* [PATCH] x86, reboot: skip DMI checks if reboot set by user @ 2012-01-17 15:16 Michael D Labriola 2012-01-17 19:58 ` Alan Cox 0 siblings, 1 reply; 11+ messages in thread From: Michael D Labriola @ 2012-01-17 15:16 UTC (permalink / raw) To: Ingo Molnar, Matthew Garrett Cc: Michael D Labriola, H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, support, Thomas Gleixner, x86 This commit causes DMI checks for vendor specific reboot quirks to be skipped if the user passed in a reboot= arg on the command line. Signed-off-by: Michael D Labriola <mlabriol@gdeb.com> --- arch/x86/kernel/reboot.c | 16 ++++++++++++++-- 1 files changed, 14 insertions(+), 2 deletions(-) diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 37a458b..4240d58 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -38,6 +38,7 @@ static const struct desc_ptr no_idt = {}; static int reboot_mode; enum reboot_type reboot_type = BOOT_ACPI; int reboot_force; +int reboot_user; #if defined(CONFIG_X86_32) && defined(CONFIG_SMP) static int reboot_cpu = -1; @@ -70,10 +71,12 @@ static int __init reboot_setup(char *str) switch (*str) { case 'w': reboot_mode = 0x1234; + reboot_user = 1; break; case 'c': reboot_mode = 0; + reboot_user = 1; break; #ifdef CONFIG_X86_32 @@ -87,6 +90,7 @@ static int __init reboot_setup(char *str) /* we will leave sorting out the final value when we are ready to reboot, since we might not have detected BSP APIC ID or smp_num_cpu */ + reboot_user = 1; break; #endif /* CONFIG_SMP */ @@ -98,10 +102,12 @@ static int __init reboot_setup(char *str) case 'e': case 'p': reboot_type = *str; + reboot_user = 1; break; case 'f': reboot_force = 1; + reboot_user = 1; break; } @@ -316,7 +322,10 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { static int __init reboot_init(void) { - dmi_check_system(reboot_dmi_table); + /* Skip the DMI check if user set reboot= on the command line */ + if (!reboot_user) { + dmi_check_system(reboot_dmi_table); + } return 0; } core_initcall(reboot_init); @@ -465,7 +474,10 @@ static struct dmi_system_id __initdata pci_reboot_dmi_table[] = { static int __init pci_reboot_init(void) { - dmi_check_system(pci_reboot_dmi_table); + /* Skip the DMI check if user set reboot= on the command line */ + if (!reboot_user) { + dmi_check_system(pci_reboot_dmi_table); + } return 0; } core_initcall(pci_reboot_init); -- 1.6.0.3 ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-17 15:16 [PATCH] x86, reboot: skip DMI checks if reboot set by user Michael D Labriola @ 2012-01-17 19:58 ` Alan Cox 2012-01-19 15:32 ` Michael D Labriola 0 siblings, 1 reply; 11+ messages in thread From: Alan Cox @ 2012-01-17 19:58 UTC (permalink / raw) To: Michael D Labriola Cc: Ingo Molnar, Matthew Garrett, H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, support, Thomas Gleixner, x86 On Tue, 17 Jan 2012 10:16:03 -0500 Michael D Labriola <mlabriol@gdeb.com> wrote: > This commit causes DMI checks for vendor specific reboot > quirks to be skipped if the user passed in a reboot= arg > on the command line. > > Signed-off-by: Michael D Labriola <mlabriol@gdeb.com> > --- > arch/x86/kernel/reboot.c | 16 ++++++++++++++-- > 1 files changed, 14 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c > index 37a458b..4240d58 100644 > --- a/arch/x86/kernel/reboot.c > +++ b/arch/x86/kernel/reboot.c > @@ -38,6 +38,7 @@ static const struct desc_ptr no_idt = {}; > static int reboot_mode; > enum reboot_type reboot_type = BOOT_ACPI; > int reboot_force; > +int reboot_user; static ... > case 'f': > reboot_force = 1; > + reboot_user = 1; Cleaner would be to add a BOOT_ACPI_FORCE perhaps so you can tell ACPI v ACPI specified by user > break; > } > > @@ -316,7 +322,10 @@ static struct dmi_system_id __initdata > reboot_dmi_table[] = { > > static int __init reboot_init(void) > { > - dmi_check_system(reboot_dmi_table); > + /* Skip the DMI check if user set reboot= on the command line */ > + if (!reboot_user) { then do if (reboot_type == BOOT_ACPI) > + dmi_check_system(reboot_dmi_table); > + } ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-17 19:58 ` Alan Cox @ 2012-01-19 15:32 ` Michael D Labriola 2012-01-19 15:45 ` Alan Cox 0 siblings, 1 reply; 11+ messages in thread From: Michael D Labriola @ 2012-01-19 15:32 UTC (permalink / raw) To: Alan Cox Cc: H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 Alan Cox <alan@lxorguk.ukuu.org.uk> wrote on 01/17/2012 02:58:41 PM: > On Tue, 17 Jan 2012 10:16:03 -0500 > Michael D Labriola <mlabriol@gdeb.com> wrote: > > > This commit causes DMI checks for vendor specific reboot > > quirks to be skipped if the user passed in a reboot= arg > > on the command line. > > > > Signed-off-by: Michael D Labriola <mlabriol@gdeb.com> > > --- > > arch/x86/kernel/reboot.c | 16 ++++++++++++++-- > > 1 files changed, 14 insertions(+), 2 deletions(-) > > > > diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c > > index 37a458b..4240d58 100644 > > --- a/arch/x86/kernel/reboot.c > > +++ b/arch/x86/kernel/reboot.c > > @@ -38,6 +38,7 @@ static const struct desc_ptr no_idt = {}; > > static int reboot_mode; > > enum reboot_type reboot_type = BOOT_ACPI; > > int reboot_force; > > +int reboot_user; > > static ... Oops, good point. > > case 'f': > > reboot_force = 1; > > + reboot_user = 1; > > Cleaner would be to add a > > BOOT_ACPI_FORCE perhaps so you can tell ACPI v ACPI specified by user > > > break; > > } > > > > @@ -316,7 +322,10 @@ static struct dmi_system_id __initdata > > reboot_dmi_table[] = { > > > > static int __init reboot_init(void) > > { > > - dmi_check_system(reboot_dmi_table); > > + /* Skip the DMI check if user set reboot= on the command line */ > > + if (!reboot_user) { > > then do > > if (reboot_type == BOOT_ACPI) > > + dmi_check_system(reboot_dmi_table); > > + } This seems a little less obvious to the casual observer than the way I implemented it... but maybe that's just me. It would indeed disable DMI checking if reboot_type's value is anything other than BOOT_ACPI (which is the default), but that does assume BOOT_ACPI is the default. What if that's ever changed? Seems likely to me that whoever changes the default in the future could very easily re-break this at that point. Thoughts? I'm going to resubmit this patche anyway, as Ingo pointed out I've got line-wrap issues. --- Michael D Labriola Electric Boat mlabriol@gdeb.com 401-848-8871 (desk) 401-848-8513 (lab) 401-316-9844 (cell) ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 15:32 ` Michael D Labriola @ 2012-01-19 15:45 ` Alan Cox 2012-01-19 15:48 ` Michael D Labriola [not found] ` <OF8642E997.5ED041E1-ON8525798A.0056983A-8525798A.0056E05B@LocalDomain> 0 siblings, 2 replies; 11+ messages in thread From: Alan Cox @ 2012-01-19 15:45 UTC (permalink / raw) To: Michael D Labriola Cc: H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 > checking if reboot_type's value is anything other than BOOT_ACPI (which > is the default), but that does assume BOOT_ACPI is the default. What if > that's ever changed? Seems likely to me that whoever changes the default > in the future could very easily re-break this at that point. > > Thoughts? Fair point - how about a BOOT_DEFAULT define value if it hasn't been set and we can just make it do the same as ACPI in the switch ? Alan ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 15:45 ` Alan Cox @ 2012-01-19 15:48 ` Michael D Labriola [not found] ` <OF8642E997.5ED041E1-ON8525798A.0056983A-8525798A.0056E05B@LocalDomain> 1 sibling, 0 replies; 11+ messages in thread From: Michael D Labriola @ 2012-01-19 15:48 UTC (permalink / raw) To: Alan Cox Cc: H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 Alan Cox <alan@lxorguk.ukuu.org.uk> wrote on 01/19/2012 10:45:04 AM: > > checking if reboot_type's value is anything other than BOOT_ACPI (which > > is the default), but that does assume BOOT_ACPI is the default. What if > > that's ever changed? Seems likely to me that whoever changes the default > > in the future could very easily re-break this at that point. > > > > Thoughts? > > Fair point - how about a BOOT_DEFAULT define value if it hasn't been set > and we can just make it do the same as ACPI in the switch ? Sounds good to me. I'll whip that up and resubmit. --- Michael D Labriola Electric Boat mlabriol@gdeb.com 401-848-8871 (desk) 401-848-8513 (lab) 401-316-9844 (cell) ^ permalink raw reply [flat|nested] 11+ messages in thread
[parent not found: <OF8642E997.5ED041E1-ON8525798A.0056983A-8525798A.0056E05B@LocalDomain>]
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user [not found] ` <OF8642E997.5ED041E1-ON8525798A.0056983A-8525798A.0056E05B@LocalDomain> @ 2012-01-19 17:46 ` Michael D Labriola 2012-01-19 17:52 ` H. Peter Anvin 0 siblings, 1 reply; 11+ messages in thread From: Michael D Labriola @ 2012-01-19 17:46 UTC (permalink / raw) To: Michael D Labriola Cc: Alan Cox, H. Peter Anvin, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 Michael D Labriola/EB/GDYN wrote on 01/19/2012 10:48:53 AM: > Alan Cox <alan@lxorguk.ukuu.org.uk> wrote on 01/19/2012 10:45:04 AM: > > > > checking if reboot_type's value is anything other than BOOT_ACPI (which > > > is the default), but that does assume BOOT_ACPI is the default. What if > > > that's ever changed? Seems likely to me that whoever changes the default > > > in the future could very easily re-break this at that point. > > > > > > Thoughts? > > > > Fair point - how about a BOOT_DEFAULT define value if it hasn't been set > > and we can just make it do the same as ACPI in the switch ? > > Sounds good to me. I'll whip that up and resubmit. Well, how about this. Using a #define won't do as far as I can tell. But I can reduce the size of the diff considerably like this. Random question... why do we have a reboot_init function that does DMI checking with reboot_dmi_table (callbacks are mostly set_bios_reboot, but there is a single set_kbd_reboot) and also a pci_reboot_init which does the DMI check again using a separate pci_reboot_dmi_table (all callbacks are to set_pci_reboot)? Wouldn't it make more sense to do a single DMI scan using one big table? diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index 37a458b..0fc5b31 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -39,6 +39,14 @@ static int reboot_mode; enum reboot_type reboot_type = BOOT_ACPI; int reboot_force; +/* This variable is used privately to keep track of whether or not + * reboot_type is still set to its default value (i.e., reboot= hasn't + * been set on the command line). This is needed so that we can + * suppress DMI scanning for reboot quirks. Without it, it's + * impossible to override a faulty reboot quirk without recompiling. + */ +static int reboot_default = 1; + #if defined(CONFIG_X86_32) && defined(CONFIG_SMP) static int reboot_cpu = -1; #endif @@ -67,6 +75,12 @@ bool port_cf9_safe = false; static int __init reboot_setup(char *str) { for (;;) { + /* Having anything passed on the command line via + * reboot= will cause us to disable DMI checking + * below. + */ + reboot_default = 0; + switch (*str) { case 'w': reboot_mode = 0x1234; @@ -316,7 +330,12 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { static int __init reboot_init(void) { - dmi_check_system(reboot_dmi_table); + /* Only do the DMI check if reboot_type hasn't been overridden + * on the command line + */ + if (reboot_default) { + dmi_check_system(reboot_dmi_table); + } return 0; } core_initcall(reboot_init); @@ -465,7 +484,12 @@ static struct dmi_system_id __initdata pci_reboot_dmi_table[] = { static int __init pci_reboot_init(void) { - dmi_check_system(pci_reboot_dmi_table); + /* Only do the DMI check if reboot_type hasn't been overridden + * on the command line + */ + if (reboot_default) { + dmi_check_system(pci_reboot_dmi_table); + } return 0; } core_initcall(pci_reboot_init); ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 17:46 ` Michael D Labriola @ 2012-01-19 17:52 ` H. Peter Anvin 2012-01-19 19:14 ` Michael D Labriola 0 siblings, 1 reply; 11+ messages in thread From: H. Peter Anvin @ 2012-01-19 17:52 UTC (permalink / raw) To: Michael D Labriola Cc: Alan Cox, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 On 01/19/2012 09:46 AM, Michael D Labriola wrote: > > Random question... why do we have a reboot_init function that does DMI > checking with reboot_dmi_table (callbacks are mostly set_bios_reboot, but > there is a single set_kbd_reboot) and also a pci_reboot_init which does > the DMI check again using a separate pci_reboot_dmi_table (all callbacks > are to set_pci_reboot)? > > Wouldn't it make more sense to do a single DMI scan using one big table? > Yes, and such a patch would be appreciated. The reason it is as it is dates back to before the 32-64 bit unification, as far as I know. (BIOS reboot is currently not supported on 64 bits, mainly.) -hpa ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 17:52 ` H. Peter Anvin @ 2012-01-19 19:14 ` Michael D Labriola 2012-01-19 19:17 ` H. Peter Anvin 2012-01-19 19:41 ` Ingo Molnar 0 siblings, 2 replies; 11+ messages in thread From: Michael D Labriola @ 2012-01-19 19:14 UTC (permalink / raw) To: H. Peter Anvin Cc: Alan Cox, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 "H. Peter Anvin" <hpa@zytor.com> wrote on 01/19/2012 12:52:15 PM: > On 01/19/2012 09:46 AM, Michael D Labriola wrote: > > > > Random question... why do we have a reboot_init function that does DMI > > checking with reboot_dmi_table (callbacks are mostly set_bios_reboot, but > > there is a single set_kbd_reboot) and also a pci_reboot_init which does > > the DMI check again using a separate pci_reboot_dmi_table (all callbacks > > are to set_pci_reboot)? > > > > Wouldn't it make more sense to do a single DMI scan using one big table? > > > > Yes, and such a patch would be appreciated. > > The reason it is as it is dates back to before the 32-64 bit > unification, as far as I know. > > (BIOS reboot is currently not supported on 64 bits, mainly.) Well, that does complicate it a bit. I'll gin something up and see what you think. I guess it will involve having an #ifdef CONFIG_X86_32 block inside a single dmi_table structure for the BIOS quirks. Actually, set_kbd_reboot is inside the current X86_32 only block, along with the one DMI callback that uses it. Is this correct? --- Michael D Labriola Electric Boat mlabriol@gdeb.com 401-848-8871 (desk) 401-848-8513 (lab) 401-316-9844 (cell) ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 19:14 ` Michael D Labriola @ 2012-01-19 19:17 ` H. Peter Anvin 2012-01-19 19:34 ` Michael D Labriola 2012-01-19 19:41 ` Ingo Molnar 1 sibling, 1 reply; 11+ messages in thread From: H. Peter Anvin @ 2012-01-19 19:17 UTC (permalink / raw) To: Michael D Labriola Cc: Alan Cox, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 On 01/19/2012 11:14 AM, Michael D Labriola wrote: >> >> Yes, and such a patch would be appreciated. >> >> The reason it is as it is dates back to before the 32-64 bit >> unification, as far as I know. >> >> (BIOS reboot is currently not supported on 64 bits, mainly.) > > Well, that does complicate it a bit. I'll gin something up and see what > you think. I guess it will involve having an #ifdef CONFIG_X86_32 block > inside a single dmi_table structure for the BIOS quirks. > > Actually, set_kbd_reboot is inside the current X86_32 only block, along > with the one DMI callback that uses it. Is this correct? > Probably not, although I suspect most of the users of that are 32-bit only systems. -hpa -- H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf. ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 19:17 ` H. Peter Anvin @ 2012-01-19 19:34 ` Michael D Labriola 0 siblings, 0 replies; 11+ messages in thread From: Michael D Labriola @ 2012-01-19 19:34 UTC (permalink / raw) To: H. Peter Anvin Cc: Alan Cox, Kushal Koolwal, linux-kernel, michael.d.labriola, Ingo Molnar, Matthew Garrett, support, Thomas Gleixner, x86 H. Peter Anvin" <hpa@zytor.com> wrote on 01/19/2012 02:17:54 PM: > On 01/19/2012 11:14 AM, Michael D Labriola wrote: > >> > >> Yes, and such a patch would be appreciated. > >> > >> The reason it is as it is dates back to before the 32-64 bit > >> unification, as far as I know. > >> > >> (BIOS reboot is currently not supported on 64 bits, mainly.) > > > > Well, that does complicate it a bit. I'll gin something up and see what > > you think. I guess it will involve having an #ifdef CONFIG_X86_32 block > > inside a single dmi_table structure for the BIOS quirks. > > > > Actually, set_kbd_reboot is inside the current X86_32 only block, along > > with the one DMI callback that uses it. Is this correct? > > > > Probably not, although I suspect most of the users of that are 32-bit > only systems. How does this look? The patch looks kinda nasty because of how much code is getting moved around... Basically, all I did was move the reboot_init method and reboot_dmi_table out of the X86_32 block they were in, put the quirks that set BIOS reboot inside an X86_32 define block, and then added all the PCI quirks into the new, single DMI table. I did also move the set_kbd_reboot method out of the X86_32 block, since all the documentation I've run into in the kernel suggests that it's valid for X86_64 as well. I even tested it by adding an entry to reboot_dmi_table for my machine and verified that behavior is the same as before the reorg. Note that this patch got generated from my test tree, so it'll have conflicts if applied against v3.2. I'll rebase it and weed that stuff out if you think it's a good idea. diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c index d840e69..e739737 100644 --- a/arch/x86/kernel/reboot.c +++ b/arch/x86/kernel/reboot.c @@ -150,6 +150,80 @@ static int __init set_bios_reboot(const struct dmi_system_id *d) return 0; } +extern const unsigned char machine_real_restart_asm[]; +extern const u64 machine_real_restart_gdt[3]; + +void machine_real_restart(unsigned int type) +{ + void *restart_va; + unsigned long restart_pa; + void (*restart_lowmem)(unsigned int); + u64 *lowmem_gdt; + + local_irq_disable(); + + /* Write zero to CMOS register number 0x0f, which the BIOS POST + routine will recognize as telling it to do a proper reboot. (Well + that's what this book in front of me says -- it may only apply to + the Phoenix BIOS though, it's not clear). At the same time, + disable NMIs by setting the top bit in the CMOS address register, + as we're about to do peculiar things to the CPU. I'm not sure if + `outb_p' is needed instead of just `outb'. Use it to be on the + safe side. (Yes, CMOS_WRITE does outb_p's. - Paul G.) + */ + spin_lock(&rtc_lock); + CMOS_WRITE(0x00, 0x8f); + spin_unlock(&rtc_lock); + + /* + * Switch back to the initial page table. + */ + load_cr3(initial_page_table); + + /* Write 0x1234 to absolute memory location 0x472. The BIOS reads + this on booting to tell it to "Bypass memory test (also warm + boot)". This seems like a fairly standard thing that gets set by + REBOOT.COM programs, and the previous reset routine did this + too. */ + *((unsigned short *)0x472) = reboot_mode; + + /* Patch the GDT in the low memory trampoline */ + lowmem_gdt = TRAMPOLINE_SYM(machine_real_restart_gdt); + + restart_va = TRAMPOLINE_SYM(machine_real_restart_asm); + restart_pa = virt_to_phys(restart_va); + restart_lowmem = (void (*)(unsigned int))restart_pa; + + /* GDT[0]: GDT self-pointer */ + lowmem_gdt[0] = + (u64)(sizeof(machine_real_restart_gdt) - 1) + + ((u64)virt_to_phys(lowmem_gdt) << 16); + /* GDT[1]: 64K real mode code segment */ + lowmem_gdt[1] = + GDT_ENTRY(0x009b, restart_pa, 0xffff); + + /* Jump to the identity-mapped low memory code */ + restart_lowmem(type); +} +#ifdef CONFIG_APM_MODULE +EXPORT_SYMBOL(machine_real_restart); +#endif + +#endif /* CONFIG_X86_32 */ + +/* + * Some Apple MacBook and MacBookPro's needs reboot=p to be able to reboot + */ +static int __init set_pci_reboot(const struct dmi_system_id *d) +{ + if (reboot_type != BOOT_CF9) { + reboot_type = BOOT_CF9; + printk(KERN_INFO "%s series board detected. " + "Selecting PCI-method for reboots.\n", d->ident); + } + return 0; +} + static int __init set_kbd_reboot(const struct dmi_system_id *d) { if (reboot_type != BOOT_KBD) { @@ -159,7 +233,11 @@ static int __init set_kbd_reboot(const struct dmi_system_id *d) return 0; } +/* This is a single dmi_table handling all reboot quirks. Note that + * REBOOT_BIOS is only available for 32bit + */ static struct dmi_system_id __initdata reboot_dmi_table[] = { +#ifdef CONFIG_X86_32 { /* Handle problems with rebooting on Dell E520's */ .callback = set_bios_reboot, .ident = "Dell E520", @@ -309,6 +387,8 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { DMI_MATCH(DMI_BOARD_NAME, "P4S800"), }, }, +#endif /* CONFIG_X86_32 */ + { /* Handle reboot issue on Acer Aspire one */ .callback = set_kbd_reboot, .ident = "Acer Aspire One A110", @@ -317,96 +397,6 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = { DMI_MATCH(DMI_PRODUCT_NAME, "AOA110"), }, }, - { } -}; - -static int __init reboot_init(void) -{ - /* Only do the DMI check if reboot_type hasn't been overridden - * on the command line - */ - if (reboot_default) { - dmi_check_system(reboot_dmi_table); - } - return 0; -} -core_initcall(reboot_init); - -extern const unsigned char machine_real_restart_asm[]; -extern const u64 machine_real_restart_gdt[3]; - -void machine_real_restart(unsigned int type) -{ - void *restart_va; - unsigned long restart_pa; - void (*restart_lowmem)(unsigned int); - u64 *lowmem_gdt; - - local_irq_disable(); - - /* Write zero to CMOS register number 0x0f, which the BIOS POST - routine will recognize as telling it to do a proper reboot. (Well - that's what this book in front of me says -- it may only apply to - the Phoenix BIOS though, it's not clear). At the same time, - disable NMIs by setting the top bit in the CMOS address register, - as we're about to do peculiar things to the CPU. I'm not sure if - `outb_p' is needed instead of just `outb'. Use it to be on the - safe side. (Yes, CMOS_WRITE does outb_p's. - Paul G.) - */ - spin_lock(&rtc_lock); - CMOS_WRITE(0x00, 0x8f); - spin_unlock(&rtc_lock); - - /* - * Switch back to the initial page table. - */ - load_cr3(initial_page_table); - - /* Write 0x1234 to absolute memory location 0x472. The BIOS reads - this on booting to tell it to "Bypass memory test (also warm - boot)". This seems like a fairly standard thing that gets set by - REBOOT.COM programs, and the previous reset routine did this - too. */ - *((unsigned short *)0x472) = reboot_mode; - - /* Patch the GDT in the low memory trampoline */ - lowmem_gdt = TRAMPOLINE_SYM(machine_real_restart_gdt); - - restart_va = TRAMPOLINE_SYM(machine_real_restart_asm); - restart_pa = virt_to_phys(restart_va); - restart_lowmem = (void (*)(unsigned int))restart_pa; - - /* GDT[0]: GDT self-pointer */ - lowmem_gdt[0] = - (u64)(sizeof(machine_real_restart_gdt) - 1) + - ((u64)virt_to_phys(lowmem_gdt) << 16); - /* GDT[1]: 64K real mode code segment */ - lowmem_gdt[1] = - GDT_ENTRY(0x009b, restart_pa, 0xffff); - - /* Jump to the identity-mapped low memory code */ - restart_lowmem(type); -} -#ifdef CONFIG_APM_MODULE -EXPORT_SYMBOL(machine_real_restart); -#endif - -#endif /* CONFIG_X86_32 */ - -/* - * Some Apple MacBook and MacBookPro's needs reboot=p to be able to reboot - */ -static int __init set_pci_reboot(const struct dmi_system_id *d) -{ - if (reboot_type != BOOT_CF9) { - reboot_type = BOOT_CF9; - printk(KERN_INFO "%s series board detected. " - "Selecting PCI-method for reboots.\n", d->ident); - } - return 0; -} - -static struct dmi_system_id __initdata pci_reboot_dmi_table[] = { { /* Handle problems with rebooting on Apple MacBook5 */ .callback = set_pci_reboot, .ident = "Apple MacBook5", @@ -474,17 +464,17 @@ static struct dmi_system_id __initdata pci_reboot_dmi_table[] = { { } }; -static int __init pci_reboot_init(void) +static int __init reboot_init(void) { /* Only do the DMI check if reboot_type hasn't been overridden * on the command line */ if (reboot_default) { - dmi_check_system(pci_reboot_dmi_table); + dmi_check_system(reboot_dmi_table); } return 0; } -core_initcall(pci_reboot_init); +core_initcall(reboot_init); static inline void kb_wait(void) { ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] x86, reboot: skip DMI checks if reboot set by user 2012-01-19 19:14 ` Michael D Labriola 2012-01-19 19:17 ` H. Peter Anvin @ 2012-01-19 19:41 ` Ingo Molnar 1 sibling, 0 replies; 11+ messages in thread From: Ingo Molnar @ 2012-01-19 19:41 UTC (permalink / raw) To: Michael D Labriola Cc: H. Peter Anvin, Alan Cox, Kushal Koolwal, linux-kernel, michael.d.labriola, Matthew Garrett, support, Thomas Gleixner, x86 * Michael D Labriola <mlabriol@gdeb.com> wrote: > > Yes, and such a patch would be appreciated. > > > > The reason it is as it is dates back to before the 32-64 bit > > unification, as far as I know. > > > > (BIOS reboot is currently not supported on 64 bits, mainly.) > > Well, that does complicate it a bit. I'll gin something up > and see what you think. I guess it will involve having an > #ifdef CONFIG_X86_32 block inside a single dmi_table structure > for the BIOS quirks. Btw., i'd suggest to keep the two changes in separate patches, so that if any causes problems we can bisect to it separately. The more patches you can reasonably split it up into, the better. Thanks, Ingo ^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2012-01-19 19:41 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-01-17 15:16 [PATCH] x86, reboot: skip DMI checks if reboot set by user Michael D Labriola
2012-01-17 19:58 ` Alan Cox
2012-01-19 15:32 ` Michael D Labriola
2012-01-19 15:45 ` Alan Cox
2012-01-19 15:48 ` Michael D Labriola
[not found] ` <OF8642E997.5ED041E1-ON8525798A.0056983A-8525798A.0056E05B@LocalDomain>
2012-01-19 17:46 ` Michael D Labriola
2012-01-19 17:52 ` H. Peter Anvin
2012-01-19 19:14 ` Michael D Labriola
2012-01-19 19:17 ` H. Peter Anvin
2012-01-19 19:34 ` Michael D Labriola
2012-01-19 19:41 ` Ingo Molnar
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).