Messages in this thread Patch in this message | | | From | Andrew Cooper <> | Subject | [PATCH] x86/smpboot: Remove 486-isms from the modern AP boot path | Date | Wed, 25 Mar 2020 10:14:31 +0000 |
| |
Linux has an implementation of the Universal Start-up Algorithm (MP spec, Appendix B.4, Application Processor Startup), which includes unconditionally writing to the Bios Data Area and CMOS registers.
The warm reset vector is only necessary in the non-integrated Local APIC case. UV and Jailhouse already have an opt-out for this behaviour, but blindly using the BDA and CMOS on a UEFI or other reduced hardware system isn't clever.
Drop the warm_reset flag from struct x86_legacy_features, and tie the warm vector modifications to the integrated-ness of the Local APIC. This has the advantage of compiling the warm reset logic out entirely for 64bit builds.
CC: Thomas Gleixner <tglx@linutronix.de> CC: Ingo Molnar <mingo@redhat.com> CC: Borislav Petkov <bp@alien8.de> CC: "H. Peter Anvin" <hpa@zytor.com> CC: x86@kernel.org CC: Jan Kiszka <jan.kiszka@siemens.com> CC: James Morris <jmorris@namei.org> CC: David Howells <dhowells@redhat.com> CC: Andrew Cooper <andrew.cooper3@citrix.com> CC: Matthew Garrett <mjg59@google.com> CC: Josh Boyer <jwboyer@redhat.com> CC: Zhenzhong Duan <zhenzhong.duan@oracle.com> CC: Steve Wahl <steve.wahl@hpe.com> CC: Mike Travis <mike.travis@hpe.com> CC: Dimitri Sivanich <dimitri.sivanich@hpe.com> CC: Arnd Bergmann <arnd@arndb.de> CC: "Peter Zijlstra (Intel)" <peterz@infradead.org> CC: Giovanni Gherdovich <ggherdovich@suse.cz> CC: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com> CC: Len Brown <len.brown@intel.com> CC: Kees Cook <keescook@chromium.org> CC: Martin Molnar <martin.molnar.programming@gmail.com> CC: Pingfan Liu <kernelfans@gmail.com> CC: linux-kernel@vger.kernel.org CC: jailhouse-dev@googlegroups.com Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> --- Thomas: I finally found the reference we were discussing in Portland. Sorry this patch took so long.
I don't have any non-integrated APIC hardware to test with. Can anyone help me out? --- arch/x86/include/asm/x86_init.h | 1 - arch/x86/kernel/apic/x2apic_uv_x.c | 1 - arch/x86/kernel/jailhouse.c | 1 - arch/x86/kernel/platform-quirks.c | 1 - arch/x86/kernel/smpboot.c | 21 ++++++++++++--------- 5 files changed, 12 insertions(+), 13 deletions(-)
diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h index 96d9cd208610..006a5d7fd7eb 100644 --- a/arch/x86/include/asm/x86_init.h +++ b/arch/x86/include/asm/x86_init.h @@ -229,7 +229,6 @@ enum x86_legacy_i8042_state { struct x86_legacy_features { enum x86_legacy_i8042_state i8042; int rtc; - int warm_reset; int no_vga; int reserve_bios_regions; struct x86_legacy_devices devices; diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index ad53b2abc859..5afcfd193592 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -343,7 +343,6 @@ static int __init uv_acpi_madt_oem_check(char *_oem_id, char *_oem_table_id) } else if (!strcmp(oem_table_id, "UVH")) { /* Only UV1 systems: */ uv_system_type = UV_NON_UNIQUE_APIC; - x86_platform.legacy.warm_reset = 0; __this_cpu_write(x2apic_extra_bits, pnodeid << uvh_apicid.s.pnode_shift); uv_set_apicid_hibit(); uv_apic = 1; diff --git a/arch/x86/kernel/jailhouse.c b/arch/x86/kernel/jailhouse.c index 6eb8b50ea07e..d628fe92d6af 100644 --- a/arch/x86/kernel/jailhouse.c +++ b/arch/x86/kernel/jailhouse.c @@ -210,7 +210,6 @@ static void __init jailhouse_init_platform(void) x86_platform.calibrate_tsc = jailhouse_get_tsc; x86_platform.get_wallclock = jailhouse_get_wallclock; x86_platform.legacy.rtc = 0; - x86_platform.legacy.warm_reset = 0; x86_platform.legacy.i8042 = X86_LEGACY_I8042_PLATFORM_ABSENT; legacy_pic = &null_legacy_pic; diff --git a/arch/x86/kernel/platform-quirks.c b/arch/x86/kernel/platform-quirks.c index b348a672f71d..d922c5e0c678 100644 --- a/arch/x86/kernel/platform-quirks.c +++ b/arch/x86/kernel/platform-quirks.c @@ -9,7 +9,6 @@ void __init x86_early_init_platform_quirks(void) { x86_platform.legacy.i8042 = X86_LEGACY_I8042_EXPECTED_PRESENT; x86_platform.legacy.rtc = 1; - x86_platform.legacy.warm_reset = 1; x86_platform.legacy.reserve_bios_regions = 0; x86_platform.legacy.devices.pnpbios = 1; diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index d85e91a8aa8c..e2ebb0be2ee3 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -1049,18 +1049,21 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle, * the targeted processor. */ - if (x86_platform.legacy.warm_reset) { + /* + * APs are typically started in one of two ways: + * + * - On 486-era hardware with a non-integrated Local APIC, a single + * INIT IPI is sent. When the AP comes out of reset, the BIOS + * follows the warm reset vector to start_ip. + * - On everything with an integrated Local APIC, the start_ip is + * provided in the Startup IPI message, sent as part of an + * INIT-SIPI-SIPI sequence. + */ + if (!APIC_INTEGRATED(boot_cpu_apic_version)) { pr_debug("Setting warm reset code and vector.\n"); smpboot_setup_warm_reset_vector(start_ip); - /* - * Be paranoid about clearing APIC errors. - */ - if (APIC_INTEGRATED(boot_cpu_apic_version)) { - apic_write(APIC_ESR, 0); - apic_read(APIC_ESR); - } } /* @@ -1118,7 +1121,7 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle, } } - if (x86_platform.legacy.warm_reset) { + if (!APIC_INTEGRATED(boot_cpu_apic_version)) { /* * Cleanup possible dangling ends... */ -- 2.11.0
| |