences:
>
> The new code does not use VIDEO_SELECT and X86_SPEEDSTEP_SMI
> (i386-only): both parts are always compiled in. The new code uses FB:
> the old code allowed setting a graphics mode, even without FB support.
> Shared config options that affect the code generation: EDD, RELOCATABLE,
> FIRMWARE_EDID, and APM (i386-only). Some code also depends on X86_ELAN
> and X86_VOYAGER.
I was considering also making the MCA code dependent on CONFIG_MCA, but
realized it would be better to have it at least be compiled, to prevent
bitrot. It could be added back in to add a handful of bytes.
> Allnoconfig realmode kernel part increases by about 3.5 kb:
> old x86_64: Boot sector 512 bytes. / Setup is 4737 bytes.
> new x86_64: Setup is 8808 bytes (padded to 9216 bytes).
> old i386: Boot sector 512 bytes. / Setup is 4842 bytes.
> new i386: Setup is 8776 bytes (padded to 9216 bytes).
>
> When all relevant options are enabled about 1.5 kb:
> old x86_64: Boot sector 512 bytes. / Setup is 7394 bytes.
> new x86_64: Setup is 9672 bytes (padded to 9728 bytes).
> old i386: Boot sector 512 bytes. / Setup is 7649 bytes.
> new i386: Setup is 9896 bytes (padded to 10240 bytes).
>
> So the code explosion is not too bad. Although... it is helps quite a
> bit that the old setup contains a chunk of slightly more than 3 kb of
> zeros which has disappeared with the new code. On the other hand, the
> new code provides a valuable debugging tool in the form of a printf
> function.
>
> $ size arch/i386/boot/setup.elf
> text data bss dec hex filename
> 9768 96 6056 15920 3e30 setup.elf
>
> How much memory should we expect to be available? Is the assert for
> 32 kb for text+data+bss in setup.ld reasonable? Is it guaranteed
> that the bootloader stores the commandline after _end?
Yes, it is, at least by a bootloader that follows protocol. I don't
know of any bootloader that are out of spec in that regard.
> The amount of code that depends on config options is relatively small.
> Would it make sense to just compile everything independent of config
> options and determine what should be done at run-time? And/or separate
> the bootcode related config-options and put them under EMBEDDED/TINY?
It doesn't, because e.g. Voyager needs special A20 code, etc. I guess
we could dynamically detect Voyager and then do that, but that would
require testing.
> If space is not that much of an issue, could you imagine putting
> the decompression/relocation/misc.c code in setup too? This would
> make it easier to start thinking of a new kernel image format without
> breaking bootloaders that use a bzImage in the intended way.
I don't think that's either desirable or practical. That code really
wants ready addressability of the whole image.
-hpa
Hello!
> As far as I could tell, "scan" simply caused the nonstandard video
> driver scan modules (unsafe probes) to be invoked. Since those modules
> are no longer present, there appeared to be no need for them. The VGA
> and VESA probes are safe.
"scan" is still useful, because it is able to find BIOS video modes with
non-standard numbers (they are still sometimes found on recent cards).
Have a nice fortnight
--
Martin `MJ' Mares <[email protected]> http://mj.ucw.cz/
Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth
Entropy isn't what it used to be.
On Thu, May 10, 2007 at 11:08:10AM -0700, H. Peter Anvin wrote:
> As far as I could tell, "scan" simply caused the nonstandard video
> driver scan modules (unsafe probes) to be invoked. Since those modules
> are no longer present, there appeared to be no need for them. The VGA
> and VESA probes are safe.
It doesn't probe the hardware in dangerous ways. (Search for mode_scan
in video.S) It works by trying to set a mode via the normal
AH=0/AL=mode/int 0x10 method for all possible values of mode. It then
checks if the bios reports the new mode as being set and reads a few
standard vga registers to determine if it is a text mode. It's
completely independent of the CONFIG_VIDEO_SVGA stuff.
> > Would it make sense to just compile everything independent of config
> > options and determine what should be done at run-time? And/or separate
> > the bootcode related config-options and put them under EMBEDDED/TINY?
>
> It doesn't, because e.g. Voyager needs special A20 code, etc. I guess
> we could dynamically detect Voyager and then do that, but that would
> require testing.
Yeah, Voyager and Elan should be detected at runtime in that case. But
if one ever wants to do that, this would be a good time... This rewrite
needs thorough testing anyhow. ;)
> > If space is not that much of an issue, could you imagine putting
> > the decompression/relocation/misc.c code in setup too? This would
> > make it easier to start thinking of a new kernel image format without
> > breaking bootloaders that use a bzImage in the intended way.
>
> I don't think that's either desirable or practical. That code really
> wants ready addressability of the whole image.
Heh. I don't know about desirable, but it is not really impractical. The
Relocatable patches already introduced a small bit of 32-bit code in
setup (git-show 968de4f0 arch/i386/boot/setup.S). This has disappeared
in your rewrite, however.
That makes me wonder: (from arch/i386/boot/pmjump.S)
37 movw $__BOOT_DS, %cx
38
39 movl %cr0, %edx
40 orb $1, %dl # Protected mode (PE) bit
41 movl %edx, %cr0
42
43 movw %cx, %ds
44 movw %cx, %es
45 movw %cx, %fs
46 movw %cx, %gs
47 movw %cx, %ss
48
49 # Jump to the 32-bit entrypoint
50 .byte 0x66, 0xea # ljmpl opcode
51 2: .long 0 # offset
52 .word __BOOT_CS # segment
I thought the 32-bit jump was required to come before the segment loads.
Does this code load values from the gdt, or are they just loaded as real
mode segments? As long as it does not crash it does not matter, because
head.S reloads them again.
Thanks,
Greetings,
Alexander
Alexander van Heukelum wrote:
> On Thu, May 10, 2007 at 11:08:10AM -0700, H. Peter Anvin wrote:
>> As far as I could tell, "scan" simply caused the nonstandard video
>> driver scan modules (unsafe probes) to be invoked. Since those modules
>> are no longer present, there appeared to be no need for them. The VGA
>> and VESA probes are safe.
>
> It doesn't probe the hardware in dangerous ways. (Search for mode_scan
> in video.S) It works by trying to set a mode via the normal
> AH=0/AL=mode/int 0x10 method for all possible values of mode. It then
> checks if the bios reports the new mode as being set and reads a few
> standard vga registers to determine if it is a text mode. It's
> completely independent of the CONFIG_VIDEO_SVGA stuff.
It's dangerous, all right (which is why it doesn't do it by default),
since you have no guarantee that the BIOS doesn't totally vomit on these
calls -- or, like my laptop, take about a minute before giving up
finding nothing.
Anyway, I re-implemented scanning and pushed it out to the git tree;
please try it out as it does absolutely nothing on any of my machines.
> That makes me wonder: (from arch/i386/boot/pmjump.S)
>
> 37 movw $__BOOT_DS, %cx
> 38
> 39 movl %cr0, %edx
> 40 orb $1, %dl # Protected mode (PE) bit
> 41 movl %edx, %cr0
> 42
> 43 movw %cx, %ds
> 44 movw %cx, %es
> 45 movw %cx, %fs
> 46 movw %cx, %gs
> 47 movw %cx, %ss
> 48
> 49 # Jump to the 32-bit entrypoint
> 50 .byte 0x66, 0xea # ljmpl opcode
> 51 2: .long 0 # offset
> 52 .word __BOOT_CS # segment
>
> I thought the 32-bit jump was required to come before the segment loads.
> Does this code load values from the gdt, or are they just loaded as real
> mode segments? As long as it does not crash it does not matter, because
> head.S reloads them again.
Once CR0.PE is set, segments are loaded from the GDT.
-hpa
Martin Mares wrote:
> Hello!
>
>> As far as I could tell, "scan" simply caused the nonstandard video
>> driver scan modules (unsafe probes) to be invoked. Since those modules
>> are no longer present, there appeared to be no need for them. The VGA
>> and VESA probes are safe.
>
> "scan" is still useful, because it is able to find BIOS video modes with
> non-standard numbers (they are still sometimes found on recent cards).
Well, I don't have a card which does anything like that, but I did just
implement the "scan" functionality and pushed it out. If anyone cares
about that functionality it would be good if they could test it out and
report if it works.
-hpa
On Tue, May 08, 2007 at 10:15:21PM -0700, H. Peter Anvin wrote:
> Hello all,
>
> I believe the x86 setup tree is now finished. I will turn it into a
> "clean patchset" later this week, but I wanted to get flamed^W feedback
> on it first.
>
> The git tree is at:
>
> http://git.kernel.org/?p=linux/kernel/git/hpa/linux-2.6-newsetup.git;a=summary
> git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-newsetup.git
> ...
>
> ... and a flat patch at ...
>
> http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
>
Wow, reading code in C is so much better than decoding assembly. :-)
Had a quick look, mainly from relocatable kernel code point of view. Yet
to dive deeper.
PHYSICAL_ALIGN needs to be 2MB on x86_64 instead of 1MB.
Thanks
Vivek
H. Peter Anvin wrote:
> Hello all,
>
> I believe the x86 setup tree is now finished. I will turn it into a
> "clean patchset" later this week, but I wanted to get flamed^W feedback
> on it first.
>
> The git tree is at:
>
> http://git.kernel.org/?p=linux/kernel/git/hpa/linux-2.6-newsetup.git;a=summary
> git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-newsetup.git
> ...
>
> ... and a flat patch at ...
>
> http://www.kernel.org/pub/linux/kernel/people/hpa/newsetup-36f021b5.patch
>
> -hpa
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>
>
Not sure if you were looking for testing, but I fuzzed it to apply to
2.6.21-git and gave it a spin. Worked just like a normal boot (which I
assume was the point).
[ 0.000000] Linux version 2.6.21-g0a3fd051-dirty (kevin@kevin) (gcc
version 4.1.2 (Gentoo 4.1.2)) #9 PREEMPT Fri May 11 20:50:02 ADT 2007
[ 0.000000] Command line: root=/dev/sda3
[ 0.000000] BIOS-provided physical RAM map:
[ 0.000000] BIOS-e820: 0000000000000000 - 000000000009f800 (usable)
[ 0.000000] BIOS-e820: 000000000009f800 - 00000000000a0000 (reserved)
[ 0.000000] BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
[ 0.000000] BIOS-e820: 0000000000100000 - 000000001fef0000 (usable)
[ 0.000000] BIOS-e820: 000000001fef0000 - 000000001fef3000 (ACPI NVS)
[ 0.000000] BIOS-e820: 000000001fef3000 - 000000001ff00000 (ACPI data)
[ 0.000000] BIOS-e820: 00000000fec00000 - 0000000100000000 (reserved)
[ 0.000000] Entering add_active_range(0, 0, 159) 0 entries of 256 used
[ 0.000000] Entering add_active_range(0, 256, 130800) 1 entries of
256 used
[ 0.000000] end_pfn_map = 1048576
[ 0.000000] DMI 2.3 present.
[ 0.000000] ACPI: RSDP 000F77D0, 0014 (r0 VIAK8T)
[ 0.000000] ACPI: RSDT 1FEF3040, 0034 (r1 VIAK8T AWRDACPI 42302E31
AWRD 0)
[ 0.000000] ACPI: FACP 1FEF30C0, 0074 (r1 VIAK8T AWRDACPI 42302E31
AWRD 0)
[ 0.000000] ACPI: DSDT 1FEF3180, 4F8A (r1 VIAK8T AWRDACPI 1000
MSFT 100000E)
[ 0.000000] ACPI: FACS 1FEF0000, 0040
[ 0.000000] ACPI: BOOT 1FEF8180, 0028 (r1 VIAK8T AWRDACPI 42302E31
AWRD 0)
[ 0.000000] ACPI: SSDT 1FEF82C0, 00B5 (r1 PTLTD POWERNOW 1
LTP 1)
[ 0.000000] ACPI: APIC 1FEF8200, 0068 (r1 VIAK8T AWRDACPI 42302E31
AWRD 0)
[ 0.000000] Entering add_active_range(0, 0, 159) 0 entries of 256 used
[ 0.000000] Entering add_active_range(0, 256, 130800) 1 entries of
256 used
[ 0.000000] Zone PFN ranges:
[ 0.000000] DMA 0 -> 4096
[ 0.000000] DMA32 4096 -> 1048576
[ 0.000000] Normal 1048576 -> 1048576
[ 0.000000] early_node_map[2] active PFN ranges
[ 0.000000] 0: 0 -> 159
[ 0.000000] 0: 256 -> 130800
[ 0.000000] On node 0 totalpages: 130703
[ 0.000000] DMA zone: 56 pages used for memmap
[ 0.000000] DMA zone: 1356 pages reserved
[ 0.000000] DMA zone: 2587 pages, LIFO batch:0
[ 0.000000] DMA32 zone: 1732 pages used for memmap
[ 0.000000] DMA32 zone: 124972 pages, LIFO batch:31
[ 0.000000] Normal zone: 0 pages used for memmap
[ 0.000000] ACPI: PM-Timer IO Port: 0x4008
[ 0.000000] ACPI: Local APIC address 0xfee00000
[ 0.000000] ACPI: LAPIC (acpi_id[0x00] lapic_id[0x00] enabled)
[ 0.000000] Processor #0 (Bootup-CPU)
[ 0.000000] ACPI: LAPIC (acpi_id[0x01] lapic_id[0x01] disabled)
[ 0.000000] ACPI: LAPIC_NMI (acpi_id[0x00] high edge lint[0x1])
[ 0.000000] ACPI: LAPIC_NMI (acpi_id[0x01] high edge lint[0x1])
[ 0.000000] ACPI: IOAPIC (id[0x02] address[0xfec00000] gsi_base[0])
[ 0.000000] IOAPIC[0]: apic_id 2, address 0xfec00000, GSI 0-23
[ 0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl dfl)
[ 0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
[ 0.000000] ACPI: IRQ0 used by override.
[ 0.000000] ACPI: IRQ2 used by override.
[ 0.000000] ACPI: IRQ9 used by override.
[ 0.000000] Setting APIC routing to flat
[ 0.000000] Using ACPI (MADT) for SMP configuration information
[ 0.000000] Allocating PCI resources starting at 20000000 (gap:
1ff00000:ded00000)
[ 0.000000] Built 1 zonelists. Total pages: 127559
[ 0.000000] Kernel command line: root=/dev/sda3
[ 0.000000] Initializing CPU#0
[ 0.000000] PID hash table entries: 2048 (order: 11, 16384 bytes)
[ 13.327158] time.c: Detected 1838.853 MHz processor.
[ 13.328328] Console: colour VGA+ 80x25
[ 13.331709] Dentry cache hash table entries: 65536 (order: 7, 524288
bytes)
[ 13.332048] Inode-cache hash table entries: 32768 (order: 6, 262144
bytes)
[ 13.332193] Checking aperture...
[ 13.332264] CPU 0: aperture @ e0000000 size 128 MB
[ 13.337619] Memory: 509000k/523200k available (3246k kernel code,
13476k reserved, 1225k data, 184k init)
[ 13.337752] SLUB: Genslabs=18, HWalign=64, Order=0-1, MinObjects=4,
Processors=1, Nodes=1
[ 13.397023] Calibrating delay using timer specific routine.. 3680.25
BogoMIPS (lpj=1840129)
[ 13.397189] Mount-cache hash table entries: 256
[ 13.397348] CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64
bytes/line)
[ 13.397424] CPU: L2 Cache: 512K (64 bytes/line)
[ 13.397505] CPU: AMD Athlon(tm) 64 Processor 3000+ stepping 00
[ 13.397659] ACPI: Core revision 20070126
[ 13.411535] Using local APIC timer interrupts.
[ 13.465907] result 12769821
[ 13.465976] Detected 12.769 MHz APIC timer.
[ 13.466993] NET: Registered protocol family 16
[ 13.467130] ACPI: bus type pci registered
[ 13.467207] PCI: Using configuration type 1
[ 13.472008] ACPI: Interpreter enabled
[ 13.472083] ACPI: (supports S0 S3 S4 S5)
[ 13.472356] ACPI: Using IOAPIC for interrupt routing
[ 13.478091] ACPI: PCI Root Bridge [PCI0] (0000:00)
[ 13.478167] PCI: Probing PCI hardware (bus 00)
[ 13.478958] ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
[ 13.530494] ACPI: PCI Interrupt Link [LNKA] (IRQs 3 4 6 7 10 11 12) *5
[ 13.531113] ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 6 7 10 *11 12)
[ 13.531689] ACPI: PCI Interrupt Link [LNKC] (IRQs 3 *4 6 7 10 11 12)
[ 13.532251] ACPI: PCI Interrupt Link [LNKD] (IRQs 3 4 6 7 10 11 12)
*0, disabled.
[ 13.532909] ACPI: PCI Interrupt Link [LNKE] (IRQs 3 4 6 7 10 11 12)
*0, disabled.
[ 13.533558] ACPI: PCI Interrupt Link [LNKF] (IRQs 3 4 6 7 10 11 12)
*0, disabled.
[ 13.534222] ACPI: PCI Interrupt Link [LNK0] (IRQs 3 4 6 7 *10 11 12)
[ 13.534768] ACPI: PCI Interrupt Link [LNK1] (IRQs 3 4 6 7 10 11 12)
*0, disabled.
[ 13.535454] ACPI: PCI Interrupt Link [ALKA] (IRQs *20)
[ 13.535798] ACPI: PCI Interrupt Link [ALKB] (IRQs *21)
[ 13.536141] ACPI: PCI Interrupt Link [ALKC] (IRQs *22)
[ 13.536512] ACPI: PCI Interrupt Link [ALKD] (IRQs *23), disabled.
[ 13.536801] Linux Plug and Play Support v0.97 (c) Adam Belay
[ 13.536883] pnp: PnP ACPI init
[ 13.536957] ACPI: bus type pnp registered
[ 13.539749] pnp: PnP ACPI: found 8 devices
[ 13.539822] ACPI: ACPI bus type pnp unregistered
[ 13.539959] SCSI subsystem initialized
[ 13.540050] libata version 2.20 loaded.
[ 13.540090] usbcore: registered new interface driver usbfs
[ 13.540184] usbcore: registered new interface driver hub
[ 13.540283] usbcore: registered new device driver usb
[ 13.540393] PCI: Using ACPI for IRQ routing
[ 13.540465] PCI: If a device doesn't work, try "pci=routeirq". If it
helps, post a report
[ 13.540574] PCI: Cannot allocate resource region 0 of device 0000:00:00.0
[ 13.540749] agpgart: Detected AGP bridge 0
[ 13.544290] agpgart: AGP aperture is 128M @ 0xe0000000
[ 13.544420] pnp: 00:00: iomem range 0xcc000-0xcffff has been reserved
[ 13.544495] pnp: 00:00: iomem range 0xf0000-0xf7fff could not be reserved
[ 13.544570] pnp: 00:00: iomem range 0xf8000-0xfbfff could not be reserved
[ 13.544645] pnp: 00:00: iomem range 0xfc000-0xfffff could not be reserved
[ 13.544723] pnp: 00:02: ioport range 0x4000-0x407f has been reserved
[ 13.544798] Time: tsc clocksource has been installed.
[ 13.544873] pnp: 00:02: ioport range 0x5000-0x500f has been reserved
[ 13.545178] PCI: Bridge: 0000:00:01.0
[ 13.545250] IO window: c000-cfff
[ 13.545321] MEM window: f8000000-f80fffff
[ 13.545394] PREFETCH window: e8000000-f7ffffff
[ 13.545476] PCI: Setting latency timer of device 0000:00:01.0 to 64
[ 13.545485] NET: Registered protocol family 2
[ 13.553759] IP route cache hash table entries: 4096 (order: 3, 32768
bytes)
[ 13.553861] TCP established hash table entries: 16384 (order: 6,
262144 bytes)
[ 13.554076] TCP bind hash table entries: 16384 (order: 5, 131072 bytes)
[ 13.554224] TCP: Hash tables configured (established 16384 bind 16384)
[ 13.554299] TCP reno registered
[ 13.556858] Simple Boot Flag at 0x37 set to 0x80
[ 13.559287] NTFS driver 2.1.28 [Flags: R/O].
[ 13.559504] io scheduler noop registered
[ 13.559657] io scheduler cfq registered (default)
[ 13.559834] Boot video device is 0000:01:00.0
[ 13.559918] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level,
low) -> IRQ 16
[ 13.560110] radeonfb: Found Intel x86 BIOS ROM Image
[ 13.560184] radeonfb: Retrieved PLL infos from BIOS
[ 13.560257] radeonfb: Reference=27.00 MHz (RefDiv=12) Memory=200.00
Mhz, System=166.00 MHz
[ 13.560359] radeonfb: PLL min 20000 max 40000
[ 13.636315] i2c-adapter i2c-1: unable to read EDID block.
[ 13.755068] i2c-adapter i2c-1: unable to read EDID block.
[ 13.873859] i2c-adapter i2c-1: unable to read EDID block.
[ 13.992712] i2c-adapter i2c-3: unable to read EDID block.
[ 14.111487] i2c-adapter i2c-3: unable to read EDID block.
[ 14.230259] i2c-adapter i2c-3: unable to read EDID block.
[ 14.483117] radeonfb: Monitor 1 type CRT found
[ 14.483187] radeonfb: EDID probed
[ 14.483257] radeonfb: Monitor 2 type no found
[ 14.483556] radeonfb (0000:01:00.0): ATI Radeon Yd
[ 14.483757] input: Power Button (FF) as /class/input/input0
[ 14.483830] ACPI: Power Button (FF) [PWRF]
[ 14.483948] input: Power Button (CM) as /class/input/input1
[ 14.484021] ACPI: Power Button (CM) [PWRB]
[ 14.485986] Real Time Clock Driver v1.12ac
[ 14.486085] Software Watchdog Timer: 0.07 initialized. soft_noboot=0
soft_margin=60 sec (nowayout= 0)
[ 14.486208] Linux agpgart interface v0.102 (c) Dave Jones
[ 14.486323] [drm] Initialized drm 1.1.0 20060810
[ 14.486505] [drm] Initialized radeon 1.27.0 20060524 on minor 0
[ 14.486589] Hangcheck: starting hangcheck timer 0.9.0 (tick is 180
seconds, margin is 60 seconds).
[ 14.486693] Hangcheck: Using get_cycles().
[ 14.486764] Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ
sharing disabled
[ 14.487136] loop: module loaded
[ 14.487240] via-rhine: Broken BIOS detected, avoid_D3 enabled.
[ 14.487363] VIA Networking Velocity Family Gigabit Ethernet Adapter
Driver Ver. 1.14
[ 14.487465] Copyright (c) 2002, 2003 VIA Networking Technologies, Inc.
[ 14.487539] Copyright (c) 2004 Red Hat Inc.
[ 14.487624] ACPI: PCI Interrupt 0000:00:0e.0[A] -> GSI 22 (level,
low) -> IRQ 22
[ 14.488499] eth0: VIA Networking Velocity Family Gigabit Ethernet Adapter
[ 14.488575] eth0: Ethernet Address: 00:50:8D:D1:28:BB
[ 14.499140] PPP generic driver version 2.4.2
[ 14.499262] PPP MPPE Compression module registered
[ 14.499421] pata_via 0000:00:0f.1: version 0.3.1
[ 14.499665] ACPI: PCI Interrupt Link [ALKA] enabled at IRQ 20
[ 14.499739] ACPI: PCI Interrupt 0000:00:0f.1[A] -> Link [ALKA] -> GSI
20 (level, low) -> IRQ 20
[ 14.500104] scsi0 : pata_via
[ 14.500215] scsi1 : pata_via
[ 14.500313] ata1: PATA max UDMA/133 cmd 0x00000000000101f0 ctl
0x00000000000103f6 bmdma 0x000000000001e300 irq 14
[ 14.500421] ata2: PATA max UDMA/133 cmd 0x0000000000010170 ctl
0x0000000000010376 bmdma 0x000000000001e308 irq 15
[ 14.955483] ata1.00: ATAPI, max UDMA/33
[ 14.955556] ata1.01: ATAPI, max UDMA/33
[ 15.112201] ata1.00: configured for UDMA/33
[ 15.268927] ata1.01: configured for UDMA/33
[ 15.457703] ata2.00: ata_hpa_resize 1: sectors = 156301488,
hpa_sectors = 156301488
[ 15.457805] ata2.00: ATA-6: WDC WD800JB-00JJA0, 05.01C05, max UDMA/100
[ 15.457879] ata2.00: 156301488 sectors, multi 16: LBA
[ 15.460688] ata2.00: ata_hpa_resize 1: sectors = 156301488,
hpa_sectors = 156301488
[ 15.461282] ata2.00: configured for UDMA/100
[ 15.461960] scsi 0:0:0:0: CD-ROM COMPAQ DVD-ROM DRD8120B
1.10 PQ: 0 ANSI: 5
[ 15.465328] sr0: scsi3-mmc drive: 0x/40x cd/rw xa/form2 cdda tray
[ 15.465404] Uniform CD-ROM driver Revision: 3.20
[ 15.465519] sr 0:0:0:0: Attached scsi CD-ROM sr0
[ 15.465561] sr 0:0:0:0: Attached scsi generic sg0 type 5
[ 15.465858] scsi 0:0:1:0: CD-ROM HL-DT-ST CD-RW GCE-8525B
1.03 PQ: 0 ANSI: 5
[ 15.469540] sr1: scsi3-mmc drive: 40x/52x writer cd/rw xa/form2 cdda tray
[ 15.469653] sr 0:0:1:0: Attached scsi CD-ROM sr1
[ 15.469691] sr 0:0:1:0: Attached scsi generic sg1 type 5
[ 15.469816] scsi 1:0:0:0: Direct-Access ATA WDC WD800JB-00JJ
05.0 PQ: 0 ANSI: 5
[ 15.469985] sd 1:0:0:0: [sda] 156301488 512-byte hardware sectors
(80026 MB)
[ 15.470067] sd 1:0:0:0: [sda] Write Protect is off
[ 15.470139] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 15.470152] sd 1:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 15.470286] sd 1:0:0:0: [sda] 156301488 512-byte hardware sectors
(80026 MB)
[ 15.470366] sd 1:0:0:0: [sda] Write Protect is off
[ 15.470440] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 15.470452] sd 1:0:0:0: [sda] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 15.470557] sda: sda1 sda2 sda3
[ 15.474730] sd 1:0:0:0: [sda] Attached SCSI disk
[ 15.474839] sd 1:0:0:0: Attached scsi generic sg2 type 0
[ 15.474970] ACPI: PCI Interrupt 0000:00:10.4[C] -> GSI 21 (level,
low) -> IRQ 21
[ 15.475125] ehci_hcd 0000:00:10.4: EHCI Host Controller
[ 15.475242] ehci_hcd 0000:00:10.4: new USB bus registered, assigned
bus number 1
[ 15.475386] ehci_hcd 0000:00:10.4: irq 21, io mem 0xf8120000
[ 15.475462] ehci_hcd 0000:00:10.4: USB 2.0 started, EHCI 1.00, driver
10 Dec 2004
[ 15.475680] usb usb1: configuration #1 chosen from 1 choice
[ 15.475784] hub 1-0:1.0: USB hub found
[ 15.475860] hub 1-0:1.0: 8 ports detected
[ 15.576246] USB Universal Host Controller Interface driver v3.0
[ 15.576413] ACPI: PCI Interrupt 0000:00:10.0[A] -> GSI 21 (level,
low) -> IRQ 21
[ 15.576559] uhci_hcd 0000:00:10.0: UHCI Host Controller
[ 15.576654] uhci_hcd 0000:00:10.0: new USB bus registered, assigned
bus number 2
[ 15.576772] uhci_hcd 0000:00:10.0: irq 21, io base 0x0000e400
[ 15.576946] usb usb2: configuration #1 chosen from 1 choice
[ 15.577047] hub 2-0:1.0: USB hub found
[ 15.577122] hub 2-0:1.0: 2 ports detected
[ 15.678072] ACPI: PCI Interrupt 0000:00:10.1[A] -> GSI 21 (level,
low) -> IRQ 21
[ 15.678215] uhci_hcd 0000:00:10.1: UHCI Host Controller
[ 15.678309] uhci_hcd 0000:00:10.1: new USB bus registered, assigned
bus number 3
[ 15.678427] uhci_hcd 0000:00:10.1: irq 21, io base 0x0000e500
[ 15.678598] usb usb3: configuration #1 chosen from 1 choice
[ 15.678697] hub 3-0:1.0: USB hub found
[ 15.678771] hub 3-0:1.0: 2 ports detected
[ 15.778891] ACPI: PCI Interrupt 0000:00:10.2[B] -> GSI 21 (level,
low) -> IRQ 21
[ 15.779034] uhci_hcd 0000:00:10.2: UHCI Host Controller
[ 15.779134] uhci_hcd 0000:00:10.2: new USB bus registered, assigned
bus number 4
[ 15.779250] uhci_hcd 0000:00:10.2: irq 21, io base 0x0000e600
[ 15.779420] usb usb4: configuration #1 chosen from 1 choice
[ 15.779525] hub 4-0:1.0: USB hub found
[ 15.779598] hub 4-0:1.0: 2 ports detected
[ 15.879716] ACPI: PCI Interrupt 0000:00:10.3[B] -> GSI 21 (level,
low) -> IRQ 21
[ 15.879858] uhci_hcd 0000:00:10.3: UHCI Host Controller
[ 15.879955] uhci_hcd 0000:00:10.3: new USB bus registered, assigned
bus number 5
[ 15.880072] uhci_hcd 0000:00:10.3: irq 21, io base 0x0000e700
[ 15.880244] usb usb5: configuration #1 chosen from 1 choice
[ 15.880347] hub 5-0:1.0: USB hub found
[ 15.880421] hub 5-0:1.0: 2 ports detected
[ 15.980534] Initializing USB Mass Storage driver...
[ 16.091298] usb 2-1: new low speed USB device using uhci_hcd and
address 2
[ 16.261733] usb 2-1: configuration #1 chosen from 1 choice
[ 16.469635] usb 2-2: new low speed USB device using uhci_hcd and
address 3
[ 16.630081] usb 2-2: configuration #1 chosen from 1 choice
[ 17.255270] usbcore: registered new interface driver usb-storage
[ 17.255343] USB Mass Storage support registered.
[ 17.255492] PNP: No PS/2 controller found. Probing ports directly.
[ 17.257625] serio: i8042 KBD port at 0x60,0x64 irq 1
[ 17.257719] serio: i8042 AUX port at 0x60,0x64 irq 12
[ 17.257838] mice: PS/2 mouse device common for all mice
[ 17.257999] i2c /dev entries driver
[ 17.258332] w83627hf: Found W83627HF chip at 0x290
[ 17.587333] w83627hf w83627hf.656: Enabling temp2, readings might not
make sense
[ 17.587466] w83627hf w83627hf.656: Enabling temp3, readings might not
make sense
[ 17.608879] abituguru: testing bank1 sensor 0
[ 17.628992] abituguru: alarm not raised during volt sensor test
[ 17.679475] abituguru: found temp sensor
[ 17.680095] abituguru: testing bank1 sensor 1
[ 17.700889] abituguru: alarm not raised during volt sensor test
[ 17.751355] abituguru: found temp sensor
[ 17.751975] abituguru: testing bank1 sensor 3
[ 17.773298] abituguru: found volt sensor
[ 17.773920] abituguru: testing bank1 sensor 4
[ 17.795252] abituguru: found volt sensor
[ 17.795871] abituguru: testing bank1 sensor 10
[ 17.817138] abituguru: found volt sensor
[ 17.817755] abituguru: testing bank1 sensor 8
[ 17.838111] abituguru: found volt sensor
[ 17.838726] abituguru: testing bank1 sensor 14
[ 17.859142] abituguru: found volt sensor
[ 17.859764] abituguru: testing bank1 sensor 2
[ 17.881061] abituguru: found volt sensor
[ 17.881675] abituguru: testing bank1 sensor 9
[ 17.901996] abituguru: found volt sensor
[ 17.902614] abituguru: testing bank1 sensor 6
[ 17.923003] abituguru: found volt sensor
[ 17.923618] abituguru: testing bank1 sensor 5
[ 17.944962] abituguru: found volt sensor
[ 17.945579] abituguru: testing bank1 sensor 11
[ 17.965922] abituguru: found volt sensor
[ 17.966540] abituguru: testing bank1 sensor 15
[ 17.986351] abituguru: alarm not raised during volt sensor test
[ 18.036859] abituguru: found temp sensor
[ 18.037481] abituguru: testing bank1 sensor 13
[ 18.058783] abituguru: found volt sensor
[ 18.059400] abituguru: testing bank1 sensor 7
[ 18.080194] abituguru: alarm not raised during volt sensor test
[ 18.130127] abituguru: alarm not raised during temp sensor test
[ 18.130849] abituguru: testing bank1 sensor 12
[ 18.151105] abituguru: alarm not raised during volt sensor test
[ 18.200985] abituguru: alarm not raised during temp sensor test
[ 18.201430] abituguru: detecting number of fan sensors
[ 18.201432] abituguru: bank2 sensor 5 does not seem to be a fan
sensor: the threshold (0) is below the minimum (5)
[ 18.201435] abituguru: found: 5 fan sensors
[ 18.201443] abituguru: detecting number of PWM outputs
[ 18.201445] abituguru: pwm channel 3 does not seem to be a pwm
channel: settings[0] = C9
[ 18.201448] abituguru: found: 3 PWM outputs
[ 18.201453] abituguru: found Abit uGuru
[ 18.215405] input: Microsoft Microsoft® Digital Media Pro Keyboard as
/class/input/input2
[ 18.215534] input: USB HID v1.11 Keyboard [Microsoft Microsoft®
Digital Media Pro Keyboard] on usb-0000:00:10.0-1
[ 18.244293] input: Microsoft Microsoft® Digital Media Pro Keyboard as
/class/input/input3
[ 18.244420] input: USB HID v1.11 Device [Microsoft Microsoft® Digital
Media Pro Keyboard] on usb-0000:00:10.0-1
[ 18.259272] input: Logitech USB Mouse as /class/input/input4
[ 18.259385] input: USB HID v1.10 Mouse [Logitech USB Mouse] on
usb-0000:00:10.0-2
[ 18.259568] usbcore: registered new interface driver usbhid
[ 18.259641] drivers/hid/usbhid/hid-core.c: v2.6:USB HID core driver
[ 18.259732] Advanced Linux Sound Architecture Driver Version
1.0.14rc4 (Wed May 09 09:51:39 2007 UTC).
[ 18.260011] ACPI: PCI Interrupt 0000:00:11.5[C] -> GSI 22 (level,
low) -> IRQ 22
[ 18.260281] PCI: Setting latency timer of device 0000:00:11.5 to 64
[ 18.773155] ALSA device list:
[ 18.773227] #0: VIA 8237 with ALC658D at 0xd800, irq 22
[ 18.773324] oprofile: using NMI interrupt.
[ 18.773400] GRE over IPv4 tunneling driver
[ 18.773539] TCP cubic registered
[ 18.773647] NET: Registered protocol family 1
[ 18.773719] NET: Registered protocol family 17
[ 18.773795] powernow-k8: Found 1 AMD Athlon(tm) 64 Processor 3000+
processors (version 2.00.00)
[ 18.773927] powernow-k8: 0 : fid 0xa (1800 MHz), vid 0x6
[ 18.774000] powernow-k8: 1 : fid 0x2 (1000 MHz), vid 0x12
[ 18.810591] kjournald starting. Commit interval 5 seconds
[ 18.810671] EXT3-fs: mounted filesystem with ordered data mode.
[ 18.810748] VFS: Mounted root (ext3 filesystem) readonly.
[ 18.810833] Freeing unused kernel memory: 184k freed
[ 21.503232] EXT3 FS on sda3, internal journal
[ 21.807016] Adding 1004052k swap on /dev/sda2. Priority:-1 extents:1
across:1004052k
[ 32.207201] agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
[ 32.207217] agpgart: Putting AGP V3 device at 0000:00:00.0 into 4x mode
[ 32.207267] agpgart: Putting AGP V3 device at 0000:01:00.0 into 4x mode
[ 32.371303] Velocity is AUTO mode
[ 32.499541] [drm] Setting GART location based on new memory map
[ 32.499617] [drm] Loading R200 Microcode
[ 32.499682] [drm] writeback test succeeded in 1 usecs
[ 33.995995] eth0: Link autonegation speed 100M bps full duplex
Kevin Winchester wrote:
> Not sure if you were looking for testing, but I fuzzed it to apply to
> 2.6.21-git and gave it a spin. Worked just like a normal boot (which I
> assume was the point).
That would be the point, yes :) Looking for breakage in video mode
detection, memory detection, and APM are probably the trickiest areas.
-hpa
On Thu, May 10, 2007 at 03:48:08PM -0700, H. Peter Anvin wrote:
> > It doesn't probe the hardware in dangerous ways. (Search for mode_scan
> > in video.S) It works by trying to set a mode via the normal
> > AH=0/AL=mode/int 0x10 method for all possible values of mode. It then
> > checks if the bios reports the new mode as being set and reads a few
> > standard vga registers to determine if it is a text mode. It's
> > completely independent of the CONFIG_VIDEO_SVGA stuff.
>
> It's dangerous, all right (which is why it doesn't do it by default),
> since you have no guarantee that the BIOS doesn't totally vomit on these
> calls -- or, like my laptop, take about a minute before giving up
> finding nothing.
I see. I think I just have too much trust in the biosses. Anyhow, the
'scan' option has always taken quite a bit of time, usually between 30
seconds and a minute.
> Anyway, I re-implemented scanning and pushed it out to the git tree;
> please try it out as it does absolutely nothing on any of my machines.
I can confirm that it works for at least one computer over here (a six
months old x86_64 machine with ATI ES1000-based on-board graphics). Some
non-vesa modes including a nice 100x30 one with 8x16 font are found by
the 'scan' option. No 100x60, however, but that is not a regression.
> > I thought the 32-bit jump was required to come before the segment loads.
> > Does this code load values from the gdt, or are they just loaded as real
> > mode segments? As long as it does not crash it does not matter, because
> > head.S reloads them again.
>
> Once CR0.PE is set, segments are loaded from the GDT.
I believe you :).
Greetings,
Alexander
Alexander van Heukelum <[email protected]> writes:
> On Thu, May 10, 2007 at 03:48:08PM -0700, H. Peter Anvin wrote:
> I can confirm that it works for at least one computer over here (a six
> months old x86_64 machine with ATI ES1000-based on-board graphics). Some
> non-vesa modes including a nice 100x30 one with 8x16 font are found by
> the 'scan' option. No 100x60, however, but that is not a regression.
>
>> > I thought the 32-bit jump was required to come before the segment loads.
>> > Does this code load values from the gdt, or are they just loaded as real
>> > mode segments? As long as it does not crash it does not matter, because
>> > head.S reloads them again.
>>
>> Once CR0.PE is set, segments are loaded from the GDT.
>
> I believe you :).
HPA is both right and wrong on this. The safe sequence for entering
protected mode requires a jump immediately after setting PE in %cr0.
To serialize the instruction stream and to be on an execution that
is tested and guaranteed to work in cpus.
On a lot of processors you can get away with more then that, but
Intel at least explicitly states in their "Software Developers Manual
Volume 3: System Programming" in 9.9.1 "Switching to Protected Mode"
that you need the jump immediately following enabling PE in cr0.
So while I was debugging and instrumenting code I would happily
place an instruction before the jump so I could get output.
However in production grade code I would not place anything
between the setting of PE in cr0 and the first jump instruction.
Eric
Eric W. Biederman wrote:
>
> HPA is both right and wrong on this. The safe sequence for entering
> protected mode requires a jump immediately after setting PE in %cr0.
> To serialize the instruction stream and to be on an execution that
> is tested and guaranteed to work in cpus.
>
Eric, that's complete nonsense. What Intel documents and what Intel
tests are two very different things.
Intel appears to be afraid of what they call "the crack", but the
semantics of the transition are quite well understood.
> On a lot of processors you can get away with more then that, but
> Intel at least explicitly states in their "Software Developers Manual
> Volume 3: System Programming" in 9.9.1 "Switching to Protected Mode"
> that you need the jump immediately following enabling PE in cr0.
>
> So while I was debugging and instrumenting code I would happily
> place an instruction before the jump so I could get output.
> However in production grade code I would not place anything
> between the setting of PE in cr0 and the first jump instruction.
A lot of the Intel manuals are boilerplate written by technical writers.
Once you're used to them you can spot it quite easily because it's
carried from generation to generation with little change, even when it's
blatantly obsoleted.
I can tell you there is tons of production code out there which *never*
reloads CS after switching into and out of protected mode. It just
plain has to work, on every x86 CPU ever made, with the obvious
semantics. With Intel's VT, you want to avoid sticking in that mode for
extended periods of time, for performance reasons, but it will work even
there (it kicks to interpreter, though.) Staying long enough to load
the segment registers is not an issue.
-hpa
"H. Peter Anvin" <[email protected]> writes:
> Eric W. Biederman wrote:
>>
>> HPA is both right and wrong on this. The safe sequence for entering
>> protected mode requires a jump immediately after setting PE in %cr0.
>> To serialize the instruction stream and to be on an execution that
>> is tested and guaranteed to work in cpus.
>>
>
> Eric, that's complete nonsense. What Intel documents and what Intel
> tests are two very different things.
Likely.
> Intel appears to be afraid of what they call "the crack", but the
> semantics of the transition are quite well understood.
Regardless not putting a jump in there violates the principle of
be conservative in what you send.
> A lot of the Intel manuals are boilerplate written by technical writers.
> Once you're used to them you can spot it quite easily because it's
> carried from generation to generation with little change, even when it's
> blatantly obsoleted.
Even on 386 and 486 class cpus?
To some extent if the rules don't change it makes sense for them to
copy the information from one generation to the next of the architecture.
Even if the current cpus don't really care.
I guess I just don't see the sense in taking chances if we don't have
to, and I don't see any real advantage of doing a data segment reload
before the jump.
Eric
Eric W. Biederman wrote:
>
> Even on 386 and 486 class cpus?
>
Yes, even on 386 and 486 class CPUs. I have personally tested this on
machines as old as the original "double sigma" 386-16.
> To some extent if the rules don't change it makes sense for them to
> copy the information from one generation to the next of the architecture.
> Even if the current cpus don't really care.
>
> I guess I just don't see the sense in taking chances if we don't have
> to, and I don't see any real advantage of doing a data segment reload
> before the jump.
It makes the code cleaner -- more debuggable -- by introducing clean
separation between 16- and 32-bit code.
-hpa
"H. Peter Anvin" <[email protected]> writes:
> Eric W. Biederman wrote:
>>
>> Even on 386 and 486 class cpus?
>>
>
> Yes, even on 386 and 486 class CPUs. I have personally tested this on
> machines as old as the original "double sigma" 386-16.
Ok. If you have tested on a wide variety of machines then I won't
worry about it.
I guess if a cr0 write has always been synchronizing things should be
a safe practice. The practical danger is if you write to cr0 and the
pipeline is not flushed and the segment loads might execute as 16bit
mode segment register loads. I think this is more of a 486 or 586/Pentium
danger actually then a 386 one.
Hmm. I'm not certain about enabling protected mode but enabling paging
at least does appear to be documented to require a jump, before the P6
core. Which may be why it is recommended for initializing protected
mode.
Eric
Eric W. Biederman wrote:
> Ok. If you have tested on a wide variety of machines then I won't
> worry about it.
>
> I guess if a cr0 write has always been synchronizing things should be
> a safe practice. The practical danger is if you write to cr0 and the
> pipeline is not flushed and the segment loads might execute as 16bit
> mode segment register loads. I think this is more of a 486 or 586/Pentium
> danger actually then a 386 one.
Writes to CR0 are synchronizing and are documented as such. This is a
fairly trivial consequence of it invoking microcode on every x86 ever
created.
> Hmm. I'm not certain about enabling protected mode but enabling paging
> at least does appear to be documented to require a jump, before the P6
> core. Which may be why it is recommended for initializing protected
> mode.
Enabling paging is possibly a different matter. I can't speak on that
particular subject with any authority.
-hpa
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
H. Peter Anvin wrote:
> Kevin Winchester wrote:
>> Not sure if you were looking for testing, but I fuzzed it to apply to
>> 2.6.21-git and gave it a spin. Worked just like a normal boot (which I
>> assume was the point).
>
> That would be the point, yes :) Looking for breakage in video mode
> detection, memory detection, and APM are probably the trickiest areas.
>
> -hpa
>
Video mode detection (including "scan") seemed to work just the same as before.
As for memory detection, my 512M x86-64 UP system is probably not what you had
in mind, and I don't really know what goes on for APM or how to test it.
As well, I've been looking through the actual code, and even with my very
limited experience with x86 assembler and operating systems, I now at least have
an idea of what is going on. So switching to C for as much of the code as
possible seems like the right idea if you want to make it easier for new people
to get involved in that area of the kernel.
Kevin
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFGRl1CKPGFQbiQ3tQRAuEcAJ93zlfNzB7YipLHgvkQP3Jrtjh0EwCfSRQ5
9emaZMVvJsuXMs1ifO4KWkY=
=G4zx
-----END PGP SIGNATURE-----
On Friday 11 May 2007 8:23 pm, H. Peter Anvin wrote:
> Kevin Winchester wrote:
> > Not sure if you were looking for testing, but I fuzzed it to apply to
> > 2.6.21-git and gave it a spin. Worked just like a normal boot (which I
> > assume was the point).
>
> That would be the point, yes :) Looking for breakage in video mode
> detection, memory detection, and APM are probably the trickiest areas.
This may not be the most useful datapoint, but applying the patch to
current -git didn't give me something that qemu could boot.
I applied the patch to current -git (well, current -hg but same thing). The
flat patch didn't quite apply cleanly. (Hunks in Documentation, video.S, and
verify_cpu.S claimed to be already applied. I told it to ignore them.) The
result compiled cleanly enough, but didn't boot under qemu. (It
said "booting from hard drive" and then hung. No decompressing message.)
I've confirmed that without the patch, current -hg does boot under qemu.
Awaiting the broken-out patchset mentioned earlier, with interest if not
skill... :)
Rob