search for: pv_altinstruct

...h/x86/include/asm/paravirt-asm.h index 8bdd50ee4bf3..a8139ea27cc1 100644 --- a/arch/x86/include/asm/paravirt-asm.h +++ b/arch/x86/include/asm/paravirt-asm.h @@ -21,6 +21,16 @@ .short clobbers; \ .popsection +#define PV_ALT_SITE(oldinstr, newinstr, ops, off, clobbers) \ + __ALTERNATIVE(.pv_altinstructions, oldinstr, newinstr, \ + X86_FEATURE_PV_OPS); \ + .pushsection .parainstructions, "a"; \ + _ASM_ALIGN; \ + _ASM_PTR 140b; \ + .byte PV_TYPE(ops, off); \ + .byte 142b-140b; \ + .short clobbers; \ + .popsection #define COND_PUSH(set, mask, reg)...

...tive > displacement, I was trying to say that pv_irq_ops always live further down) I believe the problem is this: #define PV_INDIRECT(addr) *addr(%rip) The displacement that the linker computes will be relative to the where this instruction is placed at the time of linking, which is in .pv_altinstructions (and not .text). So when we copy it into .text the displacement becomes bogus. Replacing the macro with #define PV_INDIRECT(addr) *addr // well, it's not so much indirect anymore makes things work. Or maybe it can be adjusted top be kept truly indirect. -boris

...tive > displacement, I was trying to say that pv_irq_ops always live further down) I believe the problem is this: #define PV_INDIRECT(addr) *addr(%rip) The displacement that the linker computes will be relative to the where this instruction is placed at the time of linking, which is in .pv_altinstructions (and not .text). So when we copy it into .text the displacement becomes bogus. Replacing the macro with #define PV_INDIRECT(addr) *addr // well, it's not so much indirect anymore makes things work. Or maybe it can be adjusted top be kept truly indirect. -boris

...say that pv_irq_ops always live further down) > > I believe the problem is this: > > #define PV_INDIRECT(addr) *addr(%rip) > > The displacement that the linker computes will be relative to the where > this instruction is placed at the time of linking, which is in > .pv_altinstructions (and not .text). So when we copy it into .text the > displacement becomes bogus. apply_alternatives() is supposed to adjust that displacement based on the new IP, though it could be messing that up somehow. (See patch 10/13.) -- Josh

...to say that pv_irq_ops always live further down) > > I believe the problem is this: > > #define PV_INDIRECT(addr) *addr(%rip) > > The displacement that the linker computes will be relative to the where > this instruction is placed at the time of linking, which is in > .pv_altinstructions (and not .text). So when we copy it into .text the > displacement becomes bogus. > > Replacing the macro with > > #define PV_INDIRECT(addr) *addr // well, it's not so much > indirect anymore > > makes things work. Or maybe it can be adjusted top be kept truly...

...always live further down) >> I believe the problem is this: >> >> #define PV_INDIRECT(addr) *addr(%rip) >> >> The displacement that the linker computes will be relative to the where >> this instruction is placed at the time of linking, which is in >> .pv_altinstructions (and not .text). So when we copy it into .text the >> displacement becomes bogus. > apply_alternatives() is supposed to adjust that displacement based on > the new IP, though it could be messing that up somehow. (See patch > 10/13.) > That patch doesn't take into account...

...86/paravirt: Simplify ____PVOP_CALL() x86/paravirt: Clean up paravirt_types.h x86/asm: Convert ALTERNATIVE*() assembler macros to preprocessor macros x86/alternative: Support indirect call replacement x86/paravirt: Add paravirt alternatives infrastructure objtool: Add support for new .pv_altinstructions section x86/paravirt: Convert natively patched pv ops to use paravirt alternatives arch/x86/entry/entry_32.S | 13 +- arch/x86/entry/entry_64.S | 12 +- arch/x86/entry/entry_64_compat.S | 9 +- arch/x86/entry/vdso/vdso32/system_call.S | 10 +-...

...86/paravirt: Simplify ____PVOP_CALL() x86/paravirt: Clean up paravirt_types.h x86/asm: Convert ALTERNATIVE*() assembler macros to preprocessor macros x86/alternative: Support indirect call replacement x86/paravirt: Add paravirt alternatives infrastructure objtool: Add support for new .pv_altinstructions section x86/paravirt: Convert natively patched pv ops to use paravirt alternatives arch/x86/entry/entry_32.S | 13 +- arch/x86/entry/entry_64.S | 12 +- arch/x86/entry/entry_64_compat.S | 9 +- arch/x86/entry/vdso/vdso32/system_call.S | 10 +-...

...> I believe the problem is this: > >> > >> #define PV_INDIRECT(addr) *addr(%rip) > >> > >> The displacement that the linker computes will be relative to the where > >> this instruction is placed at the time of linking, which is in > >> .pv_altinstructions (and not .text). So when we copy it into .text the > >> displacement becomes bogus. > > apply_alternatives() is supposed to adjust that displacement based on > > the new IP, though it could be messing that up somehow. (See patch > > 10/13.) > > > > That...

...> I believe the problem is this: > >> > >> #define PV_INDIRECT(addr) *addr(%rip) > >> > >> The displacement that the linker computes will be relative to the where > >> this instruction is placed at the time of linking, which is in > >> .pv_altinstructions (and not .text). So when we copy it into .text the > >> displacement becomes bogus. > > apply_alternatives() is supposed to adjust that displacement based on > > the new IP, though it could be messing that up somehow. (See patch > > 10/13.) > > > > That...

On 12/10/17 20:11, Boris Ostrovsky wrote: > On 10/06/2017 10:32 AM, Josh Poimboeuf wrote: >> On Thu, Oct 05, 2017 at 04:35:03PM -0400, Boris Ostrovsky wrote: >>>> #ifdef CONFIG_PARAVIRT >>>> +/* >>>> + * Paravirt alternatives are applied much earlier than normal alternatives. >>>> + * They are only applied when running on a hypervisor.

On 12/10/17 20:11, Boris Ostrovsky wrote: > On 10/06/2017 10:32 AM, Josh Poimboeuf wrote: >> On Thu, Oct 05, 2017 at 04:35:03PM -0400, Boris Ostrovsky wrote: >>>> #ifdef CONFIG_PARAVIRT >>>> +/* >>>> + * Paravirt alternatives are applied much earlier than normal alternatives. >>>> + * They are only applied when running on a hypervisor.