| Message ID | 20240925060101.259244-3-frediano.ziglio@cloud.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | x86/boot: Reduce assembly code | expand |
On 25/09/2024 7:00 am, Frediano Ziglio wrote: > The 2 code paths were sharing quite some common code, reuse it instead > of having duplications. > Use %dl register to store boot type before running common code. > Using a 8 bit register reduces code size. These final two lines are stale and can be dropped. > diff --git a/xen/arch/x86/boot/head.S b/xen/arch/x86/boot/head.S > index 267207e5a2..2d2f56ad22 100644 > --- a/xen/arch/x86/boot/head.S > +++ b/xen/arch/x86/boot/head.S > @@ -409,13 +411,27 @@ cs32_switch: > ELFNOTE(Xen, XEN_ELFNOTE_PHYS32_ENTRY, .long sym_offs(__pvh_start)) > > __pvh_start: > - cld > + mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ > + /* > + * Fall through into BIOS code. > + * We will use %eax to distinguish we came from PHV entry point. PVH. It occurs to me that we could actually have: mov $XEN_HVM_START_MAGIC_VALUE, %eax Given the cross-check against 0(%ebx) later, that's marginally more robust. > @@ -449,62 +458,40 @@ __pvh_start: > mov %ecx, %es > mov %ecx, %ss > > - /* Skip bootloader setup and bios setup, go straight to trampoline */ > - movb $1, sym_esi(pvh_boot) > - movb $1, sym_esi(skip_realmode) > + /* Load null selector to unused segment registers. */ > + xor %ecx, %ecx > + mov %ecx, %fs > + mov %ecx, %gs Honestly, the more I look at this, the more bizarre it is. We should just set up %fs/gs like we do %ds/es, which in this case is simply to drop the comment and the xor. > > - /* Set trampoline_phys to use mfn 1 to avoid having a mapping at VA 0 */ > - movw $0x1000, sym_esi(trampoline_phys) > - mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ > - jmp trampoline_setup > - > -#endif /* CONFIG_PVH_GUEST */ > - > -.Linitialise_bss: > /* Initialise the BSS. Preserve %eax (BOOTLOADER_MAGIC). */ > mov %eax, %ebp > - > lea sym_esi(__bss_start), %edi > lea sym_esi(__bss_end), %ecx > sub %edi, %ecx > xor %eax, %eax > shr $2, %ecx > rep stosl > - > mov %ebp, %eax Are these two dropped lines intentional? I did ask on the previous version; it's weird to introduce the code with them present, then delete them in the subsequent patch. > - ret > - > -__start: > - cld > - cli > - > - /* > - * Multiboot (both 1 and 2) specify the stack pointer as undefined > - * when entering in BIOS circumstances. This is unhelpful for > - * relocatable images, where one call (i.e. push) is required to > - * calculate the image's load address. > - * > - * This early in boot, there is one area of memory we know about with > - * reasonable confidence that it isn't overlapped by Xen, and that's > - * the Multiboot info structure in %ebx. Use it as a temporary stack. > - */ > > - /* Preserve the field we're about to clobber. */ > - mov (%ebx), %edx > - lea 4(%ebx), %esp > +#ifdef CONFIG_PVH_GUEST > + cmp $XEN_HVM_START_MAGIC_VALUE, %eax > + jne 1f > > - /* Calculate the load base address. */ > - call 1f > -1: pop %esi > - sub $sym_offs(1b), %esi > + mov %ebx, sym_esi(pvh_start_info_pa) > > - /* Restore the clobbered field. */ > - mov %edx, (%ebx) > + /* Force xen console. Will revert to user choice in init code. */ > + movb $-1, sym_esi(opt_console_xen) > > - /* Set up stack. */ > - lea STACK_SIZE - CPUINFO_sizeof + sym_esi(cpu0_stack), %esp > + /* Skip bootloader setup and bios setup, go straight to trampoline */ > + movb $1, sym_esi(pvh_boot) > + movb $1, sym_esi(skip_realmode) > > - call .Linitialise_bss > + /* Set trampoline_phys to use mfn 1 to avoid having a mapping at VA 0 */ > + movl $PAGE_SIZE, sym_esi(trampoline_phys) > + mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ This isn't needed any more. It was previously reloading %eax after REP STOSL, but that's sorted elsehow now. Overall, this bit of the diff is still hard to read, but it's the best we're going to get I think. The end result is good. I'm happy to adjust all of these on commit. ~Andrew
On 25.09.2024 21:33, Andrew Cooper wrote: > On 25/09/2024 7:00 am, Frediano Ziglio wrote: >> @@ -449,62 +458,40 @@ __pvh_start: >> mov %ecx, %es >> mov %ecx, %ss >> >> - /* Skip bootloader setup and bios setup, go straight to trampoline */ >> - movb $1, sym_esi(pvh_boot) >> - movb $1, sym_esi(skip_realmode) >> + /* Load null selector to unused segment registers. */ >> + xor %ecx, %ecx >> + mov %ecx, %fs >> + mov %ecx, %gs > > Honestly, the more I look at this, the more bizarre it is. > > We should just set up %fs/gs like we do %ds/es, which in this case is > simply to drop the comment and the xor. What's bizarre here? As long as we don't use %fs/%gs, it doesn't matter much what we set them to. So yes, they may be set to what %ds etc are set to, but they may as well be marked unusable. Documentation-wise that's cleaner imo, as down the road - when a need to use one arises - it then won't require auditing of all code to figure where the register(s) is(are) actually used (just to find: nowhere). Even if a comment to this effect was left here, I for one wouldn't trust it in a couple of years time, but rather fear it went stale. Jan
On Thu, Sep 26, 2024 at 7:50 AM Jan Beulich <jbeulich@suse.com> wrote: > > On 25.09.2024 21:33, Andrew Cooper wrote: > > On 25/09/2024 7:00 am, Frediano Ziglio wrote: > >> @@ -449,62 +458,40 @@ __pvh_start: > >> mov %ecx, %es > >> mov %ecx, %ss > >> > >> - /* Skip bootloader setup and bios setup, go straight to trampoline */ > >> - movb $1, sym_esi(pvh_boot) > >> - movb $1, sym_esi(skip_realmode) > >> + /* Load null selector to unused segment registers. */ > >> + xor %ecx, %ecx > >> + mov %ecx, %fs > >> + mov %ecx, %gs > > > > Honestly, the more I look at this, the more bizarre it is. > > > > We should just set up %fs/gs like we do %ds/es, which in this case is > > simply to drop the comment and the xor. > > What's bizarre here? As long as we don't use %fs/%gs, it doesn't matter > much what we set them to. So yes, they may be set to what %ds etc are set > to, but they may as well be marked unusable. Documentation-wise that's > cleaner imo, as down the road - when a need to use one arises - it then > won't require auditing of all code to figure where the register(s) is(are) > actually used (just to find: nowhere). Even if a comment to this effect > was left here, I for one wouldn't trust it in a couple of years time, but > rather fear it went stale. > > Jan Hi, are you against this change and asking to roll it back? Frediano
On 26.09.2024 11:24, Frediano Ziglio wrote: > On Thu, Sep 26, 2024 at 7:50 AM Jan Beulich <jbeulich@suse.com> wrote: >> >> On 25.09.2024 21:33, Andrew Cooper wrote: >>> On 25/09/2024 7:00 am, Frediano Ziglio wrote: >>>> @@ -449,62 +458,40 @@ __pvh_start: >>>> mov %ecx, %es >>>> mov %ecx, %ss >>>> >>>> - /* Skip bootloader setup and bios setup, go straight to trampoline */ >>>> - movb $1, sym_esi(pvh_boot) >>>> - movb $1, sym_esi(skip_realmode) >>>> + /* Load null selector to unused segment registers. */ >>>> + xor %ecx, %ecx >>>> + mov %ecx, %fs >>>> + mov %ecx, %gs >>> >>> Honestly, the more I look at this, the more bizarre it is. >>> >>> We should just set up %fs/gs like we do %ds/es, which in this case is >>> simply to drop the comment and the xor. >> >> What's bizarre here? As long as we don't use %fs/%gs, it doesn't matter >> much what we set them to. So yes, they may be set to what %ds etc are set >> to, but they may as well be marked unusable. Documentation-wise that's >> cleaner imo, as down the road - when a need to use one arises - it then >> won't require auditing of all code to figure where the register(s) is(are) >> actually used (just to find: nowhere). Even if a comment to this effect >> was left here, I for one wouldn't trust it in a couple of years time, but >> rather fear it went stale. > > Hi, > are you against this change and asking to roll it back? Well, first of all I'm hoping for a reply by Andrew. Maybe I'm overlooking something. As it stands right now I indeed think we'd be better off keeping nul selectors in %fs and %gs. Jan
diff --git a/xen/arch/x86/boot/head.S b/xen/arch/x86/boot/head.S index 267207e5a2..2d2f56ad22 100644 --- a/xen/arch/x86/boot/head.S +++ b/xen/arch/x86/boot/head.S @@ -25,6 +25,8 @@ #define MB2_HT(name) (MULTIBOOT2_HEADER_TAG_##name) #define MB2_TT(name) (MULTIBOOT2_TAG_TYPE_##name) +#define XEN_HVM_START_MAGIC_VALUE 0x336ec578 + .macro mb2ht_args arg:req, args:vararg .long \arg .ifnb \args @@ -409,13 +411,27 @@ cs32_switch: ELFNOTE(Xen, XEN_ELFNOTE_PHYS32_ENTRY, .long sym_offs(__pvh_start)) __pvh_start: - cld + mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ + /* + * Fall through into BIOS code. + * We will use %eax to distinguish we came from PHV entry point. + */ +#endif /* CONFIG_PVH_GUEST */ + +__start: cli + cld /* - * We need one call (i.e. push) to determine the load address. See - * __start for a discussion on how to do this safely using the PVH - * info structure. + * Multiboot (both 1 and 2) and PVH specify the stack pointer as + * undefined. This is unhelpful for relocatable images, where one + * call (i.e. push) is required to calculate the image's load address. + * + * Durig BIOS boot, there is one area of memory we know about with + * reasonable confidence that it isn't overlapped by Xen, and that's + * the Multiboot info structure in %ebx. Use it as a temporary stack. + * + * During PVH boot use info structure in %ebx. */ /* Preserve the field we're about to clobber. */ @@ -433,14 +449,7 @@ __pvh_start: /* Set up stack. */ lea STACK_SIZE - CPUINFO_sizeof + sym_esi(cpu0_stack), %esp - call .Linitialise_bss - - mov %ebx, sym_esi(pvh_start_info_pa) - - /* Force xen console. Will revert to user choice in init code. */ - movb $-1, sym_esi(opt_console_xen) - - /* Prepare gdt and segments */ + /* Initialize GDTR and basic data segments. */ add %esi, sym_esi(gdt_boot_base) lgdt sym_esi(gdt_boot_descr) @@ -449,62 +458,40 @@ __pvh_start: mov %ecx, %es mov %ecx, %ss - /* Skip bootloader setup and bios setup, go straight to trampoline */ - movb $1, sym_esi(pvh_boot) - movb $1, sym_esi(skip_realmode) + /* Load null selector to unused segment registers. */ + xor %ecx, %ecx + mov %ecx, %fs + mov %ecx, %gs - /* Set trampoline_phys to use mfn 1 to avoid having a mapping at VA 0 */ - movw $0x1000, sym_esi(trampoline_phys) - mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ - jmp trampoline_setup - -#endif /* CONFIG_PVH_GUEST */ - -.Linitialise_bss: /* Initialise the BSS. Preserve %eax (BOOTLOADER_MAGIC). */ mov %eax, %ebp - lea sym_esi(__bss_start), %edi lea sym_esi(__bss_end), %ecx sub %edi, %ecx xor %eax, %eax shr $2, %ecx rep stosl - mov %ebp, %eax - ret - -__start: - cld - cli - - /* - * Multiboot (both 1 and 2) specify the stack pointer as undefined - * when entering in BIOS circumstances. This is unhelpful for - * relocatable images, where one call (i.e. push) is required to - * calculate the image's load address. - * - * This early in boot, there is one area of memory we know about with - * reasonable confidence that it isn't overlapped by Xen, and that's - * the Multiboot info structure in %ebx. Use it as a temporary stack. - */ - /* Preserve the field we're about to clobber. */ - mov (%ebx), %edx - lea 4(%ebx), %esp +#ifdef CONFIG_PVH_GUEST + cmp $XEN_HVM_START_MAGIC_VALUE, %eax + jne 1f - /* Calculate the load base address. */ - call 1f -1: pop %esi - sub $sym_offs(1b), %esi + mov %ebx, sym_esi(pvh_start_info_pa) - /* Restore the clobbered field. */ - mov %edx, (%ebx) + /* Force xen console. Will revert to user choice in init code. */ + movb $-1, sym_esi(opt_console_xen) - /* Set up stack. */ - lea STACK_SIZE - CPUINFO_sizeof + sym_esi(cpu0_stack), %esp + /* Skip bootloader setup and bios setup, go straight to trampoline */ + movb $1, sym_esi(pvh_boot) + movb $1, sym_esi(skip_realmode) - call .Linitialise_bss + /* Set trampoline_phys to use mfn 1 to avoid having a mapping at VA 0 */ + movl $PAGE_SIZE, sym_esi(trampoline_phys) + mov (%ebx), %eax /* mov $XEN_HVM_START_MAGIC_VALUE, %eax */ + jmp trampoline_setup +1: +#endif /* CONFIG_PVH_GUEST */ /* Bootloaders may set multiboot{1,2}.mem_lower to a nonzero value. */ xor %edx,%edx @@ -563,22 +550,7 @@ __start: trampoline_bios_setup: /* * Called on legacy BIOS platforms only. - * - * Initialize GDTR and basic data segments. */ - add %esi,sym_esi(gdt_boot_base) - lgdt sym_esi(gdt_boot_descr) - - mov $BOOT_DS,%ecx - mov %ecx,%ds - mov %ecx,%es - mov %ecx,%ss - /* %esp is initialized later. */ - - /* Load null descriptor to unused segment registers. */ - xor %ecx,%ecx - mov %ecx,%fs - mov %ecx,%gs /* Set up trampoline segment 64k below EBDA */ movzwl 0x40e,%ecx /* EBDA segment */
The 2 code paths were sharing quite some common code, reuse it instead of having duplications. Use %dl register to store boot type before running common code. Using a 8 bit register reduces code size. Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com> --- Changes since v1: - use %dl instead of %ebp to reduce code size; - fold cli instruction; - update some comments. Changes since v3: - dropped %dl and constant, distinguish entry by magic. --- xen/arch/x86/boot/head.S | 108 +++++++++++++++------------------------ 1 file changed, 40 insertions(+), 68 deletions(-)