| Message ID | 1465923441-107596-1-git-send-email-glider@google.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi all, On Tue, Jun 14, 2016 at 6:57 PM, Alexander Potapenko <glider@google.com> wrote: > Add ARCH_HAS_KCOV to ARM64 config. To avoid crashes, disable > instrumentation of the following files: > > arch/arm64/boot/* > arch/arm64/kvm/hyp/* > > Signed-off-by: Alexander Potapenko <glider@google.com> > --- > v2: - disable instrumentation of arch/arm64/{boot,kvm/hyp} > - enable instrumentation of arch/arm64/lib/delay.c > --- > arch/arm64/Kconfig | 1 + > arch/arm64/boot/Makefile | 4 ++++ > arch/arm64/kvm/hyp/Makefile | 4 ++++ > 3 files changed, 9 insertions(+) > > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig > index 5a0a691..eb0b0a0 100644 > --- a/arch/arm64/Kconfig > +++ b/arch/arm64/Kconfig > @@ -7,6 +7,7 @@ config ARM64 > select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE > select ARCH_HAS_ELF_RANDOMIZE > select ARCH_HAS_GCOV_PROFILE_ALL > + select ARCH_HAS_KCOV > select ARCH_HAS_SG_CHAIN > select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST > select ARCH_USE_CMPXCHG_LOCKREF > diff --git a/arch/arm64/boot/Makefile b/arch/arm64/boot/Makefile > index 305c552..74cec89 100644 > --- a/arch/arm64/boot/Makefile > +++ b/arch/arm64/boot/Makefile > @@ -14,6 +14,10 @@ > # Based on the ia64 boot/Makefile. > # > > +# Avoid potential boot-time problems with kcov instrumentation. We are mostly > +# interested in syscall coverage, so boot code is not interesting anyway. > +KCOV_INSTRUMENT := n > + > targets := Image Image.gz > > $(obj)/Image: vmlinux FORCE > diff --git a/arch/arm64/kvm/hyp/Makefile b/arch/arm64/kvm/hyp/Makefile > index 778d0ef..0c85feb 100644 > --- a/arch/arm64/kvm/hyp/Makefile > +++ b/arch/arm64/kvm/hyp/Makefile > @@ -17,6 +17,10 @@ obj-$(CONFIG_KVM_ARM_HOST) += tlb.o > obj-$(CONFIG_KVM_ARM_HOST) += hyp-entry.o > obj-$(CONFIG_KVM_ARM_HOST) += s2-setup.o > > +# KVM code is run at a different exception code with a different map, so > +# compiler instrumentation that inserts callbacks or checks into the code may > +# cause crashes. Just disable it. > GCOV_PROFILE := n > KASAN_SANITIZE := n > UBSAN_SANITIZE := n > +KCOV_INSTRUMENT := n > -- > 2.8.0.rc3.226.g39d4020 > Following the previous patch discussion, I have dropped the delay.c exception, which isn't strictly required for kcov to function. We can decide what to do with it later. Instead I've added rules to avoid instrumentation of arch/arm64/boot and arch/arm64/kvm
On Tue, Jun 14, 2016 at 06:57:21PM +0200, Alexander Potapenko wrote: > Add ARCH_HAS_KCOV to ARM64 config. To avoid crashes, disable > instrumentation of the following files: > > arch/arm64/boot/* > arch/arm64/kvm/hyp/* > > Signed-off-by: Alexander Potapenko <glider@google.com> > --- > v2: - disable instrumentation of arch/arm64/{boot,kvm/hyp} > - enable instrumentation of arch/arm64/lib/delay.c > --- > arch/arm64/Kconfig | 1 + > arch/arm64/boot/Makefile | 4 ++++ > arch/arm64/kvm/hyp/Makefile | 4 ++++ > 3 files changed, 9 insertions(+) > > diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig > index 5a0a691..eb0b0a0 100644 > --- a/arch/arm64/Kconfig > +++ b/arch/arm64/Kconfig > @@ -7,6 +7,7 @@ config ARM64 > select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE > select ARCH_HAS_ELF_RANDOMIZE > select ARCH_HAS_GCOV_PROFILE_ALL > + select ARCH_HAS_KCOV > select ARCH_HAS_SG_CHAIN > select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST > select ARCH_USE_CMPXCHG_LOCKREF > diff --git a/arch/arm64/boot/Makefile b/arch/arm64/boot/Makefile > index 305c552..74cec89 100644 > --- a/arch/arm64/boot/Makefile > +++ b/arch/arm64/boot/Makefile > @@ -14,6 +14,10 @@ > # Based on the ia64 boot/Makefile. > # > > +# Avoid potential boot-time problems with kcov instrumentation. We are mostly > +# interested in syscall coverage, so boot code is not interesting anyway. > +KCOV_INSTRUMENT := n We have no code under our boot directory, so I don't think the changes to arch/arm64/boot are necessary. > + > targets := Image Image.gz > > $(obj)/Image: vmlinux FORCE > diff --git a/arch/arm64/kvm/hyp/Makefile b/arch/arm64/kvm/hyp/Makefile > index 778d0ef..0c85feb 100644 > --- a/arch/arm64/kvm/hyp/Makefile > +++ b/arch/arm64/kvm/hyp/Makefile > @@ -17,6 +17,10 @@ obj-$(CONFIG_KVM_ARM_HOST) += tlb.o > obj-$(CONFIG_KVM_ARM_HOST) += hyp-entry.o > obj-$(CONFIG_KVM_ARM_HOST) += s2-setup.o > > +# KVM code is run at a different exception code with a different map, so > +# compiler instrumentation that inserts callbacks or checks into the code may > +# cause crashes. Just disable it. > GCOV_PROFILE := n > KASAN_SANITIZE := n > UBSAN_SANITIZE := n > +KCOV_INSTRUMENT := n This looks sane to me. With VHE this code _may_ run in the same memory map as the kernel, but it's not something we can determine at compile time. Otherwise, I believe that the rest of the C code under arch/arm64 runs in the usual kernel memory map (including the special case of kaslr.c), and the EFI stub code has already been covered, so I'm not immediately aware of anything else that needs to be special-cased. I built and booted (via EFI) a kernel with this feature enabled (also with the boot/Makefile change removed). I haven't tested the feature itself as such, as I'm not sure how to do that. FWIW, with the boot/Makefile change removed, feel free to add: Acked-by: Mark Rutland <mark.rutland@arm.com> Thanks, Mark.
On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: > On Tue, Jun 14, 2016 at 06:57:21PM +0200, Alexander Potapenko wrote: >> Add ARCH_HAS_KCOV to ARM64 config. To avoid crashes, disable >> instrumentation of the following files: >> >> arch/arm64/boot/* >> arch/arm64/kvm/hyp/* >> >> Signed-off-by: Alexander Potapenko <glider@google.com> >> --- >> v2: - disable instrumentation of arch/arm64/{boot,kvm/hyp} >> - enable instrumentation of arch/arm64/lib/delay.c >> --- >> arch/arm64/Kconfig | 1 + >> arch/arm64/boot/Makefile | 4 ++++ >> arch/arm64/kvm/hyp/Makefile | 4 ++++ >> 3 files changed, 9 insertions(+) >> >> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig >> index 5a0a691..eb0b0a0 100644 >> --- a/arch/arm64/Kconfig >> +++ b/arch/arm64/Kconfig >> @@ -7,6 +7,7 @@ config ARM64 >> select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE >> select ARCH_HAS_ELF_RANDOMIZE >> select ARCH_HAS_GCOV_PROFILE_ALL >> + select ARCH_HAS_KCOV >> select ARCH_HAS_SG_CHAIN >> select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST >> select ARCH_USE_CMPXCHG_LOCKREF >> diff --git a/arch/arm64/boot/Makefile b/arch/arm64/boot/Makefile >> index 305c552..74cec89 100644 >> --- a/arch/arm64/boot/Makefile >> +++ b/arch/arm64/boot/Makefile >> @@ -14,6 +14,10 @@ >> # Based on the ia64 boot/Makefile. >> # >> >> +# Avoid potential boot-time problems with kcov instrumentation. We are mostly >> +# interested in syscall coverage, so boot code is not interesting anyway. >> +KCOV_INSTRUMENT := n > > We have no code under our boot directory, so I don't think the changes > to arch/arm64/boot are necessary. Indeed we don't! Removed that. >> + >> targets := Image Image.gz >> >> $(obj)/Image: vmlinux FORCE >> diff --git a/arch/arm64/kvm/hyp/Makefile b/arch/arm64/kvm/hyp/Makefile >> index 778d0ef..0c85feb 100644 >> --- a/arch/arm64/kvm/hyp/Makefile >> +++ b/arch/arm64/kvm/hyp/Makefile >> @@ -17,6 +17,10 @@ obj-$(CONFIG_KVM_ARM_HOST) += tlb.o >> obj-$(CONFIG_KVM_ARM_HOST) += hyp-entry.o >> obj-$(CONFIG_KVM_ARM_HOST) += s2-setup.o >> >> +# KVM code is run at a different exception code with a different map, so >> +# compiler instrumentation that inserts callbacks or checks into the code may >> +# cause crashes. Just disable it. >> GCOV_PROFILE := n >> KASAN_SANITIZE := n >> UBSAN_SANITIZE := n >> +KCOV_INSTRUMENT := n > > This looks sane to me. > > With VHE this code _may_ run in the same memory map as the kernel, but > it's not something we can determine at compile time. > > Otherwise, I believe that the rest of the C code under arch/arm64 runs > in the usual kernel memory map (including the special case of kaslr.c), > and the EFI stub code has already been covered, so I'm not immediately > aware of anything else that needs to be special-cased. > > I built and booted (via EFI) a kernel with this feature enabled (also > with the boot/Makefile change removed). I haven't tested the feature > itself as such, as I'm not sure how to do that. You can test it by running the test program from Documentation/kcov.txt. > FWIW, with the boot/Makefile change removed, feel free to add: > > Acked-by: Mark Rutland <mark.rutland@arm.com> Thank you! I'll wait till tomorrow for others to comment, and then will send the updated version. > Thanks, > Mark.
On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: > On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: > > I built and booted (via EFI) a kernel with this feature enabled (also > > with the boot/Makefile change removed). I haven't tested the feature > > itself as such, as I'm not sure how to do that. > You can test it by running the test program from Documentation/kcov.txt. Ah, I hadn't spotted that. If I get the chance I'll try to give that a go. > > FWIW, with the boot/Makefile change removed, feel free to add: > > > > Acked-by: Mark Rutland <mark.rutland@arm.com> > Thank you! > I'll wait till tomorrow for others to comment, and then will send the > updated version. Sounds good to me. Mark.
On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: > On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: > > On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: > > > I built and booted (via EFI) a kernel with this feature enabled (also > > > with the boot/Makefile change removed). I haven't tested the feature > > > itself as such, as I'm not sure how to do that. > > You can test it by running the test program from Documentation/kcov.txt. > > Ah, I hadn't spotted that. If I get the chance I'll try to give that a > go. I just had a go (with this applied atop of v4.7-rc3), and I get: root@ribbensteg:/home/nanook# ./kcov mmap: No such device The device exists (it was able to open the fd, evidently): root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov Strace show me: openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) It doesn't look like the error paths in kcov_mmap are hitting. Any ideas? Thanks, Mark.
On Wed, Jun 15, 2016 at 1:44 PM, Mark Rutland <mark.rutland@arm.com> wrote: > On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: >> On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: >> > On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: >> > > I built and booted (via EFI) a kernel with this feature enabled (also >> > > with the boot/Makefile change removed). I haven't tested the feature >> > > itself as such, as I'm not sure how to do that. >> > You can test it by running the test program from Documentation/kcov.txt. >> >> Ah, I hadn't spotted that. If I get the chance I'll try to give that a >> go. > > I just had a go (with this applied atop of v4.7-rc3), and I get: > > root@ribbensteg:/home/nanook# ./kcov > mmap: No such device > > The device exists (it was able to open the fd, evidently): > > root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov > -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov > > Strace show me: > > openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 > ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 > mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) > > It doesn't look like the error paths in kcov_mmap are hitting. > > Any ideas? According to Dmitry (thanks, Dmitry!) this has regressed recently, but there's a pending patch that should probably fix the problem: http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html > Thanks, > Mark.
On Wed, Jun 15, 2016 at 01:53:03PM +0200, Alexander Potapenko wrote: > On Wed, Jun 15, 2016 at 1:44 PM, Mark Rutland <mark.rutland@arm.com> wrote: > > On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: > >> On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: > >> > On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: > >> > > I built and booted (via EFI) a kernel with this feature enabled (also > >> > > with the boot/Makefile change removed). I haven't tested the feature > >> > > itself as such, as I'm not sure how to do that. > >> > You can test it by running the test program from Documentation/kcov.txt. > >> > >> Ah, I hadn't spotted that. If I get the chance I'll try to give that a > >> go. > > > > I just had a go (with this applied atop of v4.7-rc3), and I get: > > > > root@ribbensteg:/home/nanook# ./kcov > > mmap: No such device > > > > The device exists (it was able to open the fd, evidently): > > > > root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov > > -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov > > > > Strace show me: > > > > openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 > > ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 > > mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) > > > > It doesn't look like the error paths in kcov_mmap are hitting. > > > > Any ideas? > According to Dmitry (thanks, Dmitry!) this has regressed recently, but > there's a pending patch that should probably fix the problem: > http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html Thanks for the pointer! With that applied, the program runs. However, it looks like I missed a warning from the kernel build system, and my toolchain doesn't actually support -fsanitize-coverage=trace-pc, so I'm not going to be able to test that further. It would be great if we could deliberately not register the debugfs file when there was no compiler support for the feature, for those like me who miss the build time warning. We do something like that for the LSE atomics on arm64. Thanks, Mark.
On Wed, Jun 15, 2016 at 4:25 PM, Mark Rutland <mark.rutland@arm.com> wrote: >> >> > > I built and booted (via EFI) a kernel with this feature enabled (also >> >> > > with the boot/Makefile change removed). I haven't tested the feature >> >> > > itself as such, as I'm not sure how to do that. >> >> > You can test it by running the test program from Documentation/kcov.txt. >> >> >> >> Ah, I hadn't spotted that. If I get the chance I'll try to give that a >> >> go. >> > >> > I just had a go (with this applied atop of v4.7-rc3), and I get: >> > >> > root@ribbensteg:/home/nanook# ./kcov >> > mmap: No such device >> > >> > The device exists (it was able to open the fd, evidently): >> > >> > root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov >> > -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov >> > >> > Strace show me: >> > >> > openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 >> > ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 >> > mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) >> > >> > It doesn't look like the error paths in kcov_mmap are hitting. >> > >> > Any ideas? >> According to Dmitry (thanks, Dmitry!) this has regressed recently, but >> there's a pending patch that should probably fix the problem: >> http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html > > Thanks for the pointer! With that applied, the program runs. > > However, it looks like I missed a warning from the kernel build system, > and my toolchain doesn't actually support -fsanitize-coverage=trace-pc, > so I'm not going to be able to test that further. > > It would be great if we could deliberately not register the debugfs file > when there was no compiler support for the feature, for those like me > who miss the build time warning. We do something like that for the LSE > atomics on arm64. Hi Mark, It's a common problem and it would be great to detect this. But I think it's better to return ENOTSUP from open rather than not registering the file at all. This way higher level tools will be able to more easily diagnose the issue and properly report to user. A missing file looks like not mounted debugfs (which another common issue). I am not sure how to do it. Compiler does not provide any define for this option. And I am not familiar enough with kernel makefiles. Would it be possible to add a define to CLAGS in the makefile along with printing the warning?
On 15/06/16 15:25, Mark Rutland wrote: > On Wed, Jun 15, 2016 at 01:53:03PM +0200, Alexander Potapenko wrote: >> On Wed, Jun 15, 2016 at 1:44 PM, Mark Rutland <mark.rutland@arm.com> wrote: >>> On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: >>>> On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: >>>>> On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: >>>>>> I built and booted (via EFI) a kernel with this feature enabled (also >>>>>> with the boot/Makefile change removed). I haven't tested the feature >>>>>> itself as such, as I'm not sure how to do that. >>>>> You can test it by running the test program from Documentation/kcov.txt. >>>> >>>> Ah, I hadn't spotted that. If I get the chance I'll try to give that a >>>> go. >>> >>> I just had a go (with this applied atop of v4.7-rc3), and I get: >>> >>> root@ribbensteg:/home/nanook# ./kcov >>> mmap: No such device >>> >>> The device exists (it was able to open the fd, evidently): >>> >>> root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov >>> -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov >>> >>> Strace show me: >>> >>> openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 >>> ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 >>> mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) >>> >>> It doesn't look like the error paths in kcov_mmap are hitting. >>> >>> Any ideas? >> According to Dmitry (thanks, Dmitry!) this has regressed recently, but >> there's a pending patch that should probably fix the problem: >> http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html > > Thanks for the pointer! With that applied, the program runs. > > However, it looks like I missed a warning from the kernel build system, > and my toolchain doesn't actually support -fsanitize-coverage=trace-pc, > so I'm not going to be able to test that further. I dusted off a compiler that supports this, and ran the sample program under Documentation with the above unproxify patch. Tested-by: James Morse <james.morse@arm.com> Thanks, James
On Thu, Jun 16, 2016 at 12:47 PM, James Morse <james.morse@arm.com> wrote: > On 15/06/16 15:25, Mark Rutland wrote: >> On Wed, Jun 15, 2016 at 01:53:03PM +0200, Alexander Potapenko wrote: >>> On Wed, Jun 15, 2016 at 1:44 PM, Mark Rutland <mark.rutland@arm.com> wrote: >>>> On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: >>>>> On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: >>>>>> On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: >>>>>>> I built and booted (via EFI) a kernel with this feature enabled (also >>>>>>> with the boot/Makefile change removed). I haven't tested the feature >>>>>>> itself as such, as I'm not sure how to do that. >>>>>> You can test it by running the test program from Documentation/kcov.txt. >>>>> >>>>> Ah, I hadn't spotted that. If I get the chance I'll try to give that a >>>>> go. >>>> >>>> I just had a go (with this applied atop of v4.7-rc3), and I get: >>>> >>>> root@ribbensteg:/home/nanook# ./kcov >>>> mmap: No such device >>>> >>>> The device exists (it was able to open the fd, evidently): >>>> >>>> root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov >>>> -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov >>>> >>>> Strace show me: >>>> >>>> openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 >>>> ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 >>>> mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) >>>> >>>> It doesn't look like the error paths in kcov_mmap are hitting. >>>> >>>> Any ideas? >>> According to Dmitry (thanks, Dmitry!) this has regressed recently, but >>> there's a pending patch that should probably fix the problem: >>> http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html >> >> Thanks for the pointer! With that applied, the program runs. >> >> However, it looks like I missed a warning from the kernel build system, >> and my toolchain doesn't actually support -fsanitize-coverage=trace-pc, >> so I'm not going to be able to test that further. > > I dusted off a compiler that supports this, and ran the sample program under > Documentation with the above unproxify patch. > > Tested-by: James Morse <james.morse@arm.com> I think it's time to ask now :) If I receive "Tested-by" or "Acked-by" responses, do I need to send out a patch adding them, or should I rely on the maintainer taking the patch to the tree? The first option reduces the amount of work done by the maintainer, while the second one reduces the traffic in the list. Sorry, I couldn't find the answer in the manuals. Thank you, Alex > > Thanks, > > James
On Thu, Jun 16, 2016 at 05:20:03PM +0200, Alexander Potapenko wrote: > On Thu, Jun 16, 2016 at 12:47 PM, James Morse <james.morse@arm.com> wrote: > > On 15/06/16 15:25, Mark Rutland wrote: > >> On Wed, Jun 15, 2016 at 01:53:03PM +0200, Alexander Potapenko wrote: > >>> On Wed, Jun 15, 2016 at 1:44 PM, Mark Rutland <mark.rutland@arm.com> wrote: > >>>> On Wed, Jun 15, 2016 at 10:25:10AM +0100, Mark Rutland wrote: > >>>>> On Tue, Jun 14, 2016 at 08:16:08PM +0200, Alexander Potapenko wrote: > >>>>>> On Tue, Jun 14, 2016 at 7:55 PM, Mark Rutland <mark.rutland@arm.com> wrote: > >>>>>>> I built and booted (via EFI) a kernel with this feature enabled (also > >>>>>>> with the boot/Makefile change removed). I haven't tested the feature > >>>>>>> itself as such, as I'm not sure how to do that. > >>>>>> You can test it by running the test program from Documentation/kcov.txt. > >>>>> > >>>>> Ah, I hadn't spotted that. If I get the chance I'll try to give that a > >>>>> go. > >>>> > >>>> I just had a go (with this applied atop of v4.7-rc3), and I get: > >>>> > >>>> root@ribbensteg:/home/nanook# ./kcov > >>>> mmap: No such device > >>>> > >>>> The device exists (it was able to open the fd, evidently): > >>>> > >>>> root@ribbensteg:/home/nanook# ls -al /sys/kernel/debug/kcov > >>>> -rw------- 1 root root 0 Jan 1 1970 /sys/kernel/debug/kcov > >>>> > >>>> Strace show me: > >>>> > >>>> openat(AT_FDCWD, "/sys/kernel/debug/kcov", O_RDWR) = 3 > >>>> ioctl(3, CHIOMOVE or CM_IOCGATR, 0x10000) = 0 > >>>> mmap(NULL, 524288, PROT_READ|PROT_WRITE, MAP_SHARED, 3, 0) = -1 ENODEV (No such device) > >>>> > >>>> It doesn't look like the error paths in kcov_mmap are hitting. > >>>> > >>>> Any ideas? > >>> According to Dmitry (thanks, Dmitry!) this has regressed recently, but > >>> there's a pending patch that should probably fix the problem: > >>> http://lkml.iu.edu/hypermail/linux/kernel/1605.2/04379.html > >> > >> Thanks for the pointer! With that applied, the program runs. > >> > >> However, it looks like I missed a warning from the kernel build system, > >> and my toolchain doesn't actually support -fsanitize-coverage=trace-pc, > >> so I'm not going to be able to test that further. > > > > I dusted off a compiler that supports this, and ran the sample program under > > Documentation with the above unproxify patch. > > > > Tested-by: James Morse <james.morse@arm.com> > I think it's time to ask now :) > If I receive "Tested-by" or "Acked-by" responses, do I need to send > out a patch adding them, or should I rely on the maintainer taking the > patch to the tree? > The first option reduces the amount of work done by the maintainer, > while the second one reduces the traffic in the list. > Sorry, I couldn't find the answer in the manuals. It's up to the maintainer, so it varies. The best thing to do is to ask the maintainer what they'd prefer. From my experience, Catalin is usually happy to add tags, so I suspect he'd be happy to do so for this patch (assuming he's happy to pick it up). I'll leave it for him to say either way. Thanks, Mark.
On Thu, Jun 16, 2016 at 04:44:12PM +0100, Mark Rutland wrote: > On Thu, Jun 16, 2016 at 05:20:03PM +0200, Alexander Potapenko wrote: > > I think it's time to ask now :) > > If I receive "Tested-by" or "Acked-by" responses, do I need to send > > out a patch adding them, or should I rely on the maintainer taking the > > patch to the tree? > > The first option reduces the amount of work done by the maintainer, > > while the second one reduces the traffic in the list. > > Sorry, I couldn't find the answer in the manuals. > > It's up to the maintainer, so it varies. The best thing to do is to ask > the maintainer what they'd prefer. > > From my experience, Catalin is usually happy to add tags, so I suspect > he'd be happy to do so for this patch (assuming he's happy to pick it > up). I'll leave it for him to say either way. I usually cherry-pick tags that I see in reply to the *latest* version of the patch (I have a rudimentary script to do this). I noticed that there was an ack on v1 form Marc Z that's missing in v2. Maybe it no longer applies, I can't tell, but I usually expect subsequent versions of a patch to include all the previously given acks (of course, if they still apply, sometimes a patch rewrite means dropping those tags).
On Thu, Jun 16, 2016 at 05:25:31PM +0100, Catalin Marinas wrote: > On Thu, Jun 16, 2016 at 04:44:12PM +0100, Mark Rutland wrote: > > On Thu, Jun 16, 2016 at 05:20:03PM +0200, Alexander Potapenko wrote: > > > I think it's time to ask now :) > > > If I receive "Tested-by" or "Acked-by" responses, do I need to send > > > out a patch adding them, or should I rely on the maintainer taking the > > > patch to the tree? > > > The first option reduces the amount of work done by the maintainer, > > > while the second one reduces the traffic in the list. > > > Sorry, I couldn't find the answer in the manuals. > > > > It's up to the maintainer, so it varies. The best thing to do is to ask > > the maintainer what they'd prefer. > > > > From my experience, Catalin is usually happy to add tags, so I suspect > > he'd be happy to do so for this patch (assuming he's happy to pick it > > up). I'll leave it for him to say either way. > > I usually cherry-pick tags that I see in reply to the *latest* version > of the patch (I have a rudimentary script to do this). > > I noticed that there was an ack on v1 form Marc Z that's missing in v2. I believe Marc's reply [1] was to v3 [2], it's just that the version was missing form the subject, and discussions continued on v2 in the mean time. > Maybe it no longer applies, I can't tell, but I usually expect > subsequent versions of a patch to include all the previously given acks > (of course, if they still apply, sometimes a patch rewrite means > dropping those tags). I guess the simplest thing to do is for Alexander to send a v4 with the tags accumulated, assuming James's Tested-by is applicable to v3 with the boot/Makefile hunk removed. James? My ack stands. Thanks, Mark. [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-June/436551.html [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-June/436512.html
On Thu, Jun 16, 2016 at 6:32 PM, Mark Rutland <mark.rutland@arm.com> wrote: > On Thu, Jun 16, 2016 at 05:25:31PM +0100, Catalin Marinas wrote: >> On Thu, Jun 16, 2016 at 04:44:12PM +0100, Mark Rutland wrote: >> > On Thu, Jun 16, 2016 at 05:20:03PM +0200, Alexander Potapenko wrote: >> > > I think it's time to ask now :) >> > > If I receive "Tested-by" or "Acked-by" responses, do I need to send >> > > out a patch adding them, or should I rely on the maintainer taking the >> > > patch to the tree? >> > > The first option reduces the amount of work done by the maintainer, >> > > while the second one reduces the traffic in the list. >> > > Sorry, I couldn't find the answer in the manuals. >> > >> > It's up to the maintainer, so it varies. The best thing to do is to ask >> > the maintainer what they'd prefer. >> > >> > From my experience, Catalin is usually happy to add tags, so I suspect >> > he'd be happy to do so for this patch (assuming he's happy to pick it >> > up). I'll leave it for him to say either way. >> >> I usually cherry-pick tags that I see in reply to the *latest* version >> of the patch (I have a rudimentary script to do this). >> >> I noticed that there was an ack on v1 form Marc Z that's missing in v2. > > I believe Marc's reply [1] was to v3 [2], it's just that the version was > missing form the subject, and discussions continued on v2 in the mean > time. Yes, this is correct. >> Maybe it no longer applies, I can't tell, but I usually expect >> subsequent versions of a patch to include all the previously given acks >> (of course, if they still apply, sometimes a patch rewrite means >> dropping those tags). > > I guess the simplest thing to do is for Alexander to send a v4 with the > tags accumulated, assuming James's Tested-by is applicable to v3 with > the boot/Makefile hunk removed. James? I think it's safe to assume James's Tested-by is still valid, as boot/Makefile hunk did virtually nothing. I'll send the new patch version now. > My ack stands. > > Thanks, > Mark. > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-June/436551.html > [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-June/436512.html
On 16/06/16 17:36, Alexander Potapenko wrote: > On Thu, Jun 16, 2016 at 6:32 PM, Mark Rutland <mark.rutland@arm.com> wrote: >> On Thu, Jun 16, 2016 at 05:25:31PM +0100, Catalin Marinas wrote: >>> I noticed that there was an ack on v1 form Marc Z that's missing in v2. >> >> I believe Marc's reply [1] was to v3 [2], it's just that the version was >> missing form the subject, and discussions continued on v2 in the mean >> time. Ah, that's what happened... > Yes, this is correct. >>> Maybe it no longer applies, I can't tell, but I usually expect >>> subsequent versions of a patch to include all the previously given acks >>> (of course, if they still apply, sometimes a patch rewrite means >>> dropping those tags). >> >> I guess the simplest thing to do is for Alexander to send a v4 with the >> tags accumulated, assuming James's Tested-by is applicable to v3 with >> the boot/Makefile hunk removed. James? > I think it's safe to assume James's Tested-by is still valid, as > boot/Makefile hunk did virtually nothing. I agree! > I'll send the new patch version now. Thanks, James
On Thu, Jun 16, 2016 at 05:32:59PM +0100, Mark Rutland wrote: > On Thu, Jun 16, 2016 at 05:25:31PM +0100, Catalin Marinas wrote: > > On Thu, Jun 16, 2016 at 04:44:12PM +0100, Mark Rutland wrote: > > > On Thu, Jun 16, 2016 at 05:20:03PM +0200, Alexander Potapenko wrote: > > > > I think it's time to ask now :) > > > > If I receive "Tested-by" or "Acked-by" responses, do I need to send > > > > out a patch adding them, or should I rely on the maintainer taking the > > > > patch to the tree? > > > > The first option reduces the amount of work done by the maintainer, > > > > while the second one reduces the traffic in the list. > > > > Sorry, I couldn't find the answer in the manuals. > > > > > > It's up to the maintainer, so it varies. The best thing to do is to ask > > > the maintainer what they'd prefer. > > > > > > From my experience, Catalin is usually happy to add tags, so I suspect > > > he'd be happy to do so for this patch (assuming he's happy to pick it > > > up). I'll leave it for him to say either way. > > > > I usually cherry-pick tags that I see in reply to the *latest* version > > of the patch (I have a rudimentary script to do this). > > > > I noticed that there was an ack on v1 form Marc Z that's missing in v2. > > I believe Marc's reply [1] was to v3 [2], it's just that the version was > missing form the subject, and discussions continued on v2 in the mean > time. OK, thanks for clarification. I thought the one with most replies was the latest ;). > > Maybe it no longer applies, I can't tell, but I usually expect > > subsequent versions of a patch to include all the previously given acks > > (of course, if they still apply, sometimes a patch rewrite means > > dropping those tags). > > I guess the simplest thing to do is for Alexander to send a v4 with the > tags accumulated, assuming James's Tested-by is applicable to v3 with > the boot/Makefile hunk removed. James? Even better.
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 5a0a691..eb0b0a0 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -7,6 +7,7 @@ config ARM64 select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE select ARCH_HAS_ELF_RANDOMIZE select ARCH_HAS_GCOV_PROFILE_ALL + select ARCH_HAS_KCOV select ARCH_HAS_SG_CHAIN select ARCH_HAS_TICK_BROADCAST if GENERIC_CLOCKEVENTS_BROADCAST select ARCH_USE_CMPXCHG_LOCKREF diff --git a/arch/arm64/boot/Makefile b/arch/arm64/boot/Makefile index 305c552..74cec89 100644 --- a/arch/arm64/boot/Makefile +++ b/arch/arm64/boot/Makefile @@ -14,6 +14,10 @@ # Based on the ia64 boot/Makefile. # +# Avoid potential boot-time problems with kcov instrumentation. We are mostly +# interested in syscall coverage, so boot code is not interesting anyway. +KCOV_INSTRUMENT := n + targets := Image Image.gz $(obj)/Image: vmlinux FORCE diff --git a/arch/arm64/kvm/hyp/Makefile b/arch/arm64/kvm/hyp/Makefile index 778d0ef..0c85feb 100644 --- a/arch/arm64/kvm/hyp/Makefile +++ b/arch/arm64/kvm/hyp/Makefile @@ -17,6 +17,10 @@ obj-$(CONFIG_KVM_ARM_HOST) += tlb.o obj-$(CONFIG_KVM_ARM_HOST) += hyp-entry.o obj-$(CONFIG_KVM_ARM_HOST) += s2-setup.o +# KVM code is run at a different exception code with a different map, so +# compiler instrumentation that inserts callbacks or checks into the code may +# cause crashes. Just disable it. GCOV_PROFILE := n KASAN_SANITIZE := n UBSAN_SANITIZE := n +KCOV_INSTRUMENT := n
Add ARCH_HAS_KCOV to ARM64 config. To avoid crashes, disable instrumentation of the following files: arch/arm64/boot/* arch/arm64/kvm/hyp/* Signed-off-by: Alexander Potapenko <glider@google.com> --- v2: - disable instrumentation of arch/arm64/{boot,kvm/hyp} - enable instrumentation of arch/arm64/lib/delay.c --- arch/arm64/Kconfig | 1 + arch/arm64/boot/Makefile | 4 ++++ arch/arm64/kvm/hyp/Makefile | 4 ++++ 3 files changed, 9 insertions(+)