Searched hist:"185 f7e0a9d95d7fc3ed4e5e6f452a41c31d75590" (Results 1 – 3 of 3) sorted by relevance
/freebsd/sys/amd64/amd64/ |
H A D | efirt_support.S | diff 185f7e0a9d95d7fc3ed4e5e6f452a41c31d75590 Mon Jun 03 17:32:42 CEST 2019 Konstantin Belousov <kib@FreeBSD.org> amd64 ef_rt_arch_call: Preserve %rflags around call into EFI RT service.
If service code faulted, we might end up unwinding with interrupts disabled. Top-level kernel code should have interrupts enabled, which is enforced by checks.
Save %rflags before entering EFI, and restore to the known good value on return. This handles situation with disabled interrupts on fault and perhaps other potential bugs, e.g. invalid value for PSL_D.
Reported and tested by: Jan Martin Mikkelsen <janm@transactionware.com> Sponsored by: The FreeBSD Foundation MFC after: 1 week
|
H A D | genassym.c | diff 185f7e0a9d95d7fc3ed4e5e6f452a41c31d75590 Mon Jun 03 17:32:42 CEST 2019 Konstantin Belousov <kib@FreeBSD.org> amd64 ef_rt_arch_call: Preserve %rflags around call into EFI RT service.
If service code faulted, we might end up unwinding with interrupts disabled. Top-level kernel code should have interrupts enabled, which is enforced by checks.
Save %rflags before entering EFI, and restore to the known good value on return. This handles situation with disabled interrupts on fault and perhaps other potential bugs, e.g. invalid value for PSL_D.
Reported and tested by: Jan Martin Mikkelsen <janm@transactionware.com> Sponsored by: The FreeBSD Foundation MFC after: 1 week
|
/freebsd/sys/amd64/include/ |
H A D | efi.h | diff 185f7e0a9d95d7fc3ed4e5e6f452a41c31d75590 Mon Jun 03 17:32:42 CEST 2019 Konstantin Belousov <kib@FreeBSD.org> amd64 ef_rt_arch_call: Preserve %rflags around call into EFI RT service.
If service code faulted, we might end up unwinding with interrupts disabled. Top-level kernel code should have interrupts enabled, which is enforced by checks.
Save %rflags before entering EFI, and restore to the known good value on return. This handles situation with disabled interrupts on fault and perhaps other potential bugs, e.g. invalid value for PSL_D.
Reported and tested by: Jan Martin Mikkelsen <janm@transactionware.com> Sponsored by: The FreeBSD Foundation MFC after: 1 week
|