lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
Open Source and information security mailing list archives
| ||
|
Date: Fri, 24 Apr 2020 14:25:35 +0200 From: Vitaly Kuznetsov <vkuznets@...hat.com> To: x86@...nel.org Cc: Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>, "H. Peter Anvin" <hpa@...or.com>, linux-kernel@...r.kernel.org Subject: [PATCH] x86/idt: Keep spurious entries unset in system_vectors Commit dc20b2d52653 ("x86/idt: Move interrupt gate initialization to IDT code") had a side-effect of 'set_bit(i, used_vectors)' for unused entries which are being mapped to spurious entries. (user_vectors were later renamed to system_vectors). Previously, we used to count on system_vectors in arch_show_interrupts() to not print unexisting entries in /proc/interrupts. E.g. 'Hypervisor callback interrupts' should not be printed on bare metal. This is currently broken. Setting bits in system_vectors for all unused entries also makes alloc_intr_gate() fail in case someone decides to do it later. It seems this is not currently an issue because all alloc_intr_gate() users are calling it early, before we call idt_setup_apic_and_irq_gates() but this also seems wrong. Signed-off-by: Vitaly Kuznetsov <vkuznets@...hat.com> --- arch/x86/kernel/idt.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/idt.c b/arch/x86/kernel/idt.c index 87ef69a72c52..b62e9d080a3e 100644 --- a/arch/x86/kernel/idt.c +++ b/arch/x86/kernel/idt.c @@ -318,7 +318,13 @@ void __init idt_setup_apic_and_irq_gates(void) #ifdef CONFIG_X86_LOCAL_APIC for_each_clear_bit_from(i, system_vectors, NR_VECTORS) { - set_bit(i, system_vectors); + /* + * Spurious entries are left unset in system_vectors so it can + * be used to check which gates were really allocated. This also + * allows using alloc_intr_gate() and not update_intr_gate() if + * some of the currently-spurious entries are to be allocated + * later. + */ entry = spurious_entries_start + 8 * (i - FIRST_SYSTEM_VECTOR); set_intr_gate(i, entry); } -- 2.25.3
Powered by blists - more mailing lists