| Message ID | 1461591994-14918-2-git-send-email-mark.rutland@arm.com |
|---|---|
| State | New |
| Headers | show |
On 25 April 2016 at 16:15, Matt Fleming <matt@codeblueprint.co.uk> wrote: > On Mon, 25 Apr, at 03:12:01PM, Robin Murphy wrote: >> >+static void efi_call_virt_check_flags(unsigned long flags, const char *call) >> >+{ >> >+ unsigned long cur_flags; >> >+ bool mismatch; >> >+ >> >+ local_save_flags(cur_flags); >> >+ >> >+ mismatch = !!((cur_flags ^ flags) & ARCH_EFI_IRQ_FLAGS_MASK); >> >> nit: the assignment itself is already a conversion to bool, so the >> excitement is redundant here. > > This was intentional. I asked Mark to make this change so that it's > explicit for the developer that we're performing the type conversion. But replacing an implicit boolean cast with an explicit one makes little sense, no? Don't we simply want '!= 0' here if you need a boolean expression?
On Mon, Apr 25, 2016 at 03:24:35PM +0100, Matt Fleming wrote: > On Mon, 25 Apr, at 04:18:41PM, Ard Biesheuvel wrote: > > On 25 April 2016 at 16:15, Matt Fleming <matt@codeblueprint.co.uk> wrote: > > > On Mon, 25 Apr, at 03:12:01PM, Robin Murphy wrote: > > >> >+static void efi_call_virt_check_flags(unsigned long flags, const char *call) > > >> >+{ > > >> >+ unsigned long cur_flags; > > >> >+ bool mismatch; > > >> >+ > > >> >+ local_save_flags(cur_flags); > > >> >+ > > >> >+ mismatch = !!((cur_flags ^ flags) & ARCH_EFI_IRQ_FLAGS_MASK); > > >> > > >> nit: the assignment itself is already a conversion to bool, so the > > >> excitement is redundant here. > > > > > > This was intentional. I asked Mark to make this change so that it's > > > explicit for the developer that we're performing the type conversion. > > > > But replacing an implicit boolean cast with an explicit one makes > > little sense, no? Don't we simply want '!= 0' here if you need a > > boolean expression? > > Aha but '!!' is fewer characters to type!! > > I'm not that bothered as long as we don't stuff an int into a bool > without giving the programmer some idea we're doing that. It's not > about the compiler getting it wrong, more about a developer > introducing a bug when they change the code in the future. > > Unless anyone objects, I'll fix this up to use '!= 0' when I apply it. I have no strong preference so long as the code is correct. Another option is to get rid of the bool entirely: flags ^= cur_flags; if (!WARN_ON(flags & ARCH_EFI_IRQ_FLAGS_MASK)) return; Mark.
On Mon, Apr 25, 2016 at 04:59:22PM +0100, Matt Fleming wrote: > On Mon, 25 Apr, at 03:27:35PM, Mark Rutland wrote: > > > > I have no strong preference so long as the code is correct. > > > > Another option is to get rid of the bool entirely: > > > > flags ^= cur_flags; > > if (!WARN_ON(flags & ARCH_EFI_IRQ_FLAGS_MASK)) > > return; > > OK, let's do the following because we need flags to be preserved for > printing, > > --- > > unsigned long cur_flags, mismatch; > > local_save_flags(cur_flags); > > mismatch = flags ^ cur_flags; > if (!WARN_ON_ONCE(mismatch & ARCH_EFI_IRQ_FLAGS_MASK)) > return; > Sure; that looks good to me. Cheers, Mark.
diff --git a/drivers/firmware/efi/runtime-wrappers.c b/drivers/firmware/efi/runtime-wrappers.c index a2c8e70..1f0277e 100644 --- a/drivers/firmware/efi/runtime-wrappers.c +++ b/drivers/firmware/efi/runtime-wrappers.c @@ -16,23 +16,55 @@ #include <linux/bug.h> #include <linux/efi.h> +#include <linux/irqflags.h> #include <linux/mutex.h> #include <linux/spinlock.h> +#include <linux/stringify.h> #include <asm/efi.h> +/* + * Temporary scaffolding until all users provide ARCH_EFI_IRQ_FLAGS_MASK. + */ +#ifdef ARCH_EFI_IRQ_FLAGS_MASK +static void efi_call_virt_check_flags(unsigned long flags, const char *call) +{ + unsigned long cur_flags; + bool mismatch; + + local_save_flags(cur_flags); + + mismatch = !!((cur_flags ^ flags) & ARCH_EFI_IRQ_FLAGS_MASK); + if (!WARN_ON_ONCE(mismatch)) + return; + + add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_NOW_UNRELIABLE); + pr_err_ratelimited(FW_BUG "IRQ flags corrupted (0x%08lx=>0x%08lx) by EFI %s\n", + flags, cur_flags, call); + local_irq_restore(flags); +} +#else /* ARCH_EFI_IRQ_FLAGS_MASK */ +static inline void efi_call_virt_check_flags(unsigned long flags, const char *call) {} +#endif /* ARCH_EFI_IRQ_FLAGS_MASK */ + #define efi_call_virt(f, args...) \ ({ \ efi_status_t __s; \ + unsigned long flags; \ arch_efi_call_virt_setup(); \ + local_save_flags(flags); \ __s = arch_efi_call_virt(f, args); \ + efi_call_virt_check_flags(flags, __stringify(f)); \ arch_efi_call_virt_teardown(); \ __s; \ }) #define __efi_call_virt(f, args...) \ ({ \ + unsigned long flags; \ arch_efi_call_virt_setup(); \ + local_save_flags(flags); \ arch_efi_call_virt(f, args); \ + efi_call_virt_check_flags(flags, __stringify(f)); \ arch_efi_call_virt_teardown(); \ })
The UEFI spec allows runtime services to be called with interrupts masked or unmasked, and if a runtime service function needs to mask interrupts, it must restore the mask to its original state before returning (i.e. from the PoV of the OS, this does not change across a call). Firmware should never unmask exceptions, as these may then be taken by the OS unexpectedly. Unfortunately, some firmware has been seen to unmask IRQs (and potentially other maskable exceptions) across runtime services calls, leaving irq flags corrupted after returning from a runtime services function call. This may be detected by the IRQ tracing code, but often goes unnoticed, leaving a potentially disastrous bug hidden. This patch detects when the irq flags are corrupted by an EFI runtime services call, logging the call and specific corruption to the console. While restoring the expected value of the flags is insufficient to avoid problems, we do so to avoid redundant warnings from elsewhere (e.g. IRQ tracing). The set of bits in flags which we want to check is architecture-specific (e.g. we want to check FIQ on arm64, but not the zero flag on x86), so each arch must provide ARCH_EFI_IRQ_FLAGS_MASK to describe those. In the absence of this mask, the check is a no-op, and we redundantly save the flags twice, but that will be short-lived as subsequent patches will implement this and remove the scaffolding. Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org> Cc: Matt Fleming <matt@codeblueprint.co.uk> Cc: linux-efi@vger.kernel.org --- drivers/firmware/efi/runtime-wrappers.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) -- 1.9.1