Re: [PATCH 1/4] sched: dovetail: pass thread-info bits to arch_dovetail_switch_finish()

On Sat, Apr 2, 2022 at 4:23 PM Philippe Gerum via Xenomai wrote: > - arch_dovetail_switch_finish(leave_inband); > + ti_work = READ_ONCE(current_thread_info()->flags); > + arch_dovetail_switch_finish(leave_inband, ti_work); Why are you passing ti_work as parameter? AFAIU you

Re: x86 port i/o broken on recent kernel

Richard Weinberger via Xenomai writes: > Hi! > > The following test works fine on 5.4.180 (ipipe) but fails on 5.15.9 > (dovetail). > > $ ./iotest > Segmentation fault > $ dmesg > [Xenomai] switching iotest to secondary mode after exception #13 from > user-space at 0x400b8d (pid 14574) >

[PATCH 1/4] sched: dovetail: pass thread-info bits to arch_dovetail_switch_finish()

From: Philippe Gerum The out-of-band switch tail code may need the thread-info work bits to reinstate the current context appropriately, pass them to arch_dovetail_switch_finish(). Signed-off-by: Philippe Gerum --- kernel/sched/core.c | 8 ++-- 1 file changed, 6 insertions(+), 2

[PATCH 4/4] x86: dovetail: reinstate I/O bitmap on out-of-band user entry

From: Philippe Gerum We have to fix up the TSS with the proper I/O bitmap settings in arch_dovetail_switch_finish() when the current task is about to re-enter user mode on the out-of-band stage, along with reloading the fpu context if need be. This fixes an application crash observed when a

[PATCH 3/4] arm64: dovetail: fix up arch_dovetail_switch_finish() signature

From: Philippe Gerum Signed-off-by: Philippe Gerum --- arch/arm64/include/asm/dovetail.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arm64/include/asm/dovetail.h b/arch/arm64/include/asm/dovetail.h index 668679399406..f1c6605972cf 100644 ---

[PATCH 2/4] ARM: dovetail: fix up arch_dovetail_switch_finish() signature

From: Philippe Gerum Signed-off-by: Philippe Gerum --- arch/arm/include/asm/dovetail.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arm/include/asm/dovetail.h b/arch/arm/include/asm/dovetail.h index 8f3a09391d06..ff754ff4d2db 100644 ---

[PATCH 2/2] x86: irq_pipeline: resolve conflict among users of HV callback vector

From: Philippe Gerum Dovetail does not cope well with paravirtualization ATM, which is why XEN depends on !IRQ_PIPELINE. For the same reason, there is no point in adding any IRQ trampoline code for XEN_PVHVM. Also, prevent CONFIG_HYPERV from being built if interrupt pipeling is enabled, since

[PATCH 1/2] clocksource/drivers: dovetail: fix dependency on GENERIC_CLOCKSOURCE_VDSO

From: Philippe Gerum GENERIC_CLOCKSOURCE_VDSO is currently available for ARM and ARM64, but some drivers which can export their clocksource via the vDSO can still be built for unsupported architectures like x86. Use a weak reverse dependency logic for vDSO-exported clocksources, enabling

Re: [linux-dovetail PATCH 1/2] x86: fix duplication definication issue in making allyesconfig

Hongzhan Chen via Xenomai writes: > When making allyesconfig, there is duplication definication issue. > To fix it, make them exclusive each other. > > Signed-off-by: Hongzhan Chen > --- > arch/x86/kernel/irq_pipeline.c | 6 ++ > 1 file changed, 2 insertions(+), 4 deletions(-) > > diff

Re: [linux-dovetail PATCH 2/2] clocksource: GENERIC_CLOCKSOURCE_VDSO only supported on ARM

Hongzhan Chen via Xenomai writes: > Currently, GENERIC_CLOCKSOURCE_VDSO only supported well on ARM. > To fix build issue like following when make allyesconfig on x86 platform. > > arch/x86/entry/vdso/../../../../lib/vdso/gettimeofday.c:145:14: > error: implicit declaration of function

rtnet i219 setup

Hi everyone. I was trying to setup Xenomai 3.2 with rtnet, but I was stuck on i219 driver. I have same problem as https://xenomai.org/pipermail/xenomai/2020-February/042422.html. I found i219 patch, but it is for kernel 4.x. In kernel 5.x I have an error same as