Hi!
> > On Tue, 2004-01-20 at 00:39, Benjamin Herrenschmidt wrote:
> > > I see no reason why this would be needed on ppc, only the last step,
> > > that is the actual CPU state save, should matter.
> >
> > Besides saving the CPU state, the code copies the original kernel back.
> > It sort of defeats the purpose to remove that code :>
>
> Ok, you mean copying the memory pages back down ? That should be done
> with hand-made assembly or C code located specifically elsewhere then.
> I do not want to see any kind of this ugly C-generated assembly in
> arch/ppc.
FYI, this is that "ugly C-generated assembly" we are talking about. I
do not think it is so bad.
.text
/* Originally gcc generated, modified by hand */
#include <linux/linkage.h>
#include <asm/segment.h>
#include <asm/page.h>
.text
ENTRY(do_magic)
pushl %ebx
cmpl $0,8(%esp)
jne .L1450
call do_magic_suspend_1
call save_processor_state
movl %esp, saved_context_esp
movl %eax, saved_context_eax
movl %ebx, saved_context_ebx
movl %ecx, saved_context_ecx
movl %edx, saved_context_edx
movl %ebp, saved_context_ebp
movl %esi, saved_context_esi
movl %edi, saved_context_edi
pushfl ; popl saved_context_eflags
call do_magic_suspend_2
jmp .L1449
.p2align 4,,7
.L1450:
movl $swapper_pg_dir-__PAGE_OFFSET,%ecx
movl %ecx,%cr3
call do_magic_resume_1
movl $0,loop
cmpl $0,nr_copy_pages
je .L1453
.p2align 4,,7
.L1455:
movl $0,loop2
.p2align 4,,7
.L1459:
movl pagedir_nosave,%ecx
movl loop,%eax
movl loop2,%edx
sall $4,%eax
movl 4(%ecx,%eax),%ebx
movl (%ecx,%eax),%eax
movb (%edx,%eax),%al
movb %al,(%edx,%ebx)
movl %cr3, %eax;
movl %eax, %cr3; # flush TLB
movl loop2,%eax
leal 1(%eax),%edx
movl %edx,loop2
movl %edx,%eax
cmpl $4095,%eax
jbe .L1459
movl loop,%eax
leal 1(%eax),%edx
movl %edx,loop
movl %edx,%eax
cmpl nr_copy_pages,%eax
jb .L1455
.p2align 4,,7
.L1453:
movl $__USER_DS,%eax
movw %ax, %ds
movw %ax, %es
movl saved_context_esp, %esp
movl saved_context_ebp, %ebp
movl saved_context_eax, %eax
movl saved_context_ebx, %ebx
movl saved_context_ecx, %ecx
movl saved_context_edx, %edx
movl saved_context_esi, %esi
movl saved_context_edi, %edi
call restore_processor_state
pushl saved_context_eflags ; popfl
call do_magic_resume_2
.L1449:
popl %ebx
ret
.section .data.nosave
loop:
.quad 0
loop2:
.quad 0
.previous
> > It is a well defined interface: a section of memory marked nosave, with
> > variables given the matching attribute. Not my idea, by the way. If you
> > have a problem, you should be taking it up with Pavel or Linus. We
> > should also note that the interface can't be too well defined - there
> > has to be room for development over time.
>
> Still... That makes assumptions about how it's located and organised
> that plain wrong (c). Please get rid of that, at least I won't let a
> PPC
FYI, there are exactly 6 variables in "nosave" section. Two loop
variables you can see in above code, one spinlock, number of pages to
save, pointer to directory of pages to be copied, and its length.
I could probably move spinlock and length of pgdir out of there...
> > Yes. we device_suspend. Regarding the similarities with kexec, I fully
> > agree. In fact, there is also LKCD to think off. There should be a
> > synergy here.
>
> device_suspend is, imho, hairy too. We have some semantics that need
> cleanup here, I'll have to talk to Patrick about them. Putting devices
> to an idle state is what you need and what kexec need, and doesn't mean
> putting them to _sleep_. Or maybe we could pass a specific state to
Okay, I can agree that putting them into sleep is not ideal [it
puzzles users, at least]. But it is quite simple and should work
okay. I do not want another round of device_suspend changes in
2.6.X... [Well, perhaps if it was done in right&compatible way (tm),
it would be acceptable...]
Pavel
--
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
