xen: drop hypercall function tables
cover xen: drop hypercall function tables
In order to avoid indirect function calls on the hypercall path as much as possible this series is removing the hypercall function tables and is replacing the hypercall handler calls via the function array by automatically generated switch statements.
Another by-product of generating the switch statements is the automatic generating of the hypercall handler prototypes from the same data base which is used to generate the switch.
This has the additional advantage of using type safe calls of the handlers and to ensure related handler (e.g. PV and HVM ones) share the same prototypes.
A very brief performance test (parallel build of the Xen hypervisor in a 6 vcpu guest) showed a very slim improvement (less than 1%) of the performance with the patches applied. The test was performed using a PV and a PVH guest.
Juergen Gross (12):
xen: limit number of hypercall parameters to 5
xen: move do_vcpu_op() to arch specific code
xen: harmonize return types of hypercall handlers xen/x86: modify hvm_memory_op() prototype
xen: don't include asm/hypercall.h from C sources
xen: generate hypercall interface related code
xen: use generated prototypes for hypercall handlers x86/pv-shim: don't modify hypercall table
xen/x86: don't use hypercall table for calling compat hypercalls xen/x86: call hypercall handlers via switch statement xen/arm: call hypercall handlers via switch statement xen/x86: add hypercall performance counters for hvm, correct pv
.gitignore | 2 + xen/Makefile | 10 + xen/arch/arm/domain.c | 15 +- xen/arch/arm/hvm.c | 3 +- xen/arch/arm/physdev.c | 2 +- xen/arch/arm/platform_hypercall.c | 6 +- xen/arch/arm/traps.c | 124 +++-------- xen/arch/x86/compat.c | 14 +- xen/arch/x86/cpu/vpmu.c | 1 + xen/arch/x86/domain.c | 11 +- xen/arch/x86/domctl.c | 4 +- xen/arch/x86/hvm/hypercall.c | 172 ++------------ xen/arch/x86/hypercall.c | 59 ----- xen/arch/x86/mm.c | 1 - xen/arch/x86/mm/paging.c | 3 +- xen/arch/x86/platform_hypercall.c | 6 +- xen/arch/x86/pv/callback.c | 20 +- xen/arch/x86/pv/emul-priv-op.c | 2 +- xen/arch/x86/pv/hypercall.c | 188 ++-------------- xen/arch/x86/pv/iret.c | 5 +- xen/arch/x86/pv/misc-hypercalls.c | 14 +- xen/arch/x86/pv/shim.c | 48 ++-- xen/arch/x86/traps.c | 2 +- xen/arch/x86/x86_64/compat.c | 1 - xen/arch/x86/x86_64/compat/mm.c | 1 + xen/arch/x86/x86_64/domain.c | 16 +- xen/arch/x86/x86_64/mm.c | 1 - xen/arch/x86/x86_64/platform_hypercall.c | 9 +- xen/common/argo.c | 12 +- xen/common/compat/domain.c | 14 +- xen/common/compat/grant_table.c | 1 + xen/common/compat/multicall.c | 2 +- xen/common/domain.c | 11 +- xen/common/event_channel.c | 10 + xen/common/grant_table.c | 10 + xen/common/kexec.c | 6 +- xen/common/multicall.c | 1 + xen/common/trace.c | 2 +- xen/include/asm-arm/hypercall.h | 7 +- xen/include/asm-x86/hypercall.h | 195 +++------------- xen/include/asm-x86/paging.h | 3 - xen/include/asm-x86/pv/shim.h | 3 + xen/include/hypercall-defs.c | 272 +++++++++++++++++++++++ xen/include/xen/hypercall.h | 181 +-------------- xen/scripts/gen_hypercall.awk | 228 +++++++++++++++++++ 45 files changed, 769 insertions(+), 929 deletions(-) create mode 100644 xen/include/hypercall-defs.c create mode 100644 xen/scripts/gen_hypercall.awk
Signature · 1 line
2.26.2
PATCH 1/12 xen: limit number of hypercall parameters to 5
Today there is no hypercall with more than 5 parameters, while the ABI allows up to 6 parameters. Especially for the X86 32-bit case using 6 parameters would require to run without frame pointer, which isn't very fortunate. Note that for Arm the limit is 5 parameters already.
So limit the maximum number of parameters to 5 for x86, too.
--- xen/arch/x86/hvm/hypercall.c | 28 ++++++++++------------------ xen/arch/x86/pv/hypercall.c | 22 ++++++++-------------- xen/common/trace.c | 2 +- xen/include/asm-x86/hypercall.h | 2 +- 4 files changed, 20 insertions(+), 34 deletions(-)
xen/arch/x86/hvm/hypercall.c +10 -18
@@ -239,10 +239,9 @@ int hvm_hypercall(struct cpu_user_regs *regs)unsigned long rdx = regs->rdx;unsigned long r10 = regs->r10;unsigned long r8 = regs->r8; -unsigned long r9 = regs->r9; -HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%lx, %lx, %lx, %lx, %lx, %lx)" , - eax, rdi, rsi, rdx, r10, r8, r9); +HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%lx, %lx, %lx, %lx, %lx)" , + eax, rdi, rsi, rdx, r10, r8);#ifndef NDEBUG /* Deliberately corrupt parameter regs not used by this hypercall. */
@@ -252,13 +251,11 @@ int hvm_hypercall(struct cpu_user_regs *regs)case 1 : rsi =0xdeadbeefdeadf00dUL ; fallthrough;case 2 : rdx =0xdeadbeefdeadf00dUL ; fallthrough;case 3 : r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : r8 =0xdeadbeefdeadf00dUL ; fallthrough; - case 5: r9 = 0xdeadbeefdeadf00dUL; + case 4: r8 = 0xdeadbeefdeadf00dUL; }#endif - regs->rax = hvm_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8, - r9); + regs->rax = hvm_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8);#ifndef NDEBUG if ( !curr->hcall_preempted )
@@ -266,7 +263,6 @@ int hvm_hypercall(struct cpu_user_regs *regs)/* Deliberately corrupt parameter regs used by this hypercall. */ switch ( hypercall_args_table[eax].native ) { -case 6 : regs->r9 =0xdeadbeefdeadf00dUL ; fallthrough;case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough;case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough;case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough;
@@ -283,10 +279,9 @@ int hvm_hypercall(struct cpu_user_regs *regs)unsigned int edx = regs->edx;unsigned int esi = regs->esi;unsigned int edi = regs->edi; -unsigned int ebp = regs->ebp; -HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%x, %x, %x, %x, %x, %x)" , eax, - ebx, ecx, edx, esi, edi, ebp); +HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%x, %x, %x, %x, %x)" , eax, + ebx, ecx, edx, esi, edi);#ifndef NDEBUG /* Deliberately corrupt parameter regs not used by this hypercall. */
@@ -296,14 +291,12 @@ int hvm_hypercall(struct cpu_user_regs *regs)case 1 : ecx =0xdeadf00d ; fallthrough;case 2 : edx =0xdeadf00d ; fallthrough;case 3 : esi =0xdeadf00d ; fallthrough; -case 4 : edi =0xdeadf00d ; fallthrough; - case 5: ebp = 0xdeadf00d; + case 4: edi = 0xdeadf00d; }#endif curr->hcall_compat = true; - regs->rax = hvm_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi, - ebp); + regs->rax = hvm_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi); curr->hcall_compat = false;#ifndef NDEBUG
@@ -312,7 +305,6 @@ int hvm_hypercall(struct cpu_user_regs *regs)/* Deliberately corrupt parameter regs used by this hypercall. */ switch ( hypercall_args_table[eax].compat ) { -case 6 : regs->rbp =0xdeadf00d ; fallthrough;case 5 : regs->rdi =0xdeadf00d ; fallthrough;case 4 : regs->rsi =0xdeadf00d ; fallthrough;case 3 : regs->rdx =0xdeadf00d ; fallthrough;
@@ -349,7 +341,7 @@ enum mc_disposition hvm_do_multicall_call(struct mc_state *state) func =array_access_nospec (hvm_hypercall_table, call->op).native;if ( func ) call->result =func (call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ], call->args[5 ]); + call->args[3 ], call->args[4 ]);else call->result = -ENOSYS ; }
@@ -361,7 +353,7 @@ enum mc_disposition hvm_do_multicall_call(struct mc_state *state) func =array_access_nospec (hvm_hypercall_table, call->op).compat;if ( func ) call->result =func (call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ], call->args[5 ]); + call->args[3 ], call->args[4 ]);else call->result = -ENOSYS ; }
xen/arch/x86/pv/hypercall.c +8 -14
@@ -145,7 +145,6 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)unsigned long rdx = regs->rdx;unsigned long r10 = regs->r10;unsigned long r8 = regs->r8; -unsigned long r9 = regs->r9;#ifndef NDEBUG /* Deliberately corrupt parameter regs not used by this hypercall. */
@@ -155,18 +154,17 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)case 1 : rsi =0xdeadbeefdeadf00dUL ; fallthrough;case 2 : rdx =0xdeadbeefdeadf00dUL ; fallthrough;case 3 : r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : r8 =0xdeadbeefdeadf00dUL ; fallthrough; - case 5: r9 = 0xdeadbeefdeadf00dUL; + case 4: r8 = 0xdeadbeefdeadf00dUL; }#endif if (unlikely (tb_init_done) ) { -unsigned long args[6 ] = { rdi, rsi, rdx, r10, r8, r9 }; +unsigned long args[5 ] = { rdi, rsi, rdx, r10, r8 };__trace_hypercall (TRC_PV_HYPERCALL_V2 , eax, args); } - regs->rax = pv_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8, r9); + regs->rax = pv_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8);#ifndef NDEBUG if ( !curr->hcall_preempted )
@@ -174,7 +172,6 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)/* Deliberately corrupt parameter regs used by this hypercall. */ switch ( hypercall_args_table[eax].native ) { -case 6 : regs->r9 =0xdeadbeefdeadf00dUL ; fallthrough;case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough;case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough;case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough;
@@ -192,7 +189,6 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)unsigned int edx = regs->edx;unsigned int esi = regs->esi;unsigned int edi = regs->edi; -unsigned int ebp = regs->ebp;#ifndef NDEBUG /* Deliberately corrupt parameter regs not used by this hypercall. */
@@ -202,20 +198,19 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)case 1 : ecx =0xdeadf00d ; fallthrough;case 2 : edx =0xdeadf00d ; fallthrough;case 3 : esi =0xdeadf00d ; fallthrough; -case 4 : edi =0xdeadf00d ; fallthrough; - case 5: ebp = 0xdeadf00d; + case 4: edi = 0xdeadf00d; }#endif if (unlikely (tb_init_done) ) { -unsigned long args[6 ] = { ebx, ecx, edx, esi, edi, ebp }; +unsigned long args[5 ] = { ebx, ecx, edx, esi, edi };__trace_hypercall (TRC_PV_HYPERCALL_V2 , eax, args); } curr->hcall_compat = true; - regs->eax = pv_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi, ebp); + regs->eax = pv_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi); curr->hcall_compat = false;#ifndef NDEBUG
@@ -224,7 +219,6 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)/* Deliberately corrupt parameter regs used by this hypercall. */ switch ( hypercall_args_table[eax].compat ) { -case 6 : regs->ebp =0xdeadf00d ; fallthrough;case 5 : regs->edi =0xdeadf00d ; fallthrough;case 4 : regs->esi =0xdeadf00d ; fallthrough;case 3 : regs->edx =0xdeadf00d ; fallthrough;
@@ -262,7 +256,7 @@ enum mc_disposition pv_do_multicall_call(struct mc_state *state) pv_hypercall_table[op].compat ) call->result = pv_hypercall_table[op].compat ( call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ], call->args[5 ]); + call->args[3 ], call->args[4 ]); else call->result = -ENOSYS; }
@@ -276,7 +270,7 @@ enum mc_disposition pv_do_multicall_call(struct mc_state *state) pv_hypercall_table[op].native ) call->result = pv_hypercall_table[op].native ( call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ], call->args[5 ]); + call->args[3 ], call->args[4 ]); else call->result = -ENOSYS; }
xen/common/trace.c +1 -1
@@ -822,7 +822,7 @@ void __trace_hypercall(uint32_t event, unsigned long op, {struct {uint32_t op; -uint32_t args[6 ]; +uint32_t args[5 ]; } d;uint32_t *a = d.args;
xen/include/asm-x86/hypercall.h +1 -1
@@ -13,7 +13,7 @@typedef unsigned long hypercall_fn_t (unsigned long ,unsigned long ,unsigned long , -unsigned long ,unsigned long ,unsigned long ); +unsigned long ,unsigned long ); typedef struct { hypercall_fn_t *native;
Signature · 1 line
2.26.2
PATCH 2/12 xen: move do_vcpu_op() to arch specific code
In reply to Juergen GrossToday Arm is using another entry point for the vcpu_op hypercall as x86, as some of the common sub-ops are not supported on Arm. The Arm specific handler filetrs out the not supported sub-ops and then calls the common handler. This leads to the weird call hierarchy:
do_arm_vcpu_op()
do_vcpu_op()
arch_do_vcpu_op()Clean this up by renaming do_vcpu_op() to common_vcpu_op() and arch_do_vcpu_op() in each architecture to do_vcpu_op(). This way one of above calls can be avoided without restricting any potential future use of common sub-ops for Arm.
Additionally the single user of HYPERCALL_ARM() can be modified and HYPERCALL_ARM() can be removed.
--- xen/arch/arm/domain.c | 15 ++++++++------- xen/arch/arm/traps.c | 7 +------ xen/arch/x86/domain.c | 11 +++++++---- xen/arch/x86/x86_64/domain.c | 16 ++++++++++++---- xen/common/compat/domain.c | 14 ++++++-------- xen/common/domain.c | 11 ++++------- xen/include/asm-arm/hypercall.h | 2 -- xen/include/asm-x86/hypercall.h | 2 +- xen/include/xen/hypercall.h | 2 +- 9 files changed, 40 insertions(+), 40 deletions(-)
xen/arch/arm/domain.c +8 -7
@@ -1070,23 +1070,24 @@ void arch_dump_domain_info(struct domain *d) } -long do_arm_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) +long do_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) { +struct domain *d = current->domain; +struct vcpu *v; + +if ( (v =domain_vcpu (d, vcpuid)) ==NULL ) +return -ENOENT ; +switch ( cmd ) {case VCPUOP_register_vcpu_info:case VCPUOP_register_runstate_memory_area: -return do_vcpu_op (cmd, vcpuid, arg); +return common_vcpu_op (cmd, v, arg);default :return -EINVAL ; } } -long arch_do_vcpu_op (int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg) -{ -return -ENOSYS ; -} -void arch_dump_vcpu_info (struct vcpu *v) {gic_dump_info (v);
xen/arch/arm/traps.c +1 -6
@@ -1351,11 +1351,6 @@ typedef struct { .nr_args = _nr_args, \ } -#define HYPERCALL_ARM (_name, _nr_args) \ - [ __HYPERVISOR_ ## _name ] = { \ - .fn = (arm_hypercall_fn_t) &do_arm_ ## _name, \ - .nr_args = _nr_args, \ - }/* * Only use this for hypercalls which were deprecated (i.e. replaced * by something else) before Xen on ARM was created, i.e. *not* for
@@ -1386,7 +1381,7 @@ static arm_hypercall_t arm_hypercall_table[] = {#endif HYPERCALL (multicall,2 ),HYPERCALL (platform_op,1 ), -HYPERCALL_ARM (vcpu_op,3 ), +HYPERCALL (vcpu_op,3 ),HYPERCALL (vm_assist,2 ), #ifdef CONFIG_ARGO HYPERCALL(argo_op, 5),
xen/arch/x86/domain.c +7 -4
@@ -1488,11 +1488,14 @@ int arch_vcpu_reset(struct vcpu *v)return 0 ; } -long -arch_do_vcpu_op ( -int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg) +long do_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc =0 ; +struct domain *d = current->domain; +struct vcpu *v; + +if ( (v =domain_vcpu (d, vcpuid)) ==NULL ) +return -ENOENT ;switch ( cmd ) {
@@ -1544,7 +1547,7 @@ arch_do_vcpu_op( }default : - rc = -ENOSYS ; + rc =common_vcpu_op (cmd, v, arg);break ; }
xen/arch/x86/x86_64/domain.c +12 -4
@@ -13,10 +13,14 @@ CHECK_vcpu_get_physid;#undef xen_vcpu_get_physidint -arch_compat_vcpu_op ( -int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg) +compat_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) { -int rc = -ENOSYS ; +int rc; +struct domain *d = current->domain; +struct vcpu *v; + +if ( (v =domain_vcpu (d, vcpuid)) ==NULL ) +return -ENOENT ;switch ( cmd ) {
@@ -55,7 +59,11 @@ arch_compat_vcpu_op( }case VCPUOP_get_physid: - rc =arch_do_vcpu_op (cmd, v, arg); + rc =do_vcpu_op (cmd, vcpuid, arg); +break ; + +default : + rc =compat_common_vcpu_op (cmd, v, arg);break ; }
xen/common/compat/domain.c +6 -8
@@ -38,14 +38,12 @@ CHECK_vcpu_hvm_context;#endif -int compat_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) +int compat_common_vcpu_op (int cmd,struct vcpu *v, +XEN_GUEST_HANDLE_PARAM (void ) arg) { -struct domain *d = current->domain; -struct vcpu *v;int rc =0 ; - -if ( (v =domain_vcpu (d, vcpuid)) ==NULL ) -return -ENOENT ; +struct domain *d = current->domain; +unsigned int vcpuid = v->vcpu_id;switch ( cmd ) {
@@ -102,7 +100,7 @@ int compat_vcpu_op(int cmd, unsigned int vcpuid, XEN_GUEST_HANDLE_PARAM(void) arcase VCPUOP_stop_singleshot_timer:case VCPUOP_register_vcpu_info:case VCPUOP_send_nmi: - rc =do_vcpu_op (cmd, vcpuid, arg); + rc =common_vcpu_op (cmd, v, arg);break ;case VCPUOP_get_runstate_info:
@@ -133,7 +131,7 @@ int compat_vcpu_op(int cmd, unsigned int vcpuid, XEN_GUEST_HANDLE_PARAM(void) ar }default : - rc =arch_compat_vcpu_op (cmd, v, arg); + rc = -ENOSYS ;break ; }
xen/common/domain.c +4 -7
@@ -1559,14 +1559,11 @@ int default_initialise_vcpu(struct vcpu *v, XEN_GUEST_HANDLE_PARAM(void) arg)return rc; } -long do_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg) +long common_vcpu_op (int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg) { -struct domain *d = current->domain; -struct vcpu *v;long rc =0 ; - -if ( (v =domain_vcpu (d, vcpuid)) ==NULL ) -return -ENOENT ; +struct domain *d = current->domain; +unsigned int vcpuid = v->vcpu_id;switch ( cmd ) {
@@ -1736,7 +1733,7 @@ long do_vcpu_op(int cmd, unsigned int vcpuid, XEN_GUEST_HANDLE_PARAM(void) arg) }default : - rc =arch_do_vcpu_op (cmd, v, arg); + rc = -ENOSYS ;break ; }
xen/include/asm-arm/hypercall.h +0 -2
@@ -4,8 +4,6 @@#include <public/domctl.h> /* for arch_do_domctl */ int do_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg); -long do_arm_vcpu_op (int cmd,unsigned int vcpuid,XEN_GUEST_HANDLE_PARAM (void ) arg); -long subarch_do_domctl (struct xen_domctl *domctl,struct domain *d,XEN_GUEST_HANDLE_PARAM (xen_domctl_t) u_domctl);
xen/include/asm-x86/hypercall.h +1 -1
@@ -152,7 +152,7 @@ compat_physdev_op( XEN_GUEST_HANDLE_PARAM(void) arg);extern int -arch_compat_vcpu_op ( +compat_common_vcpu_op (int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg); extern int compat_mmuext_op(
xen/include/xen/hypercall.h +1 -1
@@ -110,7 +110,7 @@ do_vcpu_op(struct vcpu ;extern long -arch_do_vcpu_op (int cmd, +common_vcpu_op (int cmd,struct vcpu *v,XEN_GUEST_HANDLE_PARAM (void ) arg);
Signature · 1 line
2.26.2
PATCH 3/12 xen: harmonize return types of hypercall handlers
In reply to Juergen GrossToday most hypercall handlers have a return type of long, while the compat ones return an int. There are a few exceptions from that rule, however.
Get rid of the exceptions by letting compat handlers always return int and others always return long.
For the compat hvm case use eax instead of rax for the stored result as it should have been from the beginning.
Additionally move some prototypes to include/asm-x86/hypercall.h as they are x86 specific. Move the do_physdev_op() prototype from both architecture dependant headers to the common one. Move the compat_platform_op() prototype to the common header.
Switch some non style compliant types (u32, s32, s64) to style compliant ones.
Rename paging_domctl_continuation() to do_paging_domctl_cont() and add a matching define for the associated hypercall.
Make do_callback_op() and compat_callback_op() more similar by adding the const attribute to compat_callback_op()'s 2nd parameter.
The do_platform_op() prototype needs to be modified in order to better match its compat variant.
Change the type of the cmd parameter for [do|compat]_kexec_op() to unsigned int, as this is more appropriate for the compat case.
--- xen/arch/arm/physdev.c | 2 +- xen/arch/arm/platform_hypercall.c | 5 ++- xen/arch/x86/domctl.c | 4 +-- xen/arch/x86/hvm/hypercall.c | 8 ++--- xen/arch/x86/hypercall.c | 2 +- xen/arch/x86/mm/paging.c | 3 +- xen/arch/x86/platform_hypercall.c | 5 ++- xen/arch/x86/pv/callback.c | 20 +++++------ xen/arch/x86/pv/emul-priv-op.c | 2 +- xen/arch/x86/pv/hypercall.c | 5 +-- xen/arch/x86/pv/iret.c | 4 +-- xen/arch/x86/pv/misc-hypercalls.c | 14 +++++--- xen/arch/x86/x86_64/platform_hypercall.c | 5 ++- xen/common/argo.c | 12 +++---- xen/common/kexec.c | 6 ++-- xen/include/asm-arm/hypercall.h | 1 - xen/include/asm-x86/hypercall.h | 43 +++++++++++------------- xen/include/asm-x86/paging.h | 3 -- xen/include/xen/hypercall.h | 26 +++++++------- 19 files changed, 84 insertions(+), 86 deletions(-)
xen/arch/arm/physdev.c +1 -1
@@ -11,7 +11,7 @@#include <xen/hypercall.h> -int do_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +long do_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {#ifdef CONFIG_HAS_PCI return pci_physdev_op (cmd, arg);
xen/arch/arm/platform_hypercall.c +4 -1
@@ -17,12 +17,15 @@DEFINE_SPINLOCK (xenpf_lock); -long do_platform_op (XEN_GUEST_HANDLE_PARAM (xen_platform_op_t) u_xenpf_op) +long do_platform_op (XEN_GUEST_HANDLE_PARAM (void ) arg) {long ret; +XEN_GUEST_HANDLE_PARAM (xen_platform_op_t) u_xenpf_op;struct xen_platform_op curop, *op = &curop;struct domain *d; + u_xenpf_op =guest_handle_cast (arg, xen_platform_op_t); +if (copy_from_guest (op, u_xenpf_op,1 ) )return -EFAULT ;
xen/arch/x86/domctl.c +2 -2
@@ -221,8 +221,8 @@ long arch_do_domctl(case XEN_DOMCTL_shadow_op: ret =paging_domctl (d, &domctl->u.shadow_op, u_domctl,0 );if ( ret == -ERESTART ) -return hypercall_create_continuation (__HYPERVISOR_arch_1, -"h" , u_domctl); +return hypercall_create_continuation ( + __HYPERVISOR_paging_domctl_cont,"h" , u_domctl); copyback = true;break ;
xen/arch/x86/hvm/hypercall.c +2 -6
@@ -120,8 +120,6 @@ static long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg) [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ (hypercall_fn_t *) compat_ ## x } -#define do_arch_1 paging_domctl_continuation -static const struct {hypercall_fn_t *native, *compat; } hvm_hypercall_table[] = {
@@ -154,11 +152,9 @@ static const struct {#ifdef CONFIG_HYPFS HYPERCALL (hypfs_op), #endif -HYPERCALL (arch_1) +HYPERCALL (paging_domctl_cont) }; -#undef do_arch_1 -#undef HYPERCALL#undef HVM_CALL#undef COMPAT_CALL
@@ -296,7 +292,7 @@ int hvm_hypercall(struct cpu_user_regs *regs)#endif curr->hcall_compat = true; - regs->rax = hvm_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi); + regs->eax = hvm_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi); curr->hcall_compat = false;#ifndef NDEBUG
xen/arch/x86/hypercall.c +1 -1
@@ -75,7 +75,7 @@ const hypercall_args_t hypercall_args_table[NR_hypercalls] =ARGS (dm_op,3 ),ARGS (hypfs_op,5 ),ARGS (mca,1 ), - ARGS(arch_1, 1), + ARGS(paging_domctl_cont, 1), };#undef COMP
xen/arch/x86/mm/paging.c +2 -1
@@ -21,6 +21,7 @@#include <xen/init.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <asm/paging.h> #include <asm/shadow.h> #include <asm/p2m.h>
@@ -756,7 +757,7 @@ int paging_domctl(struct domain *d, struct xen_domctl_shadow_op *sc,return shadow_domctl (d, sc, u_domctl); } -long paging_domctl_continuation (XEN_GUEST_HANDLE_PARAM (xen_domctl_t) u_domctl) +long do_paging_domctl_cont (XEN_GUEST_HANDLE_PARAM (xen_domctl_t) u_domctl) {struct xen_domctl op;struct domain *d;
xen/arch/x86/platform_hypercall.c +4 -1
@@ -210,11 +210,14 @@ void resource_access(void *info) }#endif -ret_t do_platform_op (XEN_GUEST_HANDLE_PARAM (xen_platform_op_t) u_xenpf_op) +ret_t do_platform_op (XEN_GUEST_HANDLE_PARAM (void ) arg) {ret_t ret; +XEN_GUEST_HANDLE_PARAM (xen_platform_op_t) u_xenpf_op;struct xen_platform_op curop, *op = &curop; + u_xenpf_op =guest_handle_cast (arg, xen_platform_op_t); +if (copy_from_guest (op, u_xenpf_op,1 ) )return -EFAULT ;
xen/arch/x86/pv/callback.c +10 -10
@@ -207,9 +207,9 @@ long do_set_callbacks(unsigned long event_address,#include <compat/callback.h> #include <compat/nmi.h> -static long compat_register_guest_callback (struct compat_callback_register *reg) +static int compat_register_guest_callback (struct compat_callback_register *reg) { -long ret =0 ; +int ret =0 ;struct vcpu *curr = current;fixup_guest_code_selector (curr->domain, reg->address.cs);
@@ -256,10 +256,10 @@ static long compat_register_guest_callback(struct compat_callback_register *reg)return ret; } -static long compat_unregister_guest_callback ( +static int compat_unregister_guest_callback (struct compat_callback_unregister *unreg) { -long ret; +int ret;switch ( unreg->type ) {
@@ -283,9 +283,9 @@ static long compat_unregister_guest_callback(return ret; } -long compat_callback_op (int cmd,XEN_GUEST_HANDLE (void ) arg) +int compat_callback_op (int cmd,XEN_GUEST_HANDLE (const_void) arg) { -long ret; +int ret;switch ( cmd ) {
@@ -321,10 +321,10 @@ long compat_callback_op(int cmd, XEN_GUEST_HANDLE(void) arg)return ret; } -long compat_set_callbacks (unsigned long event_selector, -unsigned long event_address, -unsigned long failsafe_selector, -unsigned long failsafe_address) +int compat_set_callbacks (unsigned long event_selector, +unsigned long event_address, +unsigned long failsafe_selector, +unsigned long failsafe_address) {struct compat_callback_register event = { .type = CALLBACKTYPE_event,
xen/arch/x86/pv/emul-priv-op.c +1 -1
@@ -22,12 +22,12 @@#include <xen/domain_page.h> #include <xen/event.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/iocap.h> #include <asm/amd.h> #include <asm/debugreg.h> #include <asm/hpet.h> -#include <asm/hypercall.h> #include <asm/mc146818rtc.h> #include <asm/pv/domain.h> #include <asm/pv/trace.h>
xen/arch/x86/pv/hypercall.c +1 -4
@@ -40,8 +40,6 @@#define COMPAT_CALL (x) HYPERCALL(x)#endif -#define do_arch_1 paging_domctl_continuation -const pv_hypercall_table_t pv_hypercall_table[] = {COMPAT_CALL (set_trap_table),HYPERCALL (mmu_update),
@@ -102,11 +100,10 @@ const pv_hypercall_table_t pv_hypercall_table[] = {#endif HYPERCALL(mca), #ifndef CONFIG_PV_SHIM_EXCLUSIVE - HYPERCALL(arch_1), + HYPERCALL(paging_domctl_cont), #endif }; -#undef do_arch_1#undef COMPAT_CALL#undef HYPERCALL
xen/arch/x86/pv/iret.c +2 -2
@@ -48,7 +48,7 @@ static void async_exception_cleanup(struct vcpu *curr) curr->arch.async_exception_state (trap).old_mask; } -unsigned long do_iret (void ) +long do_iret (void ) {struct cpu_user_regs *regs =guest_cpu_user_regs ();struct iret_context iret_saved;
@@ -105,7 +105,7 @@ unsigned long do_iret(void) }#ifdef CONFIG_PV32 -unsigned int compat_iret (void ) +int compat_iret (void ) {struct cpu_user_regs *regs =guest_cpu_user_regs ();struct vcpu *v = current;
xen/arch/x86/pv/misc-hypercalls.c +9 -5
@@ -28,12 +28,16 @@ long do_set_debugreg(int reg, unsigned long value)return set_debugreg (current, reg, value); } -unsigned long do_get_debugreg (int reg) +long do_get_debugreg (int reg) { -unsigned long val; -int res =x86emul_read_dr (reg, &val,NULL ); - -return res ==X86EMUL_OKAY ? val : -ENODEV ; +/* Avoid undefined behavior due to casting an unsigned long to long. */ +union { +unsigned long val; +long ret; + } u; +int res =x86emul_read_dr (reg, &u.val,NULL ); + +return res ==X86EMUL_OKAY ? u.ret : -ENODEV ; }long do_fpu_taskswitch (int set)
xen/arch/x86/x86_64/platform_hypercall.c +4 -1
@@ -4,13 +4,16 @@EMIT_FILE ; +#include <xen/types.h> +#include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/lib.h> #include <compat/platform.h> DEFINE_XEN_GUEST_HANDLE (compat_platform_op_t);#define xen_platform_op compat_platform_op#define xen_platform_op_t compat_platform_op_t -#define do_platform_op (x) compat_platform_op(_##x) +#define do_platform_op (x) compat_platform_op(x)#define efi_get_info efi_compat_get_info#define efi_runtime_call (x) efi_compat_runtime_call(x)
xen/common/argo.c +6 -6
@@ -2207,13 +2207,13 @@ do_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, }#ifdef CONFIG_COMPAT -long -compat_argo_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg1, -XEN_GUEST_HANDLE_PARAM (void ) arg2,unsigned long arg3, -unsigned long arg4) +int compat_argo_op (unsigned int cmd, +XEN_GUEST_HANDLE_PARAM (void ) arg1, +XEN_GUEST_HANDLE_PARAM (void ) arg2, +unsigned long arg3,unsigned long arg4) {struct domain *currd = current->domain; -long rc; +int rc;xen_argo_send_addr_t send_addr;xen_argo_iov_t iovs[XEN_ARGO_MAXIOV ];compat_argo_iov_t compat_iovs[XEN_ARGO_MAXIOV ];
@@ -2267,7 +2267,7 @@ compat_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, rc =sendv (currd, &send_addr.src, &send_addr.dst, iovs, niov, arg4); out: -argo_dprintk ("<-compat_argo_op(%u)=%ld\n" , cmd, rc); +argo_dprintk ("<-compat_argo_op(%u)=%d\n" , cmd, rc);return rc; }
xen/common/kexec.c +3 -3
@@ -1213,7 +1213,7 @@ static int kexec_status(XEN_GUEST_HANDLE_PARAM(void) uarg)return !!test_bit (bit, &kexec_flags); } -static int do_kexec_op_internal (unsigned long op, +static int do_kexec_op_internal (unsigned int op,XEN_GUEST_HANDLE_PARAM (void ) uarg,bool_t compat) {
@@ -1265,13 +1265,13 @@ static int do_kexec_op_internal(unsigned long op,return ret; } -long do_kexec_op (unsigned long op,XEN_GUEST_HANDLE_PARAM (void ) uarg) +long do_kexec_op (unsigned int op,XEN_GUEST_HANDLE_PARAM (void ) uarg) {return do_kexec_op_internal (op, uarg,0 ); }#ifdef CONFIG_COMPAT -int compat_kexec_op (unsigned long op,XEN_GUEST_HANDLE_PARAM (void ) uarg) +int compat_kexec_op (unsigned int op,XEN_GUEST_HANDLE_PARAM (void ) uarg) {return do_kexec_op_internal (op, uarg,1 ); }
xen/include/asm-arm/hypercall.h +0 -1
@@ -2,7 +2,6 @@#define __ASM_ARM_HYPERCALL_H__#include <public/domctl.h> /* for arch_do_domctl */ -int do_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg);long subarch_do_domctl (struct xen_domctl *domctl,struct domain *d,XEN_GUEST_HANDLE_PARAM (xen_domctl_t) u_domctl);
xen/include/asm-x86/hypercall.h +20 -23
103 lines not on this page · raw message
xen/include/asm-x86/paging.h +0 -3
@@ -236,9 +236,6 @@ int paging_domctl(struct domain *d, struct xen_domctl_shadow_op *sc,XEN_GUEST_HANDLE_PARAM (xen_domctl_t) u_domctl, bool_t resuming); -/* Helper hypercall for dealing with continuations. */ -long paging_domctl_continuation (XEN_GUEST_HANDLE_PARAM (xen_domctl_t )); -/* Call when destroying a vcpu/domain */ void paging_vcpu_teardown (struct vcpu *v);int paging_teardown (struct domain *d);
xen/include/xen/hypercall.h +12 -14
@@ -43,7 +43,7 @@ arch_do_sysctl(extern long do_platform_op ( -XEN_GUEST_HANDLE_PARAM (xen_platform_op_t) u_xenpf_op); +XEN_GUEST_HANDLE_PARAM (void ) u_xenpf_op);extern long pci_physdev_op(
@@ -114,11 +114,6 @@ common_vcpu_op(int cmd, struct vcpu *v, XEN_GUEST_HANDLE_PARAM(void) arg); -extern long -do_nmi_op ( -unsigned int cmd, -XEN_GUEST_HANDLE_PARAM (void ) arg); - extern long do_hvm_op( unsigned long op,
@@ -126,13 +121,15 @@ do_hvm_op(extern long do_kexec_op ( -unsigned long op, +unsigned int op,XEN_GUEST_HANDLE_PARAM (void ) uarg);extern long do_xsm_op (XEN_GUEST_HANDLE_PARAM (void ) u_xsm_op); +long do_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg); +#ifdef CONFIG_ARGO extern long do_argo_op( unsigned int cmd,
@@ -145,9 +142,6 @@ extern long do_argo_op(extern long do_xenoprof_op (int op,XEN_GUEST_HANDLE_PARAM (void ) arg); -extern long -do_xenpmu_op (unsigned int op,XEN_GUEST_HANDLE_PARAM (xen_pmu_params_t) arg); - extern long do_dm_op( domid_t domid,
@@ -198,21 +192,25 @@ compat_sched_op(extern int compat_set_timer_op ( -u32 lo, -s32 hi); +uint32_t lo, +int32_t hi);extern int compat_xsm_op (XEN_GUEST_HANDLE_PARAM (void ) op); -extern int compat_kexec_op (unsigned long op,XEN_GUEST_HANDLE_PARAM (void ) uarg); +extern int compat_kexec_op (unsigned int op,XEN_GUEST_HANDLE_PARAM (void ) uarg);DEFINE_XEN_GUEST_HANDLE (multicall_entry_compat_t);extern int compat_multicall (XEN_GUEST_HANDLE_PARAM (multicall_entry_compat_t) call_list,uint32_t nr_calls); +int compat_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg); + +int compat_platform_op (XEN_GUEST_HANDLE_PARAM (void ) u_xenpf_op); +#ifdef CONFIG_ARGO -extern long compat_argo_op( +extern int compat_argo_op( unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, XEN_GUEST_HANDLE_PARAM(void) arg2,
Signature · 1 line
2.26.2
Re: [PATCH 03/12] xen: harmonize return types of hypercall handlers
Quoted 18 lines · Juergen Gross
Today most hypercall handlers have a return type of long, while the compat ones return an int. There are a few exceptions from that rule, however.
Get rid of the exceptions by letting compat handlers always return int and others always return long.
For the compat hvm case use eax instead of rax for the stored result as it should have been from the beginning.
Additionally move some prototypes to
include/asm-x86/hypercall.has they are x86 specific. Move thedo_physdev_op()prototype from both architecture dependant headers to the common one. Move thecompat_platform_op()prototype to the common header.Switch some non style compliant types (u32, s32, s64) to style compliant ones.
Quoted 8 lines
Rename
paging_domctl_continuation()todo_paging_domctl_cont()and add a matching define for the associated hypercall.Make
do_callback_op()andcompat_callback_op()more similar by adding the const attribute to compat_callback_op()'s 2nd parameter.The
do_platform_op()prototype needs to be modified in order to better match its compat variant.
"Better" in what direction? So far both have been using typed handles, which I consider better than void ones. You also don't seem to have had a reason to switch e.g. multicall or dm_op, where (different) typed handles are also in use. So I wonder what the reason for this change is.
Quoted 2 lines
Change the type of the cmd parameter for [do|compat]_kexec_op() to unsigned int, as this is more appropriate for the compat case.
The change for the compat case is fine, but for native you change behavior for callers passing values equaling valid KEXEC_CMD_* modulo 2³².
Quoted 4 lines
xen/arch/x86/pv/misc-hypercalls.c +0
-0@@ -28,12 +28,16 @@ long do_set_debugreg(int reg, unsigned long value)return set_debugreg (current, reg, value);
xen/arch/x86/pv/misc-hypercalls.c long do_set_debugreg(int reg, unsigned long value) · line 33
Quoted 10 lines
xen/arch/x86/pv/misc-hypercalls.c +2
-5} -unsigned long do_get_debugreg (int reg) +long do_get_debugreg (int reg) { -unsigned long val; -int res =x86emul_read_dr (reg, &val,NULL ); - -return res ==X86EMUL_OKAY ? val : -ENODEV ; +/* Avoid undefined behavior due to casting an unsigned long to long. */
Nit: unsigned -> signed conversion is implementation-defined, not undefined.
Quoted 4 lines
xen/common/argo.c +0
-0@@ -2207,13 +2207,13 @@ do_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, }
xen/common/argo.c do_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, · line 2213
Quoted 10 lines
xen/common/argo.c +4
-4#ifdef CONFIG_COMPAT -long -compat_argo_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg1, -XEN_GUEST_HANDLE_PARAM (void ) arg2,unsigned long arg3, -unsigned long arg4) +int compat_argo_op (unsigned int cmd, +XEN_GUEST_HANDLE_PARAM (void ) arg1, +XEN_GUEST_HANDLE_PARAM (void ) arg2, +unsigned long arg3,unsigned long arg4)
Strictly speaking arg3 and arg4 also ought to be unsigned int here. But that's perhaps for a separate patch at another time.
Jan
Re: [PATCH 03/12] xen: harmonize return types of hypercall handlers
Quoted 25 lines · Jan Beulich
Quoted 24 lines · Juergen Gross
Today most hypercall handlers have a return type of long, while the compat ones return an int. There are a few exceptions from that rule, however.
Get rid of the exceptions by letting compat handlers always return int and others always return long.
For the compat hvm case use eax instead of rax for the stored result as it should have been from the beginning.
Additionally move some prototypes to
include/asm-x86/hypercall.has they are x86 specific. Move thedo_physdev_op()prototype from both architecture dependant headers to the common one. Move thecompat_platform_op()prototype to the common header.Switch some non style compliant types (u32, s32, s64) to style compliant ones.
Rename
paging_domctl_continuation()todo_paging_domctl_cont()and add a matching define for the associated hypercall.Make
do_callback_op()andcompat_callback_op()more similar by adding the const attribute to compat_callback_op()'s 2nd parameter.
Quoted 8 lines
Quoted 2 lines
The
do_platform_op()prototype needs to be modified in order to better match its compat variant."Better" in what direction? So far both have been using typed handles, which I consider better than void ones. You also don't seem to have had a reason to switch e.g. multicall or
dm_op, where (different) typed handles are also in use. So I wonder what the reason for this change is.
I had some problems making this build. Before my patches platform_hypercall.c didn't include the header with the prototype, so there was no mismatch detected.
Thanks for the pointers above. dm_op is no good example, as the compat variant is explicitly a different implementation compared to the normal one. But with the multicall example I can have another try converting platform_op to a type safe variant using non-void handles.
Quoted 7 lines
Quoted 3 lines
Change the type of the cmd parameter for [do|compat]_kexec_op() to unsigned int, as this is more appropriate for the compat case.
The change for the compat case is fine, but for native you change behavior for callers passing values equaling valid KEXEC_CMD_* modulo 2³².
TBH, I don't think this is really a problem. Or do you think there really is a user of this interface relying on a -ENOSYS in this case?
Quoted 7 lines
Quoted 7 lines
xen/arch/x86/pv/misc-hypercalls.c +0
-0@@ -28,12 +28,16 @@ long do_set_debugreg(int reg, unsigned long value)return set_debugreg (current, reg, value); }
Quoted 11 lines
xen/arch/x86/pv/misc-hypercalls.c long do_set_debugreg(int reg, unsigned long value) · line 33
Quoted 8 lines
xen/arch/x86/pv/misc-hypercalls.c +2
-5-unsigned long do_get_debugreg (int reg) +long do_get_debugreg (int reg) { -unsigned long val; -int res =x86emul_read_dr (reg, &val,NULL ); - -return res ==X86EMUL_OKAY ? val : -ENODEV ; +/* Avoid undefined behavior due to casting an unsigned long to long. */ Nit: unsigned -> signed conversion is implementation-defined, not undefined.
Okay, will change the comment.
Quoted 6 lines
Quoted 6 lines
xen/common/argo.c +0
-0@@ -2207,13 +2207,13 @@ do_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, }
Quoted 12 lines
xen/common/argo.c do_argo_op(unsigned int cmd, XEN_GUEST_HANDLE_PARAM(void) arg1, · line 2213
Quoted 9 lines
xen/common/argo.c +4
-4#ifdef CONFIG_COMPAT -long -compat_argo_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg1, -XEN_GUEST_HANDLE_PARAM (void ) arg2,unsigned long arg3, -unsigned long arg4) +int compat_argo_op (unsigned int cmd, +XEN_GUEST_HANDLE_PARAM (void ) arg1, +XEN_GUEST_HANDLE_PARAM (void ) arg2, +unsigned long arg3,unsigned long arg4)Strictly speaking arg3 and arg4 also ought to be unsigned int here. But that's perhaps for a separate patch at another time.
I'd rather leave it as is, as this way I can use the same definition for both cases in patch 6. And I don't see how anything could go wrong this way, as expanding a 32-bit unsigned value to 64 bits is in no way ambiguous.
Juergen
Re: [PATCH 03/12] xen: harmonize return types of hypercall handlers
Quoted 5 lines · Juergen Gross
Quoted 4 lines · Jan Beulich
Quoted 2 lines · Juergen Gross
Change the type of the cmd parameter for [do|compat]_kexec_op() to unsigned int, as this is more appropriate for the compat case.
Quoted 7 lines
Quoted 3 lines
The change for the compat case is fine, but for native you change behavior for callers passing values equaling valid KEXEC_CMD_* modulo 2³².
TBH, I don't think this is really a problem. Or do you think there really is a user of this interface relying on a
-ENOSYSin this case?
That's a secondary consideration of mine only. The primary one is that invoking with an invalid sub-op should fail, such that in the future we can assign meaning to the upper bits, if need be. See their use for continuations in memory-op, for example.
Jan
Re: [PATCH 03/12] xen: harmonize return types of hypercall handlers
Quoted 10 lines · Jan Beulich
Quoted 9 lines · Juergen Gross
Quoted 7 lines · Jan Beulich
Quoted 2 lines · Juergen Gross
Change the type of the cmd parameter for [do|compat]_kexec_op() to unsigned int, as this is more appropriate for the compat case.
The change for the compat case is fine, but for native you change behavior for callers passing values equaling valid KEXEC_CMD_* modulo 2³².
Quoted 8 lines
Quoted 3 lines
TBH, I don't think this is really a problem. Or do you think there really is a user of this interface relying on a
-ENOSYSin this case?That's a secondary consideration of mine only. The primary one is that invoking with an invalid sub-op should fail, such that in the future we can assign meaning to the upper bits, if need be. See their use for continuations in memory-op, for example.
But this would mean to exclude such usage of upper bits for 32 bit systems in case this difference should matter.
Juergen
PATCH 4/12 xen/x86: modify hvm_memory_op() prototype
In reply to Juergen Grosshvm_memory_op() should take an unsigned long as cmd, like do_memory_op().
As hvm_memory_op() is basically just calling do_memory_op() (or compat_memory_op()) passing through the parameters the cmd parameter should have no smaller size than that of the called functions.
--- xen/arch/x86/hvm/hypercall.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
xen/arch/x86/hvm/hypercall.c +1 -1
@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h> -static long hvm_memory_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;
Signature · 1 line
2.26.2
Re: [PATCH 04/12] xen/x86: modify hvm_memory_op() prototype
Quoted 8 lines · Juergen Gross
hvm_memory_op()should take an unsigned long as cmd, likedo_memory_op().As
hvm_memory_op()is basically just callingdo_memory_op()(orcompat_memory_op()) passing through the parameters the cmd parameter should have no smaller size than that of the called functions.Signed-off-by Juergen Gross <[email protected]>
Nevertheless ...
xen/arch/x86/hvm/hypercall.c line 36
Quoted 10 lines
xen/arch/x86/hvm/hypercall.c +1
-1@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h> -static long hvm_memory_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;
... I think this would even better be dealt with by splitting the function into a native one (using unsigned long) and a compat one (using unsigned int).
Jan
Re: [PATCH 04/12] xen/x86: modify hvm_memory_op() prototype
Quoted 20 lines · Jan Beulich
Quoted 8 lines · Juergen Gross
hvm_memory_op()should take an unsigned long as cmd, likedo_memory_op().As
hvm_memory_op()is basically just callingdo_memory_op()(orcompat_memory_op()) passing through the parameters the cmd parameter should have no smaller size than that of the called functions.Signed-off-by Juergen Gross <[email protected]>Reviewed-by Jan Beulich <[email protected]>Nevertheless ...
Quoted 6 lines
xen/arch/x86/hvm/hypercall.c +0
-0@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h>
Quoted 8 lines
xen/arch/x86/hvm/hypercall.c line 36
Quoted 4 lines
xen/arch/x86/hvm/hypercall.c +1
-1-static long hvm_memory_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;... I think this would even better be dealt with by splitting the function into a native one (using unsigned long) and a compat one (using unsigned int).
Why? In 32-bit case the value is naturally limited to 32 bits width zero-extending perfectly fine to unsigned long.
Otherwise I couldn't use the same definition later.
Juergen
Re: [PATCH 04/12] xen/x86: modify hvm_memory_op() prototype
Quoted 30 lines · Juergen Gross
Quoted 20 lines · Jan Beulich
Quoted 8 lines · Juergen Gross
hvm_memory_op()should take an unsigned long as cmd, likedo_memory_op().As
hvm_memory_op()is basically just callingdo_memory_op()(orcompat_memory_op()) passing through the parameters the cmd parameter should have no smaller size than that of the called functions.Signed-off-by Juergen Gross <[email protected]>Reviewed-by Jan Beulich <[email protected]>Nevertheless ...
Quoted 6 lines
xen/arch/x86/hvm/hypercall.c +0
-0@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h> Quoted 8 lines
xen/arch/x86/hvm/hypercall.c line 36
Quoted 4 lines
xen/arch/x86/hvm/hypercall.c +1
-1-static long hvm_memory_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;... I think this would even better be dealt with by splitting the function into a native one (using unsigned long) and a compat one (using unsigned int).
Quoted 2 lines
Why? In 32-bit case the value is naturally limited to 32 bits width zero-extending perfectly fine to unsigned long.
It all ends up working fine, yes. Else I wouldn't have given R-b. But the .compat slot of the hypercall table really should use a prototype without unsigned long, and then the calls wouldn't zero-extend the arguments anymore. And then the declaration would be wrong, as then it would need to be the callee to zero-extend if it wants to use 64-bit values.
Quoted 1 line
Otherwise I couldn't use the same definition later.
Right. And this will be less of a problem once the function pointer tables are gone, as then the compiler sees the real parameter types for the individual functions.
Jan
Re: [PATCH 04/12] xen/x86: modify hvm_memory_op() prototype
Quoted 41 lines · Jan Beulich
Quoted 30 lines · Juergen Gross
Quoted 20 lines · Jan Beulich
Quoted 8 lines · Juergen Gross
hvm_memory_op()should take an unsigned long as cmd, likedo_memory_op().As
hvm_memory_op()is basically just callingdo_memory_op()(orcompat_memory_op()) passing through the parameters the cmd parameter should have no smaller size than that of the called functions.Signed-off-by Juergen Gross <[email protected]>Reviewed-by Jan Beulich <[email protected]>Nevertheless ...
Quoted 6 lines
xen/arch/x86/hvm/hypercall.c +0
-0@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h> Quoted 8 lines
xen/arch/x86/hvm/hypercall.c line 36
Quoted 4 lines
xen/arch/x86/hvm/hypercall.c +1
-1-static long hvm_memory_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;... I think this would even better be dealt with by splitting the function into a native one (using unsigned long) and a compat one (using unsigned int).
Quoted 2 lines
Why? In 32-bit case the value is naturally limited to 32 bits width zero-extending perfectly fine to unsigned long.
It all ends up working fine, yes. Else I wouldn't have given R-b. But the .compat slot of the hypercall table really should use a prototype without unsigned long, and then the calls wouldn't zero-extend the arguments anymore. And then the declaration would be wrong, as then it would need to be the callee to zero-extend if it wants to use 64-bit values.
Quoted 5 lines
Quoted 1 line
Otherwise I couldn't use the same definition later.
Right. And this will be less of a problem once the function pointer tables are gone, as then the compiler sees the real parameter types for the individual functions.
Okay, I understand that.
I'd prefer to do that as a followup patch (series) then.
Juergen
PATCH 5/12 xen: don't include asm/hypercall.h from C sources
In reply to Juergen GrossInstead of including asm/hypercall.h always use xen/hypercall.h. Additionally include xen/hypercall.h from all sources containing a hypercall handler.
This prepares for generating the handlers' prototypes at build time.
Add a guard in asm/hypercall.h to catch direct inclusion.
--- xen/arch/arm/hvm.c | 3 +-- xen/arch/arm/platform_hypercall.c | 1 + xen/arch/x86/cpu/vpmu.c | 1 + xen/arch/x86/mm.c | 1 - xen/arch/x86/platform_hypercall.c | 1 + xen/arch/x86/pv/iret.c | 1 + xen/arch/x86/traps.c | 2 +- xen/arch/x86/x86_64/compat/mm.c | 1 + xen/arch/x86/x86_64/mm.c | 1 - xen/arch/x86/x86_64/platform_hypercall.c | 4 ++-- xen/common/compat/grant_table.c | 1 + xen/common/compat/multicall.c | 2 +- xen/common/event_channel.c | 1 + xen/common/grant_table.c | 1 + xen/common/multicall.c | 1 + xen/include/asm-arm/hypercall.h | 4 ++++ xen/include/asm-x86/hypercall.h | 4 ++++ 17 files changed, 22 insertions(+), 8 deletions(-)
xen/arch/arm/hvm.c +1 -2
@@ -20,6 +20,7 @@#include <xen/lib.h> #include <xen/errno.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/sched.h> #include <xen/monitor.h>
@@ -29,8 +30,6 @@#include <public/hvm/params.h> #include <public/hvm/hvm_op.h> -#include <asm/hypercall.h> -static int hvm_allow_set_param (const struct domain *d,unsigned int param) { switch ( param )
xen/arch/arm/platform_hypercall.c +1 -0
@@ -9,6 +9,7 @@#include <xen/types.h> #include <xen/sched.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/spinlock.h> #include <public/platform.h> #include <xsm/xsm.h>
xen/arch/x86/cpu/vpmu.c +1 -0
@@ -21,6 +21,7 @@#include <xen/xenoprof.h> #include <xen/event.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/cpu.h> #include <xen/param.h> #include <asm/regs.h>
xen/arch/x86/mm.c +0 -1
@@ -129,7 +129,6 @@#include <asm/ldt.h> #include <asm/x86_emulate.h> #include <asm/e820.h> -#include <asm/hypercall.h> #include <asm/shared.h> #include <asm/mem_sharing.h> #include <public/memory.h>
xen/arch/x86/platform_hypercall.c +1 -0
@@ -17,6 +17,7 @@#include <xen/console.h> #include <xen/iocap.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/acpi.h> #include <xen/efi.h> #include <xen/cpu.h>
xen/arch/x86/pv/iret.c +1 -0
@@ -18,6 +18,7 @@*/ #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/lib.h> #include <xen/sched.h>
xen/arch/x86/traps.c +1 -1
@@ -29,6 +29,7 @@#include <xen/lib.h> #include <xen/err.h> #include <xen/errno.h> +#include <xen/hypercall.h> #include <xen/mm.h> #include <xen/param.h> #include <xen/console.h>
@@ -70,7 +71,6 @@#include <asm/x86_emulate.h> #include <asm/traps.h> #include <asm/hvm/vpt.h> -#include <asm/hypercall.h> #include <asm/mce.h> #include <asm/apic.h> #include <asm/mc146818rtc.h>
xen/arch/x86/x86_64/compat/mm.c +1 -0
@@ -1,4 +1,5 @@#include <xen/event.h> +#include <xen/hypercall.h> #include <xen/mem_access.h> #include <xen/multicall.h> #include <compat/memory.h>
xen/arch/x86/x86_64/mm.c +0 -1
@@ -32,7 +32,6 @@ EMIT_FILE;#include <asm/page.h> #include <asm/flushtlb.h> #include <asm/fixmap.h> -#include <asm/hypercall.h> #include <asm/msr.h> #include <asm/pv/domain.h> #include <asm/setup.h>
xen/arch/x86/x86_64/platform_hypercall.c +2 -2
@@ -41,8 +41,8 @@ CHECK_pf_resource_entry;#undef xen_pf_resource_entry#define COMPAT -#define _XEN_GUEST_HANDLE (t) XEN_GUEST_HANDLE(t) -#define _XEN_GUEST_HANDLE_PARAM (t) XEN_GUEST_HANDLE_PARAM(t) +#undef guest_handle_okay +#define guest_handle_okay compat_handle_okaytypedef int ret_t ;#include "../platform_hypercall.c"
xen/common/compat/grant_table.c +1 -0
@@ -4,6 +4,7 @@*/ #include <compat/grant_table.h> +#include <xen/hypercall.h> #define xen_grant_entry_v1 grant_entry_v1 CHECK_grant_entry_v1;
xen/common/compat/multicall.c +1 -1
@@ -4,6 +4,7 @@EMIT_FILE ; +#include <xen/hypercall.h> #include <xen/types.h> #include <xen/multicall.h> #include <xen/trace.h>
@@ -19,7 +20,6 @@ static inline void xlat_multicall_entry(struct mc_state *mcs) mcs->compat_call.args[i] = mcs->call.args[i]; } -DEFINE_XEN_GUEST_HANDLE (multicall_entry_compat_t);#define multicall_entry compat_multicall_entry#define multicall_entry_t multicall_entry_compat_t#define do_multicall_call compat_multicall_call
xen/common/event_channel.c +1 -0
@@ -24,6 +24,7 @@#include <xen/iocap.h> #include <xen/compat.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/keyhandler.h> #include <asm/current.h>
xen/common/grant_table.c +1 -0
@@ -33,6 +33,7 @@#include <xen/trace.h> #include <xen/grant_table.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/domain_page.h> #include <xen/iommu.h> #include <xen/paging.h>
xen/common/multicall.c +1 -0
@@ -9,6 +9,7 @@#include <xen/event.h> #include <xen/multicall.h> #include <xen/guest_access.h> +#include <xen/hypercall.h> #include <xen/perfc.h> #include <xen/trace.h> #include <asm/current.h>
xen/include/asm-arm/hypercall.h +4 -0
@@ -1,3 +1,7 @@ +#ifndef __XEN_HYPERCALL_H__ +#error "asm/hypercall.h should not be included directly - include xen/hypercall.h instead" +#endif +#ifndef __ASM_ARM_HYPERCALL_H__#define __ASM_ARM_HYPERCALL_H__
xen/include/asm-x86/hypercall.h +4 -0
@@ -2,6 +2,10 @@* asm-x86/hypercall.h */ +#ifndef __XEN_HYPERCALL_H__ +#error "asm/hypercall.h should not be included directly - include xen/hypercall.h instead" +#endif +#ifndef __ASM_X86_HYPERCALL_H__#define __ASM_X86_HYPERCALL_H__
Signature · 1 line
2.26.2
Re: [PATCH 05/12] xen: don't include asm/hypercall.h from C sources
xen/arch/x86/x86_64/mm.c EMIT_FILE; · line 37
Quoted 10 lines · Juergen Gross
xen/arch/x86/x86_64/mm.c +0
-1@@ -32,7 +32,6 @@ EMIT_FILE;#include <asm/page.h> #include <asm/flushtlb.h> #include <asm/fixmap.h> -#include <asm/hypercall.h> #include <asm/msr.h> #include <asm/pv/domain.h> #include <asm/setup.h>
Afaict xen/hypercall.h also doesn't need including here. But I can agree that this is at best a partly related change.
xen/arch/x86/x86_64/platform_hypercall.c CHECK_pf_resource_entry; · line 45
Quoted 10 lines
xen/arch/x86/x86_64/platform_hypercall.c +2
-2@@ -41,8 +41,8 @@ CHECK_pf_resource_entry;#undef xen_pf_resource_entry#define COMPAT -#define _XEN_GUEST_HANDLE (t) XEN_GUEST_HANDLE(t) -#define _XEN_GUEST_HANDLE_PARAM (t) XEN_GUEST_HANDLE_PARAM(t) +#undef guest_handle_okay +#define guest_handle_okay compat_handle_okay
Is this a change that slipped in here from patch 3, accompanying
-#define do_platform_op(x) compat_platform_op(_##x)
+#define do_platform_op(x) compat_platform_op(x)
there? Or does that change belong here? I have to admit anyway that it's not entirely clear to me why this adjustment needs making.
xen/common/compat/grant_table.c line 7
Quoted 7 lines
xen/common/compat/grant_table.c +1
-0@@ -4,6 +4,7 @@*/ #include <compat/grant_table.h> +#include <xen/hypercall.h>
Nit: Generally compat/*.h go last.
Jan
Re: [PATCH 05/12] xen: don't include asm/hypercall.h from C sources
Quoted 14 lines · Jan Beulich
xen/arch/x86/x86_64/mm.c EMIT_FILE; · line 37
Quoted 10 lines · Juergen Gross
xen/arch/x86/x86_64/mm.c +0
-1@@ -32,7 +32,6 @@ EMIT_FILE;#include <asm/page.h> #include <asm/flushtlb.h> #include <asm/fixmap.h> -#include <asm/hypercall.h> #include <asm/msr.h> #include <asm/pv/domain.h> #include <asm/setup.h> Afaict
xen/hypercall.halso doesn't need including here. But I can agree that this is at best a partly related change.
Will check and delete it in case not needed.
Quoted 17 lines
xen/arch/x86/x86_64/platform_hypercall.c CHECK_pf_resource_entry; · line 45
Quoted 11 lines
xen/arch/x86/x86_64/platform_hypercall.c +2
-2@@ -41,8 +41,8 @@ CHECK_pf_resource_entry;#undef xen_pf_resource_entry#define COMPAT -#define _XEN_GUEST_HANDLE (t) XEN_GUEST_HANDLE(t) -#define _XEN_GUEST_HANDLE_PARAM (t) XEN_GUEST_HANDLE_PARAM(t) +#undef guest_handle_okay +#define guest_handle_okay compat_handle_okayIs this a change that slipped in here from patch 3, accompanying
(quoted diff) +1
-1-#define do_platform_op(x) compat_platform_op(_##x) +#define do_platform_op(x) compat_platform_op(x)
Quoted 2 lines
there? Or does that change belong here? I have to admit anyway that it's not entirely clear to me why this adjustment needs making.
Oh yes, this is a leftover to make things work with prototypes being made seen in platform_hypercall.c.
Quoted 10 lines
xen/common/compat/grant_table.c line 7
Quoted 8 lines
xen/common/compat/grant_table.c +1
-0@@ -4,6 +4,7 @@*/ #include <compat/grant_table.h> +#include <xen/hypercall.h> Nit: Generally compat/*.h go last.
Okay.
Juergen
PATCH 6/12 xen: generate hypercall interface related code
In reply to Juergen GrossInstead of repeating similar data multiple times use a single source file and a generator script for producing prototypes and call sequences of the hypercalls.
As the script already knows the number of parameters used add generating a macro for populating an array with the number of parameters per hypercall.
--- .gitignore | 2 + xen/Makefile | 10 ++ xen/include/hypercall-defs.c | 272 ++++++++++++++++++++++++++++++++++ xen/scripts/gen_hypercall.awk | 228 ++++++++++++++++++++++++++++ 4 files changed, 512 insertions(+) create mode 100644 xen/include/hypercall-defs.c create mode 100644 xen/scripts/gen_hypercall.awk
.gitignore +2 -0
@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.h xen/include/compat/* xen/include/config/ xen/include/generated/ +xen/include//hypercall-defs.i xen/include/public/public xen/include/xen/*.new xen/include/xen/acm_policy.h xen/include/xen/compile.h +xen/include/xen/hypercall-defs.h xen/include/xen/lib/x86/cpuid-autogen.h xen/test/livepatch/config.h xen/test/livepatch/expect_config.h
xen/Makefile +10 -0
@@ -383,6 +383,7 @@ _clean: delete-unfresh-files -o -name "*.gcno" -o -name ".*.cmd" -o -name "lib.a" \) -exec rm -f {} \; rm -f include/asm $(TARGET) $(TARGET).gz $(TARGET).efi $(TARGET).efi.map $(TARGET)-syms $(TARGET)-syms.map *~ core rm -f asm-offsets.s include/asm-*/asm-offsets.h + rm -f include/xen/hypercall-defs.h include/hypercall-defs.i rm -f .banner .allconfig.tmp .PHONY: _distclean
@@ -405,6 +406,7 @@ $(TARGET): delete-unfresh-files $(MAKE) -f $(BASEDIR)/Rules.mk -C include $(MAKE) -f $(BASEDIR)/Rules.mk -C arch/$(TARGET_ARCH) include $(MAKE) -f $(BASEDIR)/Rules.mk include/asm-$(TARGET_ARCH)/asm-offsets.h + $(MAKE) -f $(BASEDIR)/Rules.mk include/xen/hypercall-defs.h $(MAKE) -f $(BASEDIR)/Rules.mk -C arch/$(TARGET_ARCH) $@ # drivers/char/console.o contains static banner/compile info. Blow it away.
@@ -466,6 +468,14 @@ include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s echo ""; \ echo "#endif") <$< >$@ +quiet_cmd_genhyp = GEN $@ +define cmd_genhyp + awk -f scripts/gen_hypercall.awk <$< >$@ +endef + +include/xen/hypercall-defs.h: include/hypercall-defs.i scripts/gen_hypercall.awk FORCE + $(call if_changed,genhyp) + SUBDIRS = xsm arch/$(TARGET_ARCH) common drivers lib test define all_sources ( find include/asm-$(TARGET_ARCH) -name '*.h' -print; \
xen/include/hypercall-defs.c new file +272 -0
272 lines not on this page · raw message
xen/scripts/gen_hypercall.awk new file +228 -0
228 lines not on this page · raw message
Signature · 1 line
2.26.2
Re: [PATCH 06/12] xen: generate hypercall interface related code
Quoted 7 lines · Juergen Gross
Instead of repeating similar data multiple times use a single source file and a generator script for producing prototypes and call sequences of the hypercalls.
As the script already knows the number of parameters used add generating a macro for populating an array with the number of parameters per hypercall.
Isn't that array intended to go away?
.gitignore xen/include/asm-x86/asm-macros.h · line 335
Quoted 7 lines
.gitignore +1
-0@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.h xen/include/compat/* xen/include/config/ xen/include/generated/ +xen/include//hypercall-defs.i
Nit: Stray double slash (unless this has a meaning I'm unaware of).
Yet then I wonder: Shouldn't *.i be among the patterns at the top of the file, like *.o is?
.gitignore include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s · line 477
Quoted 11 lines
.gitignore +7
-0@@ -466,6 +468,14 @@ include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s echo ""; \ echo "#endif") <$< >$@ +quiet_cmd_genhyp = GEN $@ +define cmd_genhyp + awk -f scripts/gen_hypercall.awk <$< >$@ +endef + +include/xen/hypercall-defs.h: include/hypercall-defs.i scripts/gen_hypercall.awk FORCE + $(call if_changed,genhyp)
As per patch 5 there are quite a few sources including xen/hypercall.h and hence (in one of the next patches) the header generated here. If this one gets re-generated for a benign reason (i.e. without changing the header), all dependents will get rebuilt for no reason. Use $(move-if-changed ...)?
Quoted 41 lines
.gitignore +41
-0+prefix: do +set_timer_op(s_time_t timeout) +console_io(unsigned int cmd, unsigned int count, char *buffer) +vm_assist(unsigned int cmd, unsigned int type) +event_channel_op(int cmd, void *arg) +mmuext_op(mmuext_op_t *uops, unsigned int count, unsigned int *pdone, unsigned int foreigndom) +multicall(multicall_entry_t *call_list, unsigned int nr_calls) +#ifdef CONFIG_PV +mmu_update(mmu_update_t *ureqs, unsigned int count, unsigned int *pdone, unsigned int foreigndom) +stack_switch(unsigned long ss, unsigned long esp) +fpu_taskswitch(int set) +set_debugreg(int reg, unsigned long value) +get_debugreg(int reg) +set_segment_base(unsigned int which, unsigned long base) +mca(xen_mc_t *u_xen_mc) +set_trap_table(const_trap_info_t *traps) +set_gdt(xen_ulong_t *frame_list, unsigned int entries) +set_callbacks(unsigned long event_address, unsigned long failsafe_address, unsigned long syscall_address) +update_descriptor(uint64_t gaddr, seg_desc_t desc) +update_va_mapping(unsigned long va, uint64_t val64, unsigned long flags) +update_va_mapping_otherdomain(unsigned long va, uint64_t val64, unsigned long flags, domid_t domid) +#endif +#ifdef CONFIG_IOREQ_SERVER +dm_op(domid_t domid, unsigned int nr_bufs, xen_dm_op_buf_t *bufs) +#endif +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +sysctl(xen_sysctl_t *u_sysctl) +domctl(xen_domctl_t *u_domctl) +paging_domctl_cont(xen_domctl_t *u_domctl) +#endif +#ifdef CONFIG_HVM +hvm_op(unsigned long op, void *arg) +#endif +#ifdef CONFIG_HYPFS +hypfs_op(unsigned int cmd, const char *arg1, unsigned long arg2, void *arg3, unsigned long arg4) +#endif +#ifdef CONFIG_X86 +xenpmu_op(unsigned int op, xen_pmu_params_t *arg) +#endif + +#ifdef CONFIG_PV
.gitignore include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s
Quoted 10 lines
.gitignore +10
-0+caller: pv64 +#ifdef CONFIG_PV32 +caller: pv32 +#endif +#endif +#ifdef CONFIG_HVM +caller: hvm64 +#ifdef CONFIG_COMPAT +caller: hvm32 +#endif
HVM selects COMPAT, so I don't see a point in this inner conditional.
Quoted 62 lines
.gitignore +62
-0+#endif +#ifdef CONFIG_ARM +caller: arm +#endif + +table: pv32 pv64 hvm32 hvm64 arm +set_trap_table compat do - - - +mmu_update do do - - - +set_gdt compat do - - - +stack_switch do do - - - +set_callbacks compat do - - - +fpu_taskswitch do do - - - +sched_op_compat do do - - dep +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +platform_op compat do compat do do +#endif +set_debugreg do do - - - +get_debugreg do do - - - +update_descriptor compat do - - - +memory_op compat do hvm hvm do +multicall compat do compat do do +update_va_mapping compat do - - - +set_timer_op compat do compat do - +event_channel_op_compat do do - - dep +xen_version compat do compat do do +console_io do do do do do +physdev_op_compat compat do - - dep +#if defined(CONFIG_GRANT_TABLE) || defined(CONFIG_PV_SHIM) +grant_table_op compat do hvm hvm do +#endif +vm_assist do do do do do +update_va_mapping_otherdomain compat do - - - +iret compat do - - - +vcpu_op compat do compat do do +set_segment_base do do - - - +#ifdef CONFIG_PV +mmuext_op compat do compat do - +#endif +xsm_op compat do compat do do +nmi_op compat do - - - +sched_op compat do compat do do +callback_op compat do - - - +#ifdef CONFIG_XENOPROF +xenoprof_op compat do - - - +#endif +event_channel_op do do do do do +physdev_op compat do hvm hvm do +#ifdef CONFIG_HVM +hvm_op do do do do do +#endif +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +sysctl do do do do do +domctl do do do do do +#endif +#ifdef CONFIG_KEXEC +kexec_op compat do - - - +#endif +tmem_op - - - - - +#ifdef CONFIG_ARGO +argo_op compat do compat do do +#endif +xenpmu_op do do do do -
.gitignore
Quoted 10 lines
.gitignore +10
-0+#ifdef CONFIG_IOREQ_SERVER +dm_op compat do compat do do +#endif +#ifdef CONFIG_HYPFS +hypfs_op do do do do do +#endif +mca do do - - - +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +paging_domctl_cont do do do do - +#endif
I assume the intention here is to sort by hypercall number. This results in 3 PV_SHIM_EXCLUSIVE conditionals when one might do. Just a remark for consideration, not strictly a request to change anything.
Jan
Re: [PATCH 06/12] xen: generate hypercall interface related code
Quoted 10 lines · Jan Beulich
Quoted 7 lines · Juergen Gross
Instead of repeating similar data multiple times use a single source file and a generator script for producing prototypes and call sequences of the hypercalls.
As the script already knows the number of parameters used add generating a macro for populating an array with the number of parameters per hypercall.
Isn't that array intended to go away?
I thought so, yes, but on Arm there is a case where it is needed.
So generating it from the available data is the sensible thing to do IMO.
Quoted 10 lines
.gitignore xen/include/asm-x86/asm-macros.h · line 335
Quoted 8 lines
.gitignore +1
-0@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.h xen/include/compat/* xen/include/config/ xen/include/generated/ +xen/include//hypercall-defs.iNit: Stray double slash (unless this has a meaning I'm unaware of).
Oh, right. No special meaning AFAIK.
Quoted 2 lines
Yet then I wonder: Shouldn't *.i be among the patterns at the top of the file, like *.o is?
Yes, I can do that. Probably via a separate patch then.
Quoted 4 lines
Quoted 4 lines
.gitignore +0
-0@@ -466,6 +468,14 @@ include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s echo ""; \ echo "#endif") <$< >$@
Quoted 13 lines
.gitignore include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s · line 477
Quoted 7 lines
.gitignore +7
-0+quiet_cmd_genhyp = GEN $@ +define cmd_genhyp + awk -f scripts/gen_hypercall.awk <$< >$@ +endef + +include/xen/hypercall-defs.h: include/hypercall-defs.i scripts/gen_hypercall.awk FORCE + $(call if_changed,genhyp)As per patch 5 there are quite a few sources including
xen/hypercall.hand hence (in one of the next patches) the header generated here. If this one gets re-generated for a benign reason (i.e. without changing the header), all dependents will get rebuilt for no reason. Use $(move-if-changed ...)?
The reasons for re-generating are quite limited. The most probable case is a .config change, which will trigger quite some rebuild anyway.
If you think its is really worth the effort, I can use move-if-changed.
Quoted 46 lines
Quoted 46 lines
.gitignore +45
-0+prefix: do +set_timer_op(s_time_t timeout) +console_io(unsigned int cmd, unsigned int count, char *buffer) +vm_assist(unsigned int cmd, unsigned int type) +event_channel_op(int cmd, void *arg) +mmuext_op(mmuext_op_t *uops, unsigned int count, unsigned int *pdone, unsigned int foreigndom) +multicall(multicall_entry_t *call_list, unsigned int nr_calls) +#ifdef CONFIG_PV +mmu_update(mmu_update_t *ureqs, unsigned int count, unsigned int *pdone, unsigned int foreigndom) +stack_switch(unsigned long ss, unsigned long esp) +fpu_taskswitch(int set) +set_debugreg(int reg, unsigned long value) +get_debugreg(int reg) +set_segment_base(unsigned int which, unsigned long base) +mca(xen_mc_t *u_xen_mc) +set_trap_table(const_trap_info_t *traps) +set_gdt(xen_ulong_t *frame_list, unsigned int entries) +set_callbacks(unsigned long event_address, unsigned long failsafe_address, unsigned long syscall_address) +update_descriptor(uint64_t gaddr, seg_desc_t desc) +update_va_mapping(unsigned long va, uint64_t val64, unsigned long flags) +update_va_mapping_otherdomain(unsigned long va, uint64_t val64, unsigned long flags, domid_t domid) +#endif +#ifdef CONFIG_IOREQ_SERVER +dm_op(domid_t domid, unsigned int nr_bufs, xen_dm_op_buf_t *bufs) +#endif +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +sysctl(xen_sysctl_t *u_sysctl) +domctl(xen_domctl_t *u_domctl) +paging_domctl_cont(xen_domctl_t *u_domctl) +#endif +#ifdef CONFIG_HVM +hvm_op(unsigned long op, void *arg) +#endif +#ifdef CONFIG_HYPFS +hypfs_op(unsigned int cmd, const char *arg1, unsigned long arg2, void *arg3, unsigned long arg4) +#endif +#ifdef CONFIG_X86 +xenpmu_op(unsigned int op, xen_pmu_params_t *arg) +#endif + +#ifdef CONFIG_PV +caller: pv64 +#ifdef CONFIG_PV32 +caller: pv32 +#endif
Quoted 8 lines
.gitignore
Quoted 6 lines
.gitignore +6
-0+#endif +#ifdef CONFIG_HVM +caller: hvm64 +#ifdef CONFIG_COMPAT +caller: hvm32 +#endifHVM selects COMPAT, so I don't see a point in this inner conditional.
Ah, indeed.
Quoted 69 lines
Quoted 69 lines
.gitignore +68
-0+#endif +#ifdef CONFIG_ARM +caller: arm +#endif + +table: pv32 pv64 hvm32 hvm64 arm +set_trap_table compat do - - - +mmu_update do do - - - +set_gdt compat do - - - +stack_switch do do - - - +set_callbacks compat do - - - +fpu_taskswitch do do - - - +sched_op_compat do do - - dep +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +platform_op compat do compat do do +#endif +set_debugreg do do - - - +get_debugreg do do - - - +update_descriptor compat do - - - +memory_op compat do hvm hvm do +multicall compat do compat do do +update_va_mapping compat do - - - +set_timer_op compat do compat do - +event_channel_op_compat do do - - dep +xen_version compat do compat do do +console_io do do do do do +physdev_op_compat compat do - - dep +#if defined(CONFIG_GRANT_TABLE) || defined(CONFIG_PV_SHIM) +grant_table_op compat do hvm hvm do +#endif +vm_assist do do do do do +update_va_mapping_otherdomain compat do - - - +iret compat do - - - +vcpu_op compat do compat do do +set_segment_base do do - - - +#ifdef CONFIG_PV +mmuext_op compat do compat do - +#endif +xsm_op compat do compat do do +nmi_op compat do - - - +sched_op compat do compat do do +callback_op compat do - - - +#ifdef CONFIG_XENOPROF +xenoprof_op compat do - - - +#endif +event_channel_op do do do do do +physdev_op compat do hvm hvm do +#ifdef CONFIG_HVM +hvm_op do do do do do +#endif +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +sysctl do do do do do +domctl do do do do do +#endif +#ifdef CONFIG_KEXEC +kexec_op compat do - - - +#endif +tmem_op - - - - - +#ifdef CONFIG_ARGO +argo_op compat do compat do do +#endif +xenpmu_op do do do do - +#ifdef CONFIG_IOREQ_SERVER +dm_op compat do compat do do +#endif +#ifdef CONFIG_HYPFS +hypfs_op do do do do do +#endif
Quoted 8 lines
.gitignore
Quoted 4 lines
.gitignore +4
-0+mca do do - - - +#ifndef CONFIG_PV_SHIM_EXCLUSIVE +paging_domctl_cont do do do do - +#endifI assume the intention here is to sort by hypercall number. This results in 3
PV_SHIM_EXCLUSIVEconditionals when one might do. Just a remark for consideration, not strictly a request to change anything.
I intentionally sorted this by hypercall number, yes.
This makes it much easier to spot any missing case IMO. I can change that if wanted.
Juergen
Re: [PATCH 06/12] xen: generate hypercall interface related code
Quoted 10 lines · Juergen Gross
Quoted 9 lines · Jan Beulich
Quoted 7 lines · Juergen Gross
Instead of repeating similar data multiple times use a single source file and a generator script for producing prototypes and call sequences of the hypercalls.
As the script already knows the number of parameters used add generating a macro for populating an array with the number of parameters per hypercall.
Quoted 6 lines
Quoted 1 line
Isn't that array intended to go away?
I thought so, yes, but on Arm there is a case where it is needed.
So generating it from the available data is the sensible thing to do IMO.
Absolutely, if such a table continues to be needed.
Quoted 18 lines
Quoted 4 lines
Quoted 4 lines
xen/Makefile +0
-0@@ -466,6 +468,14 @@ include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s echo ""; \ echo "#endif") <$< >$@Quoted 13 lines
xen/Makefile include/asm-$(TARGET_ARCH)/asm-offsets.h: asm-offsets.s · line 477
Quoted 7 lines
xen/Makefile +7
-0+quiet_cmd_genhyp = GEN $@ +define cmd_genhyp + awk -f scripts/gen_hypercall.awk <$< >$@ +endef + +include/xen/hypercall-defs.h: include/hypercall-defs.i scripts/gen_hypercall.awk FORCE + $(call if_changed,genhyp)As per patch 5 there are quite a few sources including
xen/hypercall.hand hence (in one of the next patches) the header generated here. If this one gets re-generated for a benign reason (i.e. without changing the header), all dependents will get rebuilt for no reason. Use $(move-if-changed ...)?
Quoted 2 lines
The reasons for re-generating are quite limited. The most probable case is a .config change, which will trigger quite some rebuild anyway.
Oh, good point - I should also have considered the dependencies here, which are pretty limited. Please disregard my remark then.
Jan
Re: [PATCH 06/12] xen: generate hypercall interface related code
In reply to Juergen GrossQuoted 12 lines · Juergen Gross
Quoted 10 lines · Jan Beulich
Quoted 7 lines · Juergen Gross
.gitignore +0
-0@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.hxen/include/compat/*
xen/include/config/
xen/include/generated/
+xen/include//hypercall-defs.iNit: Stray double slash (unless this has a meaning I'm unaware of).
Quoted 6 lines
Oh, right. No special meaning AFAIK.
Quoted 2 lines
Yet then I wonder: Shouldn't *.i be among the patterns at the top of the file, like *.o is?
Yes, I can do that. Probably via a separate patch then.
I can't do that, as we have one source file in our git tree matching this pattern: tools/libs/stat/bindings/swig/xenstat.i is used as an input file for swig for generating perl and python bindings. And the .i suffix seems to be the common one for swig input files.
Juergen
Re: [PATCH 06/12] xen: generate hypercall interface related code
Quoted 18 lines · Juergen Gross
Quoted 17 lines · Juergen Gross
Quoted 10 lines · Jan Beulich
Quoted 7 lines · Juergen Gross
.gitignore +0
-0@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.hxen/include/compat/*
xen/include/config/
xen/include/generated/
+xen/include//hypercall-defs.iNit: Stray double slash (unless this has a meaning I'm unaware of).
Oh, right. No special meaning AFAIK.
Quoted 2 lines
Yet then I wonder: Shouldn't *.i be among the patterns at the top of the file, like *.o is?
Quoted 6 lines
Quoted 1 line
Yes, I can do that. Probably via a separate patch then.
I can't do that, as we have one source file in our git tree matching this pattern:
tools/libs/stat/bindings/swig/xenstat.iis used as an input file for swig for generating perl and python bindings. And the .i suffix seems to be the common one for swig input files.
Ugly. Since we have a rule to produce *.i in xen/Rules.mk, I think we really should have these ignored. Perhaps a good enough reason to put *.i in xen/.gitignore? And while at it perhaps also *.s? Unless there's a way to specify a pattern for an entire subtree - it's not clear to me whether xen/*.i in ./.gitignore would cover subdirs of xen/ as well ...
Jan
Re: [PATCH 06/12] xen: generate hypercall interface related code
Quoted 26 lines · Jan Beulich
Quoted 18 lines · Juergen Gross
Quoted 17 lines · Juergen Gross
Quoted 10 lines · Jan Beulich
.gitignore xen/include/asm-x86/asm-macros.h · line 335
Quoted 7 lines · Juergen Gross
.gitignore +1
-0@@ -332,10 +332,12 @@ xen/include/asm-x86/asm-macros.h xen/include/compat/* xen/include/config/ xen/include/generated/ +xen/include//hypercall-defs.iNit: Stray double slash (unless this has a meaning I'm unaware of).
Oh, right. No special meaning AFAIK.
Quoted 2 lines
Yet then I wonder: Shouldn't *.i be among the patterns at the top of the file, like *.o is?
Quoted 6 lines
Quoted 1 line
Yes, I can do that. Probably via a separate patch then.
I can't do that, as we have one source file in our git tree matching this pattern:
tools/libs/stat/bindings/swig/xenstat.iis used as an input file for swig for generating perl and python bindings. And the .i suffix seems to be the common one for swig input files.
Quoted 6 lines
Ugly. Since we have a rule to produce *.i in
xen/Rules.mk, I think we really should have these ignored. Perhaps a good enough reason to put *.i in xen/.gitignore? And while at it perhaps also *.s? Unless there's a way to specify a pattern for an entire subtree - it's not clear to me whether xen/*.i in ./.gitignore would cover subdirs of xen/ as well ...
xen/**/*.i will do that. From the gitignore syntax description:
A slash followed by two consecutive asterisks then a slash matches zero or more directories. For example, "a/**/b" matches "a/b", "a/x/b", "a/x/y/b" and so on.
So I'll go with adding xen/**/*.i and xen/**/*.s to .gitignore.
Juergen
PATCH 7/12 xen: use generated prototypes for hypercall handlers
In reply to Juergen GrossRemove the hypercall handler's prototypes in the related header files and use the generated ones instead.
Some handlers having been static before need to be made globally visible.
--- xen/arch/x86/hvm/hypercall.c | 6 +- xen/include/asm-x86/hypercall.h | 133 ------------------------ xen/include/xen/hypercall.h | 177 +------------------------------- 3 files changed, 5 insertions(+), 311 deletions(-)
xen/arch/x86/hvm/hypercall.c +3 -3
@@ -31,7 +31,7 @@#include <public/hvm/hvm_op.h> #include <public/hvm/params.h> -static long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +long hvm_memory_op (unsigned long cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {long rc;
@@ -51,7 +51,7 @@ static long hvm_memory_op(unsigned long cmd, XEN_GUEST_HANDLE_PARAM(void) arg) }#ifdef CONFIG_GRANT_TABLE -static long hvm_grant_table_op ( +long hvm_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop,unsigned int count) { switch ( cmd )
@@ -77,7 +77,7 @@ static long hvm_grant_table_op( }#endif -static long hvm_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +long hvm_physdev_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {const struct vcpu *curr = current;const struct domain *currd = curr->domain;
xen/include/asm-x86/hypercall.h +0 -133
145 lines not on this page · raw message
xen/include/xen/hypercall.h +2 -175
205 lines not on this page · raw message
Signature · 1 line
2.26.2
Re: [PATCH 07/12] xen: use generated prototypes for hypercall handlers
Quoted 7 lines · Juergen Gross
Remove the hypercall handler's prototypes in the related header files and use the generated ones instead.
Some handlers having been static before need to be made globally visible.
Signed-off-by Juergen Gross <[email protected]>
PATCH 8/12 x86/pv-shim: don't modify hypercall table
In reply to Juergen GrossWhen running as pv-shim the hypercall is modified today in order to replace the functions for __HYPERVISOR_event_channel_op and __HYPERVISOR_grant_table_op hypercalls.
Change this to call the related functions from the normal handlers instead when running as shim. The performance implications are not really relevant, as a normal production hypervisor will not be configured to support shim mode, so the related calls will be dropped due to optimization of the compiler.
Note that for the CONFIG_PV_SHIM_EXCLUSIVE case there is a dummy wrapper do_grant_table_op() needed, as in this case grant_table.c isn't being built.
--- xen/arch/x86/pv/shim.c | 48 +++++++++++++++-------------------- xen/common/event_channel.c | 9 +++++++ xen/common/grant_table.c | 9 +++++++ xen/include/asm-x86/pv/shim.h | 3 +++ 4 files changed, 42 insertions(+), 27 deletions(-)
xen/arch/x86/pv/shim.c +21 -27
@@ -56,11 +56,6 @@ static DEFINE_SPINLOCK(balloon_lock);static struct platform_bad_page __initdata reserved_pages[2 ]; -static long pv_shim_event_channel_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg); -static long pv_shim_grant_table_op (unsigned int cmd, -XEN_GUEST_HANDLE_PARAM (void ) uop, -unsigned int count); -/* * By default give the shim 1MB of free memory slack. Some users may wish to * tune this constants for better memory utilization. This can be achieved
@@ -203,7 +198,6 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start, start_info_t *si) {bool compat =is_pv_32bit_domain (d); -pv_hypercall_table_t *rw_pv_hypercall_table;uint64_t param =0 ;long rc;
@@ -249,23 +243,6 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,consoled_set_ring_addr (page); } -/* -* Locate pv_hypercall_table[] (usually .rodata) in the directmap (which -* is writeable) and insert some shim-specific hypercall handlers. -*/ - rw_pv_hypercall_table =__va (__pa (pv_hypercall_table)); - rw_pv_hypercall_table[__HYPERVISOR_event_channel_op].native = - (hypercall_fn_t *)pv_shim_event_channel_op; - rw_pv_hypercall_table[__HYPERVISOR_grant_table_op].native = - (hypercall_fn_t *)pv_shim_grant_table_op; - -#ifdef CONFIG_PV32 - rw_pv_hypercall_table[__HYPERVISOR_event_channel_op].compat = - (hypercall_fn_t *)pv_shim_event_channel_op; - rw_pv_hypercall_table[__HYPERVISOR_grant_table_op].compat = - (hypercall_fn_t *)pv_shim_grant_table_op; -#endif - guest = d;/*
@@ -435,7 +412,7 @@ int pv_shim_shutdown(uint8_t reason)return 0 ; } -static long pv_shim_event_channel_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) +long pv_shim_event_channel_op (int cmd,XEN_GUEST_HANDLE_PARAM (void ) arg) {struct domain *d = current->domain;struct evtchn_close close;
@@ -683,9 +660,9 @@ void pv_shim_inject_evtchn(unsigned int port)# define compat_handle_okay guest_handle_okay#endif -static long pv_shim_grant_table_op (unsigned int cmd, -XEN_GUEST_HANDLE_PARAM (void ) uop, -unsigned int count) +long pv_shim_grant_table_op (unsigned int cmd, +XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) {struct domain *d = current->domain;long rc =0 ;
@@ -845,6 +822,23 @@ static long pv_shim_grant_table_op(unsigned int cmd,return rc; } +#ifndef CONFIG_GRANT_TABLE +/* Thin wrapper(s) needed. */ +long do_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{ +return pv_shim_grant_table_op (cmd, uop, count); +} + +#ifdef CONFIG_PV32 +int compat_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{ +return pv_shim_grant_table_op (cmd, uop, count); +} +#endif +#endif +long pv_shim_cpu_up (void *data) {struct vcpu *v = data;
xen/common/event_channel.c +9 -0
@@ -32,6 +32,10 @@#include <public/event_channel.h> #include <xsm/xsm.h> +#ifdef CONFIG_PV_SHIM +#include <asm/guest.h> +#endif +#define ERROR_EXIT (_errno) \ do { \ gdprintk(XENLOG_WARNING, \
@@ -1190,6 +1194,11 @@ long do_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg) {int rc; +#ifdef CONFIG_PV_SHIM +if ( pv_shim ) +return pv_shim_event_channel_op (cmd, arg); +#endif +switch ( cmd ) {case EVTCHNOP_alloc_unbound: {
xen/common/grant_table.c +9 -0
@@ -45,6 +45,10 @@#include <asm/flushtlb.h> #include <asm/guest_atomics.h> +#ifdef CONFIG_PV_SHIM +#include <asm/guest.h> +#endif +/* Per-domain grant information. */ struct grant_table {/*
@@ -3526,6 +3530,11 @@ do_grant_table_op(long rc;unsigned int opaque_in = cmd &GNTTABOP_ARG_MASK , opaque_out =0 ; +#ifdef CONFIG_PV_SHIM +if ( pv_shim ) +return pv_shim_grant_table_op (cmd, uop, count); +#endif +if ( (int )count <0 )return -EINVAL ;
xen/include/asm-x86/pv/shim.h +3 -0
@@ -19,6 +19,7 @@#ifndef __X86_PV_SHIM_H__#define __X86_PV_SHIM_H__ +#include <xen/hypercall.h> #include <xen/types.h> #if defined(CONFIG_PV_SHIM_EXCLUSIVE )
@@ -45,6 +46,8 @@ domid_t get_initial_domain_id(void);uint64_t pv_shim_mem (uint64_t avail);void pv_shim_fixup_e820 (struct e820map *e820);const struct platform_bad_page *pv_shim_reserved_pages (unsigned int *size); +typeof (do_event_channel_op) pv_shim_event_channel_op; +typeof (do_grant_table_op) pv_shim_grant_table_op;#else
Signature · 1 line
2.26.2
Re: [PATCH 08/12] x86/pv-shim: don't modify hypercall table
Quoted 10 lines · Juergen Gross
When running as pv-shim the hypercall is modified today in order to replace the functions for
__HYPERVISOR_event_channel_opand__HYPERVISOR_grant_table_ophypercalls.Change this to call the related functions from the normal handlers instead when running as shim. The performance implications are not really relevant, as a normal production hypervisor will not be configured to support shim mode, so the related calls will be dropped due to optimization of the compiler.
Quoted 3 lines
Note that for the
CONFIG_PV_SHIM_EXCLUSIVEcase there is a dummy wrapperdo_grant_table_op()needed, as in this casegrant_table.cisn't being built.
While you say CONFIG_PV_SHIM_EXCLUSIVE here, ...
xen/arch/x86/pv/shim.c static long pv_shim_grant_table_op(unsigned int cmd, · line 825
Quoted 5 lines
xen/arch/x86/pv/shim.c +1
-0@@ -845,6 +822,23 @@ static long pv_shim_grant_table_op(unsigned int cmd,return rc; } +#ifndef CONFIG_GRANT_TABLE
... you don't actually enforce this here. I also don't see why it would be needed in the "exclusive" case only. A binary usable both ways would still need these, wouldn't it?
Quoted 5 lines
xen/arch/x86/pv/shim.c +5
-0+/* Thin wrapper(s) needed. */ +long do_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{ +return pv_shim_grant_table_op (cmd, uop, count);
xen/arch/x86/pv/shim.c static long pv_shim_grant_table_op(unsigned int cmd, · line 840
Quoted 10 lines
xen/arch/x86/pv/shim.c +10
-0+} + +#ifdef CONFIG_PV32 +int compat_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{ +return pv_shim_grant_table_op (cmd, uop, count); +} +#endif +#endif
Don't you also need to adjust the respective #ifdef in pv_hypercall_table[]? Otherwise I don't see how, at this point of the series, the functions would actually get hooked up. In a bi-modal binary further guarding will then be needed inside the wrappers, I think. (While the table is going to go away, that guarding is going to be needed even at the end of this series, I believe.)
Talking of wrappers - do you need actual code to be emitted for this? Can't you simply put in place an alias each, which is permitted now that pv_shim_grant_table_op() isn't static anymore? (Albeit that's partly moot if guarding code gets added to the functions - in that case only compat_grant_table_op() could become an alias of do_grant_table_op().)
Jan
Re: [PATCH 08/12] x86/pv-shim: don't modify hypercall table
Quoted 21 lines · Jan Beulich
Quoted 10 lines · Juergen Gross
When running as pv-shim the hypercall is modified today in order to replace the functions for
__HYPERVISOR_event_channel_opand__HYPERVISOR_grant_table_ophypercalls.Change this to call the related functions from the normal handlers instead when running as shim. The performance implications are not really relevant, as a normal production hypervisor will not be configured to support shim mode, so the related calls will be dropped due to optimization of the compiler.
Quoted 3 lines
Note that for the
CONFIG_PV_SHIM_EXCLUSIVEcase there is a dummy wrapperdo_grant_table_op()needed, as in this casegrant_table.cisn't being built.While you say
CONFIG_PV_SHIM_EXCLUSIVEhere, ...Quoted 4 lines
xen/arch/x86/pv/shim.c +0
-0@@ -845,6 +822,23 @@ static long pv_shim_grant_table_op(unsigned int cmd,return rc; }
Quoted 5 lines
xen/arch/x86/pv/shim.c static long pv_shim_grant_table_op(unsigned int cmd, · line 825
Quoted 1 line
xen/arch/x86/pv/shim.c +1
-0+#ifndef CONFIG_GRANT_TABLE ... you don't actually enforce this here. I also don't see why it would be needed in the "exclusive" case only. A binary usable both ways would still need these, wouldn't it?
The "exclusive" case is normally the one where CONFIG_GRANT_TABLE is not set. I highlighted this as it is a common situation.
Quoted 12 lines
Quoted 12 lines
xen/arch/x86/pv/shim.c +11
-0+/* Thin wrapper(s) needed. */ +long do_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{ +return pv_shim_grant_table_op (cmd, uop, count); +} + +#ifdef CONFIG_PV32 +int compat_grant_table_op (unsigned int cmd,XEN_GUEST_HANDLE_PARAM (void ) uop, +unsigned int count) +{
Quoted 8 lines
xen/arch/x86/pv/shim.c static long pv_shim_grant_table_op(unsigned int cmd, · line 840
Quoted 4 lines
xen/arch/x86/pv/shim.c +4
-0+return pv_shim_grant_table_op (cmd, uop, count); +} +#endif +#endif Don't you also need to adjust the respective #ifdef in pv_hypercall_table[]? Otherwise I don't see how, at this point of the series, the functions would actually get hooked up.
Ah, right.
Quoted 4 lines
In a bi-modal
binary further guarding will then be needed inside the wrappers, I think. (While the table is going to go away, that guarding is going to be needed even at the end of this series, I believe.)
Oh, you mean for the weird case of !pv_shim? Yes, I need to return -ENOSYS in this case.
Quoted 6 lines
Talking of wrappers - do you need actual code to be emitted for this? Can't you simply put in place an alias each, which is permitted now that
pv_shim_grant_table_op()isn't static anymore? (Albeit that's partly moot if guarding code gets added to the functions - in that case onlycompat_grant_table_op()could become an alias ofdo_grant_table_op().)
I didn't think of an alias. But I don't think I can make compat_grant_table_op() an alias of do_grant_table_op(), as they have different return types.
Juergen
PATCH 9/12 xen/x86: don't use hypercall table for calling compat hypercalls
In reply to Juergen GrossToday the *_op_compat hypercalls call the modern handler functions by using the entries from the hypercall table. This is resulting in a not needed indirect function call which can be avoided by using the correct handler function directly. This is basically a revert of commit 1252e282311734 ("86/pv: Export pv_hypercall_table[] rather than working around it in several ways"), which reasoning no longer applies, as shim no longer modifies the hypercall table.
The hypercall table can now be made static as there is no external reference to it any longer.
Commit 834cb8761051f7 ("x86/PV32: fix physdev_op_compat handling") can be reverted, too, as using the direct call of the correct handler is already handled fine without that patch.
--- xen/arch/x86/compat.c | 14 ++++---------- xen/arch/x86/pv/hypercall.c | 9 ++++++++- xen/arch/x86/x86_64/compat.c | 1 - xen/include/asm-x86/hypercall.h | 8 -------- 4 files changed, 12 insertions(+), 20 deletions(-)
xen/arch/x86/compat.c +4 -10
@@ -17,14 +17,12 @@ typedef long ret_t;/* Legacy hypercall (as of 0x00030202). */ ret_t do_physdev_op_compat (XEN_GUEST_HANDLE_PARAM (physdev_op_t) uop) { -typeof (do_physdev_op) *fn = - (void *)pv_hypercall_table[__HYPERVISOR_physdev_op].native;struct physdev_op op;if (unlikely (copy_from_guest (&op, uop,1 ) !=0 ) )return -EFAULT ; -return fn (op.cmd,guest_handle_from_ptr (&uop.p->u, void)); +return do_physdev_op (op.cmd,guest_handle_from_ptr (&uop.p->u, void)); }#ifndef COMPAT
@@ -32,14 +30,11 @@ ret_t do_physdev_op_compat(XEN_GUEST_HANDLE_PARAM(physdev_op_t) uop)/* Legacy hypercall (as of 0x00030101). */ long do_sched_op_compat (int cmd,unsigned long arg) { -typeof (do_sched_op) *fn = - (void *)pv_hypercall_table[__HYPERVISOR_sched_op].native; -switch ( cmd ) {case SCHEDOP_yield:case SCHEDOP_block: -return fn (cmd,guest_handle_from_ptr (NULL , void)); +return do_sched_op (cmd,guest_handle_from_ptr (NULL , void));case SCHEDOP_shutdown: TRACE_3D(TRC_SCHED_SHUTDOWN,
@@ -57,8 +52,6 @@ long do_sched_op_compat(int cmd, unsigned long arg)/* Legacy hypercall (as of 0x00030202). */ long do_event_channel_op_compat (XEN_GUEST_HANDLE_PARAM (evtchn_op_t) uop) { -typeof (do_event_channel_op) *fn = - (void *)pv_hypercall_table[__HYPERVISOR_event_channel_op].native;struct evtchn_op op;if (unlikely (copy_from_guest (&op, uop,1 ) !=0 ) )
@@ -76,7 +69,8 @@ long do_event_channel_op_compat(XEN_GUEST_HANDLE_PARAM(evtchn_op_t) uop)case EVTCHNOP_bind_ipi:case EVTCHNOP_bind_vcpu:case EVTCHNOP_unmask: -return fn (op.cmd,guest_handle_from_ptr (&uop.p->u, void)); +return do_event_channel_op (op.cmd, +guest_handle_from_ptr (&uop.p->u, void));default :return -ENOSYS ;
xen/arch/x86/pv/hypercall.c +8 -1
@@ -27,6 +27,13 @@#include <asm/multicall.h> #include <irq_vectors.h> +typedef struct { +hypercall_fn_t *native; +#ifdef CONFIG_PV32 +hypercall_fn_t *compat; +#endif +}pv_hypercall_table_t ; +#ifdef CONFIG_PV32 #define HYPERCALL (x) \ [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \
@@ -40,7 +47,7 @@#define COMPAT_CALL (x) HYPERCALL(x)#endif -const pv_hypercall_table_t pv_hypercall_table[] = { +static const pv_hypercall_table_t pv_hypercall_table[] = {COMPAT_CALL (set_trap_table),HYPERCALL (mmu_update),COMPAT_CALL (set_gdt),
xen/arch/x86/x86_64/compat.c +0 -1
@@ -12,7 +12,6 @@ EMIT_FILE;#define physdev_op_t physdev_op_compat_t#define do_physdev_op compat_physdev_op#define do_physdev_op_compat (x) compat_physdev_op_compat(_##x) -#define native compat#define COMPAT #define _XEN_GUEST_HANDLE (t) XEN_GUEST_HANDLE(t)
xen/include/asm-x86/hypercall.h +0 -8
@@ -21,13 +21,6 @@ typedef unsigned long hypercall_fn_t( unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); -typedef struct { -hypercall_fn_t *native; -#ifdef CONFIG_PV32 -hypercall_fn_t *compat; -#endif -}pv_hypercall_table_t ; - typedef struct { uint8_t native; #ifdef CONFIG_COMPAT
@@ -38,7 +31,6 @@ typedef struct {extern const hypercall_args_t hypercall_args_table[NR_hypercalls];#ifdef CONFIG_PV -extern const pv_hypercall_table_t pv_hypercall_table[];void pv_hypercall (struct cpu_user_regs *regs);#endif
Signature · 1 line
2.26.2
PATCH 10/12 xen/x86: call hypercall handlers via switch statement
In reply to Juergen GrossInstead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This is beneficial to performance and avoids speculation issues.
With calling the handlers using the correct number of parameters now it is possible to do the parameter register clobbering in the NDEBUG case after returning from the handler. This in turn removes the only users of hypercall_args_table[] which can be removed now.
--- xen/arch/x86/hvm/hypercall.c | 144 +++---------------------- xen/arch/x86/hypercall.c | 59 ----------- xen/arch/x86/pv/hypercall.c | 180 +++----------------------------- xen/include/asm-x86/hypercall.h | 43 +++++--- 4 files changed, 60 insertions(+), 366 deletions(-)
xen/arch/x86/hvm/hypercall.c +13 -131
@@ -108,56 +108,10 @@ long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)return compat_physdev_op (cmd, arg); } -#define HYPERCALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) do_ ## x } - -#define HVM_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) hvm_ ## x, \ - (hypercall_fn_t *) hvm_ ## x } - -#define COMPAT_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) compat_ ## x } - -static const struct { -hypercall_fn_t *native, *compat; -} hvm_hypercall_table[] = { -HVM_CALL (memory_op), -COMPAT_CALL (multicall), -#ifdef CONFIG_GRANT_TABLE -HVM_CALL (grant_table_op), -#endif -HYPERCALL (vm_assist), -COMPAT_CALL (vcpu_op), -HVM_CALL (physdev_op), -COMPAT_CALL (xen_version), -HYPERCALL (console_io), -HYPERCALL (event_channel_op), -COMPAT_CALL (sched_op), -COMPAT_CALL (set_timer_op), -COMPAT_CALL (xsm_op), -HYPERCALL (hvm_op), -HYPERCALL (sysctl), -HYPERCALL (domctl), -#ifdef CONFIG_ARGO -COMPAT_CALL (argo_op), -#endif -COMPAT_CALL (platform_op), -#ifdef CONFIG_PV -COMPAT_CALL (mmuext_op), -#endif -HYPERCALL (xenpmu_op), -COMPAT_CALL (dm_op), -#ifdef CONFIG_HYPFS -HYPERCALL (hypfs_op), +#ifndef NDEBUG +static unsigned char hypercall_args_64[] = hypercall_args_hvm64; +static unsigned char hypercall_args_32[] = hypercall_args_hvm32;#endif -HYPERCALL (paging_domctl_cont) -}; - -#undef HYPERCALL -#undef HVM_CALL -#undef COMPAT_CALLint hvm_hypercall (struct cpu_user_regs *regs) {
@@ -203,23 +157,6 @@ int hvm_hypercall(struct cpu_user_regs *regs)return ret; } -BUILD_BUG_ON (ARRAY_SIZE (hvm_hypercall_table) > -ARRAY_SIZE (hypercall_args_table)); - -if ( eax >=ARRAY_SIZE (hvm_hypercall_table) ) - { - regs->rax = -ENOSYS ; -return HVM_HCALL_completed; - } - - eax =array_index_nospec (eax,ARRAY_SIZE (hvm_hypercall_table)); - -if ( !hvm_hypercall_table[eax].native ) - { - regs->rax = -ENOSYS ; -return HVM_HCALL_completed; - } -/* * Caching is intended for instruction emulation only. Disable it * for any accesses by hypercall argument copy-in / copy-out.
@@ -239,33 +176,11 @@ int hvm_hypercall(struct cpu_user_regs *regs)HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%lx, %lx, %lx, %lx, %lx)" , eax, rdi, rsi, rdx, r10, r8); -#ifndef NDEBUG -/* Deliberately corrupt parameter regs not used by this hypercall. */ -switch ( hypercall_args_table[eax].native ) - { -case 0 : rdi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : r8 =0xdeadbeefdeadf00dUL ; - } -#endif - - regs->rax = hvm_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8); +call_handlers_hvm64 (eax, regs->rax, rdi, rsi, rdx, r10, r8);#ifndef NDEBUG -if ( !curr->hcall_preempted ) - { -/* Deliberately corrupt parameter regs used by this hypercall. */ -switch ( hypercall_args_table[eax].native ) - { -case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : regs->rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : regs->rdi =0xdeadbeefdeadf00dUL ; - } - } +if ( !curr->hcall_preempted && regs->rax != -ENOSYS ) +clobber_regs (regs, hypercall_args_64[eax]);#endif }else
@@ -279,35 +194,13 @@ int hvm_hypercall(struct cpu_user_regs *regs)HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%x, %x, %x, %x, %x)" , eax, ebx, ecx, edx, esi, edi); -#ifndef NDEBUG -/* Deliberately corrupt parameter regs not used by this hypercall. */ -switch ( hypercall_args_table[eax].compat ) - { -case 0 : ebx =0xdeadf00d ; fallthrough; -case 1 : ecx =0xdeadf00d ; fallthrough; -case 2 : edx =0xdeadf00d ; fallthrough; -case 3 : esi =0xdeadf00d ; fallthrough; -case 4 : edi =0xdeadf00d ; - } -#endif - curr->hcall_compat = true; - regs->eax = hvm_hypercall_table[eax].compat (ebx, ecx, edx, esi, edi); +call_handlers_hvm32 (eax, regs->eax, ebx, ecx, edx, esi, edi); curr->hcall_compat = false;#ifndef NDEBUG -if ( !curr->hcall_preempted ) - { -/* Deliberately corrupt parameter regs used by this hypercall. */ -switch ( hypercall_args_table[eax].compat ) - { -case 5 : regs->rdi =0xdeadf00d ; fallthrough; -case 4 : regs->rsi =0xdeadf00d ; fallthrough; -case 3 : regs->rdx =0xdeadf00d ; fallthrough; -case 2 : regs->rcx =0xdeadf00d ; fallthrough; -case 1 : regs->rbx =0xdeadf00d ; - } - } +if ( !curr->hcall_preempted && regs->eax != -ENOSYS ) +clobber_regs32 (regs, hypercall_args_32[eax]);#endif }
@@ -327,31 +220,20 @@ int hvm_hypercall(struct cpu_user_regs *regs)enum mc_disposition hvm_do_multicall_call (struct mc_state *state) {struct vcpu *curr = current; -hypercall_fn_t *func =NULL ;if (hvm_guest_x86_mode (curr) ==8 ) {struct multicall_entry *call = &state->call; -if ( call->op <ARRAY_SIZE (hvm_hypercall_table) ) - func =array_access_nospec (hvm_hypercall_table, call->op).native; -if ( func ) - call->result =func (call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ]); -else - call->result = -ENOSYS ; +call_handlers_hvm64 (call->op, call->result, call->args[0 ], call->args[1 ], + call->args[2 ], call->args[3 ], call->args[4 ]); }else {struct compat_multicall_entry *call = &state->compat_call; -if ( call->op <ARRAY_SIZE (hvm_hypercall_table) ) - func =array_access_nospec (hvm_hypercall_table, call->op).compat; -if ( func ) - call->result =func (call->args[0 ], call->args[1 ], call->args[2 ], - call->args[3 ], call->args[4 ]); -else - call->result = -ENOSYS ; +call_handlers_hvm32 (call->op, call->result, call->args[0 ], call->args[1 ], + call->args[2 ], call->args[3 ], call->args[4 ]); }return !hvm_get_cpl (curr) ? mc_continue : mc_preempt;
xen/arch/x86/hypercall.c +0 -59
65 lines not on this page · raw message
xen/arch/x86/pv/hypercall.c +17 -163
239 lines not on this page · raw message
xen/include/asm-x86/hypercall.h +30 -13
53 lines not on this page · raw message
Signature · 1 line
2.26.2
Re: [PATCH 10/12] xen/x86: call hypercall handlers via switch statement
Quoted 9 lines · Juergen Gross
Instead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This is beneficial to performance and avoids speculation issues.
With calling the handlers using the correct number of parameters now it is possible to do the parameter register clobbering in the NDEBUG case after returning from the handler. This in turn removes the only users of hypercall_args_table[] which can be removed now.
"removed" reads misleading to me: You really replace it by new tables, using script-generated initializers. Also it looks like you're doubling the data, as the same sets were previously used by pv64/hvm64 and pv32/hvm32 respectively.
Quoted 42 lines
xen/arch/x86/hvm/hypercall.c +0
-36@@ -108,56 +108,10 @@ long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)return compat_physdev_op (cmd, arg); } -#define HYPERCALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) do_ ## x } - -#define HVM_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) hvm_ ## x, \ - (hypercall_fn_t *) hvm_ ## x } - -#define COMPAT_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) compat_ ## x } - -static const struct { -hypercall_fn_t *native, *compat; -} hvm_hypercall_table[] = { -HVM_CALL (memory_op), -COMPAT_CALL (multicall), -#ifdef CONFIG_GRANT_TABLE -HVM_CALL (grant_table_op), -#endif -HYPERCALL (vm_assist), -COMPAT_CALL (vcpu_op), -HVM_CALL (physdev_op), -COMPAT_CALL (xen_version), -HYPERCALL (console_io), -HYPERCALL (event_channel_op), -COMPAT_CALL (sched_op), -COMPAT_CALL (set_timer_op), -COMPAT_CALL (xsm_op), -HYPERCALL (hvm_op), -HYPERCALL (sysctl), -HYPERCALL (domctl), -#ifdef CONFIG_ARGO -COMPAT_CALL (argo_op), -#endif -COMPAT_CALL (platform_op),
xen/arch/x86/hvm/hypercall.c long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg) · line 113
Quoted 10 lines
xen/arch/x86/hvm/hypercall.c +3
-7-#ifdef CONFIG_PV -COMPAT_CALL (mmuext_op), -#endif -HYPERCALL (xenpmu_op), -COMPAT_CALL (dm_op), -#ifdef CONFIG_HYPFS - HYPERCALL(hypfs_op), +#ifndef NDEBUG +static unsigned char hypercall_args_64[] = hypercall_args_hvm64; +static unsigned char hypercall_args_32[] = hypercall_args_hvm32;
Irrespective of this being debugging-only: Const?
Quoted 24 lines
xen/arch/x86/hvm/hypercall.c +1
-17@@ -239,33 +176,11 @@ int hvm_hypercall(struct cpu_user_regs *regs)HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%lx, %lx, %lx, %lx, %lx)" , eax, rdi, rsi, rdx, r10, r8); -#ifndef NDEBUG -/* Deliberately corrupt parameter regs not used by this hypercall. */ -switch ( hypercall_args_table[eax].native ) - { -case 0 : rdi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : r8 =0xdeadbeefdeadf00dUL ; - } -#endif - - regs->rax = hvm_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8); +call_handlers_hvm64 (eax, regs->rax, rdi, rsi, rdx, r10, r8);#ifndef NDEBUG -if ( !curr->hcall_preempted ) - { -/* Deliberately corrupt parameter regs used by this hypercall. */ - switch ( hypercall_args_table[eax].native )
xen/arch/x86/hvm/hypercall.c int hvm_hypercall(struct cpu_user_regs *regs) · line 183
Quoted 10 lines
xen/arch/x86/hvm/hypercall.c +2
-8- { -case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : regs->rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : regs->rdi =0xdeadbeefdeadf00dUL ; - } - } +if ( !curr->hcall_preempted && regs->rax != -ENOSYS ) +clobber_regs (regs, hypercall_args_64[eax]);
I'm not fundamentally opposed, but sadly -ENOSYS comes back also in undue situations, e.g. various hypercalls still produce this for "unknown sub-function". Hence the weakened clobbering wants at least mentioning, perhaps also justifying, in the description.
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 45
Quoted 5 lines
xen/arch/x86/hvm/hypercall.c +1
-0@@ -55,4 +42,34 @@ compat_common_vcpu_op(#endif /* CONFIG_COMPAT */ +#ifndef NDEBUG
Hmm, I was actuall hoping for the conditional to actually live ...
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 48
Quoted 3 lines
xen/arch/x86/hvm/hypercall.c +3
-0+static inline void clobber_regs (struct cpu_user_regs *regs, +unsigned int nargs) +{
... here and ...
Quoted 4 lines
xen/arch/x86/hvm/hypercall.c +4
-0+/* Deliberately corrupt used parameter regs. */ +switch ( nargs ) + { +case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough;
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 62
Quoted 10 lines
xen/arch/x86/hvm/hypercall.c +10
-0+case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough; +case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough; +case 2 : regs->rsi =0xdeadbeefdeadf00dUL ; fallthrough; +case 1 : regs->rdi =0xdeadbeefdeadf00dUL ; + } +} + +static inline void clobber_regs32 (struct cpu_user_regs *regs, +unsigned int nargs) +{
... here, such that the conditionals in the .c files could go away altogether.
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 70
Quoted 8 lines
xen/arch/x86/hvm/hypercall.c +8
-0+/* Deliberately corrupt used parameter regs. */ +switch ( nargs ) + { +case 5 : regs->edi =0xdeadf00dUL ; fallthrough; +case 4 : regs->esi =0xdeadf00dUL ; fallthrough; +case 3 : regs->edx =0xdeadf00dUL ; fallthrough; +case 2 : regs->ecx =0xdeadf00dUL ; fallthrough; +case 1 : regs->ebx =0xdeadf00dUL ;
No need for the UL suffixes here afaics; U ones may want to be there.
Overall, besides these mainly cosmetic aspects the main thing missing is an approach to prioritize the handful most frequently used functions, for them to be pulled out of the switch() so we don't depend on the compiler's choice for the order of comparisons done.
Jan
Re: [PATCH 10/12] xen/x86: call hypercall handlers via switch statement
Quoted 11 lines · Jan Beulich
Quoted 9 lines · Juergen Gross
Instead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This is beneficial to performance and avoids speculation issues.
With calling the handlers using the correct number of parameters now it is possible to do the parameter register clobbering in the NDEBUG case after returning from the handler. This in turn removes the only users of hypercall_args_table[] which can be removed now.
Quoted 4 lines
"removed" reads misleading to me: You really replace it by new tables, using script-generated initializers. Also it looks like you're doubling the data, as the same sets were previously used by pv64/hvm64 and pv32/hvm32 respectively.
Yes, I'll change that paragraph.
Regarding having 4 tables on x86 now: merging the pv/hvm tables would be possible, but this would add some complexity to the script generating the tables (it should test whether the number of parameters of pv and hvm match). As the tables are present in debug build only I don't think this is a real issue.
Quoted 47 lines
Quoted 47 lines
xen/arch/x86/hvm/hypercall.c +0
-40@@ -108,56 +108,10 @@ long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)return compat_physdev_op (cmd, arg); } -#define HYPERCALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) do_ ## x } - -#define HVM_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) hvm_ ## x, \ - (hypercall_fn_t *) hvm_ ## x } - -#define COMPAT_CALL (x) \ - [ __HYPERVISOR_ ## x ] = { (hypercall_fn_t *) do_ ## x, \ - (hypercall_fn_t *) compat_ ## x } - -static const struct { -hypercall_fn_t *native, *compat; -} hvm_hypercall_table[] = { -HVM_CALL (memory_op), -COMPAT_CALL (multicall), -#ifdef CONFIG_GRANT_TABLE -HVM_CALL (grant_table_op), -#endif -HYPERCALL (vm_assist), -COMPAT_CALL (vcpu_op), -HVM_CALL (physdev_op), -COMPAT_CALL (xen_version), -HYPERCALL (console_io), -HYPERCALL (event_channel_op), -COMPAT_CALL (sched_op), -COMPAT_CALL (set_timer_op), -COMPAT_CALL (xsm_op), -HYPERCALL (hvm_op), -HYPERCALL (sysctl), -HYPERCALL (domctl), -#ifdef CONFIG_ARGO -COMPAT_CALL (argo_op), -#endif -COMPAT_CALL (platform_op), -#ifdef CONFIG_PV -COMPAT_CALL (mmuext_op), -#endif -HYPERCALL (xenpmu_op),
Quoted 8 lines
xen/arch/x86/hvm/hypercall.c long hvm_physdev_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg) · line 113
Quoted 6 lines
xen/arch/x86/hvm/hypercall.c +3
-3- COMPAT_CALL(dm_op), -#ifdef CONFIG_HYPFS - HYPERCALL(hypfs_op), +#ifndef NDEBUG +static unsigned char hypercall_args_64[] = hypercall_args_hvm64; +static unsigned char hypercall_args_32[] = hypercall_args_hvm32;Irrespective of this being debugging-only: Const?
Yes.
Quoted 32 lines
Quoted 32 lines
xen/arch/x86/hvm/hypercall.c +1
-24@@ -239,33 +176,11 @@ int hvm_hypercall(struct cpu_user_regs *regs)HVM_DBG_LOG (DBG_LEVEL_HCALL ,"hcall%lu(%lx, %lx, %lx, %lx, %lx)" , eax, rdi, rsi, rdx, r10, r8); -#ifndef NDEBUG -/* Deliberately corrupt parameter regs not used by this hypercall. */ -switch ( hypercall_args_table[eax].native ) - { -case 0 : rdi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : r8 =0xdeadbeefdeadf00dUL ; - } -#endif - - regs->rax = hvm_hypercall_table[eax].native (rdi, rsi, rdx, r10, r8); +call_handlers_hvm64 (eax, regs->rax, rdi, rsi, rdx, r10, r8);#ifndef NDEBUG -if ( !curr->hcall_preempted ) - { -/* Deliberately corrupt parameter regs used by this hypercall. */ -switch ( hypercall_args_table[eax].native ) - { -case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough; -case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough; -case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough; -case 2 : regs->rsi =0xdeadbeefdeadf00dUL ; fallthrough; -case 1 : regs->rdi =0xdeadbeefdeadf00dUL ; - }
Quoted 8 lines
xen/arch/x86/hvm/hypercall.c int hvm_hypercall(struct cpu_user_regs *regs) · line 183
Quoted 3 lines
xen/arch/x86/hvm/hypercall.c +2
-1- } +if ( !curr->hcall_preempted && regs->rax != -ENOSYS ) +clobber_regs (regs, hypercall_args_64[eax]);I'm not fundamentally opposed, but sadly
-ENOSYScomes back also in undue situations, e.g. various hypercalls still produce this for "unknown sub-function". Hence the weakened clobbering wants at least mentioning, perhaps also justifying, in the description.
Okay.
Quoted 24 lines
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 45
Quoted 6 lines
xen/arch/x86/hvm/hypercall.c +1
-0@@ -55,4 +42,34 @@ compat_common_vcpu_op(#endif /* CONFIG_COMPAT */ +#ifndef NDEBUG Hmm, I was actuall hoping for the conditional to actually live ...
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 48
Quoted 3 lines
xen/arch/x86/hvm/hypercall.c +3
-0+static inline void clobber_regs (struct cpu_user_regs *regs, +unsigned int nargs) +{... here and ...
Quoted 9 lines
xen/arch/x86/hvm/hypercall.c +9
-0+/* Deliberately corrupt used parameter regs. */ +switch ( nargs ) + { +case 5 : regs->r8 =0xdeadbeefdeadf00dUL ; fallthrough; +case 4 : regs->r10 =0xdeadbeefdeadf00dUL ; fallthrough; +case 3 : regs->rdx =0xdeadbeefdeadf00dUL ; fallthrough; +case 2 : regs->rsi =0xdeadbeefdeadf00dUL ; fallthrough; +case 1 : regs->rdi =0xdeadbeefdeadf00dUL ; + }
Quoted 8 lines
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 62
Quoted 5 lines
xen/arch/x86/hvm/hypercall.c +5
-0+} + +static inline void clobber_regs32 (struct cpu_user_regs *regs, +unsigned int nargs) +{... here, such that the conditionals in the .c files could go away altogether.
I didn't do that in order to be able to have the tables with the number of parameters inside #ifndef NDEBUG sections.
I think I can change that by using a macro for reading the table values.
Quoted 11 lines
xen/arch/x86/hvm/hypercall.c compat_common_vcpu_op( · line 70
Quoted 9 lines
xen/arch/x86/hvm/hypercall.c +8
-0+/* Deliberately corrupt used parameter regs. */ +switch ( nargs ) + { +case 5 : regs->edi =0xdeadf00dUL ; fallthrough; +case 4 : regs->esi =0xdeadf00dUL ; fallthrough; +case 3 : regs->edx =0xdeadf00dUL ; fallthrough; +case 2 : regs->ecx =0xdeadf00dUL ; fallthrough; +case 1 : regs->ebx =0xdeadf00dUL ;No need for the UL suffixes here afaics; U ones may want to be there.
Okay.
Quoted 4 lines
Overall, besides these mainly cosmetic aspects the main thing missing is an approach to prioritize the handful most frequently used functions, for them to be pulled out of the
switch()so we don't depend on the compiler's choice for the order of comparisons done.
I have already prepared that step by generating the complete call sequence, so any change for prioritizing some hypercalls can be local to the generator script and the used input data.
The main question is how to do that. I've collected some hypercall statistics data for PV and PVH guests running some simple tests (once a build of the Xen hypervisor, and once a scp of a large file). The data is split between guest and dom0 (PV) counts. There is no clear "winner" which hypercall should be fastest, but several hypercalls are clearly not important.
Here is the data:
PV-hypercall PV-guest build PV-guest scp dom0 build dom0 scpmmu_update 186175729 2865 20936 33725stack_switch 1273311 62381 108589 270764
multicall 2182803 50 302 524update_va_mapping 571868 10 60 80xen_version 73061 850 859 5432
grant_table_op 0 0 35557 139110 iret 75673006 484132 268157 757958 vcpu_op 453037 71199 138224 334988 set_segment_base 1650249 62387 108645 270823 mmuext_op 11225681 188 7239 3426 sched_op 280153 134645 70729 137943 event_channel_op 192327 66204 71409 214191
physdev_op 0 0 7721 4315
(the dom0 values are for the guest running the build or scp test, so dom0 acting as backend)
HVM-hypercall PVH-guest build PVH-guest scpvcpu_op 277684 2324event_channel_op 350233 57383
(the related dom0 counter values are in the same range as with the test running in the PV guest)
It should be noted that during boot of the guests the numbers for the PV guest are more like the ones for the build test with the exception of iret and sched_op being higher, while for PVH sched_op is by far the most often used hypercall.
I'm not sure how to translate those numbers into a good algorithm for generating the call sequence.
I could add priorities to each hypercall in hypercall-defs.c and have a cascade of if (likely(foo)) call_foo; else if (likely(bla)) ... else switch(rest).
Or I could have groups of hypercalls with a priority for each group and:
mask = 1ULL << num;
if (likely(mask & prio_1_mask)) switch(num) ... else if (likely(mask & prio_2_mask)) switch (num) ...
...
else switch (num) ...
Or I could combine those approaches using the mask variant for cases of multiple entries having the same priority and the direct call variant for the cases of only a single entry having a specific priority.
And then there is the problem to set the priorities (fairly simple for HVM, PV is more diffcult).
Juergen
Re: [PATCH 10/12] xen/x86: call hypercall handlers via switch statement
Quoted 17 lines · Juergen Gross
Quoted 11 lines · Jan Beulich
Quoted 9 lines · Juergen Gross
Instead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This is beneficial to performance and avoids speculation issues.
With calling the handlers using the correct number of parameters now it is possible to do the parameter register clobbering in the NDEBUG case after returning from the handler. This in turn removes the only users of hypercall_args_table[] which can be removed now.
Quoted 4 lines
"removed" reads misleading to me: You really replace it by new tables, using script-generated initializers. Also it looks like you're doubling the data, as the same sets were previously used by pv64/hvm64 and pv32/hvm32 respectively.
Quoted 7 lines
Yes, I'll change that paragraph.
Regarding having 4 tables on x86 now: merging the pv/hvm tables would be possible, but this would add some complexity to the script generating the tables (it should test whether the number of parameters of pv and hvm match). As the tables are present in debug build only I don't think this is a real issue.
Sure, but that imo wants saying in the description.
Quoted 16 lines
Quoted 4 lines
Overall, besides these mainly cosmetic aspects the main thing missing is an approach to prioritize the handful most frequently used functions, for them to be pulled out of the
switch()so we don't depend on the compiler's choice for the order of comparisons done.I have already prepared that step by generating the complete call sequence, so any change for prioritizing some hypercalls can be local to the generator script and the used input data.
The main question is how to do that. I've collected some hypercall statistics data for PV and PVH guests running some simple tests (once a build of the Xen hypervisor, and once a scp of a large file). The data is split between guest and dom0 (PV) counts. There is no clear "winner" which hypercall should be fastest, but several hypercalls are clearly not important.
Quoted 4 lines
Here is the data:
PV-hypercall PV-guest build PV-guest scp dom0 build dom0 scp
mmu_update186175729 2865 20936 33725
Builds should be local to the guest and I/O should involve gnttab ops but no mmu-update. Hence I have a hard time seeing where the huge difference here would be coming from. Did you have any thoughts here?
Quoted 2 lines
stack_switch1273311 62381 108589 270764
multicall 2182803 50 302 524
A fair amount of the mmu-updates is going to be coming through muticalls, I would guess. Priorities therefore may even differ for the two separate dispatch points.
Quoted 4 lines
update_va_mapping571868 10 60 80xen_version73061 850 859 5432grant_table_op0 0 35557 139110
iret 75673006 484132 268157 757958
The huge differences for builds is puzzling mere here ...
Quoted 3 lines
vcpu_op 453037 71199 138224 334988 set_segment_base 1650249 62387 108645 270823
mmuext_op11225681 188 7239 3426
... and here as well. Did Dom0 and DomU use identical numbers of vCPU-s and identical -j make option values?
Quoted 12 lines
sched_op 280153 134645 70729 137943 event_channel_op 192327 66204 71409 214191
physdev_op0 0 7721 4315
(the dom0 values are for the guest running the build or scp test, so dom0 acting as backend)HVM-hypercall PVH-guest build PVH-guest scp
vcpu_op277684 2324event_channel_op350233 57383
(the related dom0 counter values are in the same range as with the test running in the PV guest)
Quoted 7 lines
It should be noted that during boot of the guests the numbers for the PV guest are more like the ones for the build test with the exception of iret and
sched_opbeing higher, while for PVHsched_opis by far the most often used hypercall.I'm not sure how to translate those numbers into a good algorithm for generating the call sequence.
Well, there's never going to be a clear cut fitting everything, I suppose.
Quoted 3 lines
I could add priorities to each hypercall in
hypercall-defs.cand have a cascade of if (likely(foo))call_foo; else if (likely(bla)) ... else switch(rest).
Personally I'd lean to an approach like this one; perhaps there's not even a need to specify priorities for every hypercall, but just the ones we deem most frequently used?
Jan
Quoted 18 lines
Or I could have groups of hypercalls with a priority for each group and:
mask = 1ULL << num;
if (likely(mask & prio_1_mask)) switch(num) ... else if (likely(mask & prio_2_mask)) switch (num) ...
...
else switch (num) ...Or I could combine those approaches using the mask variant for cases of multiple entries having the same priority and the direct call variant for the cases of only a single entry having a specific priority.
And then there is the problem to set the priorities (fairly simple for HVM, PV is more diffcult).
Juergen
Re: [PATCH 10/12] xen/x86: call hypercall handlers via switch statement
Quoted 44 lines · Jan Beulich
Quoted 17 lines · Juergen Gross
Quoted 11 lines · Jan Beulich
Quoted 9 lines · Juergen Gross
Instead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This is beneficial to performance and avoids speculation issues.
With calling the handlers using the correct number of parameters now it is possible to do the parameter register clobbering in the NDEBUG case after returning from the handler. This in turn removes the only users of hypercall_args_table[] which can be removed now.
Quoted 4 lines
"removed" reads misleading to me: You really replace it by new tables, using script-generated initializers. Also it looks like you're doubling the data, as the same sets were previously used by pv64/hvm64 and pv32/hvm32 respectively.
Quoted 7 lines
Yes, I'll change that paragraph.
Regarding having 4 tables on x86 now: merging the pv/hvm tables would be possible, but this would add some complexity to the script generating the tables (it should test whether the number of parameters of pv and hvm match). As the tables are present in debug build only I don't think this is a real issue.
Sure, but that imo wants saying in the description.
Quoted 16 lines
Quoted 4 lines
Overall, besides these mainly cosmetic aspects the main thing missing is an approach to prioritize the handful most frequently used functions, for them to be pulled out of the
switch()so we don't depend on the compiler's choice for the order of comparisons done.I have already prepared that step by generating the complete call sequence, so any change for prioritizing some hypercalls can be local to the generator script and the used input data.
The main question is how to do that. I've collected some hypercall statistics data for PV and PVH guests running some simple tests (once a build of the Xen hypervisor, and once a scp of a large file). The data is split between guest and dom0 (PV) counts. There is no clear "winner" which hypercall should be fastest, but several hypercalls are clearly not important.
Quoted 8 lines
Quoted 4 lines
Here is the data:
PV-hypercall PV-guest build PV-guest scp dom0 build dom0 scp
mmu_update186175729 2865 20936 33725Builds should be local to the guest and I/O should involve gnttab ops but no mmu-update. Hence I have a hard time seeing where the huge difference here would be coming from. Did you have any thoughts here?
I think you misunderstood the columns.
The first column of data is the build job running in domU and the number of hypercalls done by that domU. The 3rd data column is the same test (build running in domU), but the number of hypercalls done by dom0 (so pure backend hypercall activity).
The missing gnttab ops on domU side are fine, as granting a page doesn't require a hypercall.
Quoted 7 lines
Quoted 3 lines
stack_switch1273311 62381 108589 270764
multicall 2182803 50 302 524A fair amount of the mmu-updates is going to be coming through muticalls, I would guess. Priorities therefore may even differ for the two separate dispatch points.
I can look into collecting some data here.
Quoted 38 lines
Quoted 5 lines
update_va_mapping571868 10 60 80xen_version73061 850 859 5432grant_table_op0 0 35557 139110
iret 75673006 484132 268157 757958The huge differences for builds is puzzling mere here ...
Quoted 3 lines
vcpu_op 453037 71199 138224 334988 set_segment_base 1650249 62387 108645 270823
mmuext_op11225681 188 7239 3426... and here as well. Did Dom0 and DomU use identical numbers of vCPU-s and identical -j make option values?
Quoted 12 lines
sched_op 280153 134645 70729 137943 event_channel_op 192327 66204 71409 214191
physdev_op0 0 7721 4315
(the dom0 values are for the guest running the build or scp test, so dom0 acting as backend)HVM-hypercall PVH-guest build PVH-guest scp
vcpu_op277684 2324event_channel_op350233 57383
(the related dom0 counter values are in the same range as with the test running in the PV guest)Quoted 7 lines
It should be noted that during boot of the guests the numbers for the PV guest are more like the ones for the build test with the exception of iret and
sched_opbeing higher, while for PVHsched_opis by far the most often used hypercall.I'm not sure how to translate those numbers into a good algorithm for generating the call sequence.
Well, there's never going to be a clear cut fitting everything, I suppose.
Quoted 7 lines
Quoted 3 lines
I could add priorities to each hypercall in
hypercall-defs.cand have a cascade of if (likely(foo))call_foo; else if (likely(bla)) ... else switch(rest).Personally I'd lean to an approach like this one; perhaps there's not even a need to specify priorities for every hypercall, but just the ones we deem most frequently used?
See my new series.
Juergen
PATCH 11/12 xen/arm: call hypercall handlers via switch statement
In reply to Juergen GrossInstead of using a function table use the generated switch statement macros for calling the appropriate hypercall handlers.
This makes the calls of the handlers type safe.
For deprecated hypercalls define stub functions.
--- xen/arch/arm/traps.c | 119 ++++++++++--------------------------------- 1 file changed, 26 insertions(+), 93 deletions(-)
xen/arch/arm/traps.c +26 -93
@@ -1337,62 +1337,20 @@ static register_t do_deprecated_hypercall(void)return -ENOSYS ; } -typedef register_t (*arm_hypercall_fn_t )( -register_t ,register_t ,register_t ,register_t ,register_t ); - -typedef struct { -arm_hypercall_fn_t fn; -int nr_args; -}arm_hypercall_t ; - -#define HYPERCALL (_name, _nr_args) \ - [ __HYPERVISOR_ ## _name ] = { \ - .fn = (arm_hypercall_fn_t) &do_ ## _name, \ - .nr_args = _nr_args, \ - } +long dep_sched_op_compat (int cmd,unsigned long arg) +{ +return do_deprecated_hypercall (); +} -/* -* Only use this for hypercalls which were deprecated (i.e. replaced -* by something else) before Xen on ARM was created, i.e. *not* for -* hypercalls which are simply not yet used on ARM. -*/ -#define HYPERCALL_DEPRECATED (_name, _nr_args) \ - [ __HYPERVISOR_##_name ] = { \ - .fn = (arm_hypercall_fn_t) &do_deprecated_hypercall, \ - .nr_args = _nr_args, \ - } +long dep_event_channel_op_compat (XEN_GUEST_HANDLE_PARAM (evtchn_op_t) uop) +{ +return do_deprecated_hypercall (); +} -static arm_hypercall_t arm_hypercall_table[] = { -HYPERCALL (memory_op,2 ), -HYPERCALL (domctl,1 ), -HYPERCALL (sched_op,2 ), -HYPERCALL_DEPRECATED (sched_op_compat,2 ), -HYPERCALL (console_io,3 ), -HYPERCALL (xen_version,2 ), -HYPERCALL (xsm_op,1 ), -HYPERCALL (event_channel_op,2 ), -HYPERCALL_DEPRECATED (event_channel_op_compat,1 ), -HYPERCALL (physdev_op,2 ), -HYPERCALL_DEPRECATED (physdev_op_compat,1 ), -HYPERCALL (sysctl,2 ), -HYPERCALL (hvm_op,2 ), -#ifdef CONFIG_GRANT_TABLE -HYPERCALL (grant_table_op,3 ), -#endif -HYPERCALL (multicall,2 ), -HYPERCALL (platform_op,1 ), -HYPERCALL (vcpu_op,3 ), -HYPERCALL (vm_assist,2 ), -#ifdef CONFIG_ARGO -HYPERCALL (argo_op,5 ), -#endif -#ifdef CONFIG_HYPFS -HYPERCALL (hypfs_op,5 ), -#endif -#ifdef CONFIG_IOREQ_SERVER -HYPERCALL (dm_op,3 ), -#endif -}; +long dep_physdev_op_compat (XEN_GUEST_HANDLE_PARAM (physdev_op_t) uop) +{ +return do_deprecated_hypercall (); +}#ifndef NDEBUG static void do_debug_trap (struct cpu_user_regs *regs,unsigned int code)
@@ -1431,7 +1389,6 @@ static void do_debug_trap(struct cpu_user_regs *regs, unsigned int code)#define HYPERCALL_ARG3 (r) (r)->x2#define HYPERCALL_ARG4 (r) (r)->x3#define HYPERCALL_ARG5 (r) (r)->x4 -#define HYPERCALL_ARGS (r) (r)->x0, (r)->x1, (r)->x2, (r)->x3, (r)->x4#else #define HYPERCALL_RESULT_REG (r) (r)->r0#define HYPERCALL_ARG1 (r) (r)->r0
@@ -1439,52 +1396,40 @@ static void do_debug_trap(struct cpu_user_regs *regs, unsigned int code)#define HYPERCALL_ARG3 (r) (r)->r2#define HYPERCALL_ARG4 (r) (r)->r3#define HYPERCALL_ARG5 (r) (r)->r4 -#define HYPERCALL_ARGS (r) (r)->r0, (r)->r1, (r)->r2, (r)->r3, (r)->r4#endif +static unsigned char hypercall_args[] = hypercall_args_arm; +static void do_trap_hypercall (struct cpu_user_regs *regs,register_t *nr,const union hsr hsr) { -arm_hypercall_fn_t call =NULL ;struct vcpu *curr = current; -BUILD_BUG_ON (NR_hypercalls <ARRAY_SIZE (arm_hypercall_table) ); -if ( hsr.iss !=XEN_HYPERCALL_TAG ) {gprintk (XENLOG_WARNING ,"Invalid HVC imm 0x%x\n" , hsr.iss);return inject_undef_exception (regs, hsr); } -if ( *nr >=ARRAY_SIZE (arm_hypercall_table) ) - { -perfc_incr (invalid_hypercalls); -HYPERCALL_RESULT_REG (regs) = -ENOSYS ; -return ; - } - curr->hcall_preempted = false;perfc_incra (hypercalls, *nr); - call = arm_hypercall_table[*nr].fn; -if ( call ==NULL ) - { -HYPERCALL_RESULT_REG (regs) = -ENOSYS ; -return ; - } -HYPERCALL_RESULT_REG (regs) =call (HYPERCALL_ARGS (regs)); +call_handlers_arm (*nr,HYPERCALL_RESULT_REG (regs),HYPERCALL_ARG1 (regs), +HYPERCALL_ARG2 (regs),HYPERCALL_ARG3 (regs), +HYPERCALL_ARG4 (regs),HYPERCALL_ARG5 (regs));#ifndef NDEBUG -if ( !curr->hcall_preempted ) +if ( !curr->hcall_preempted &&HYPERCALL_RESULT_REG (regs) != -ENOSYS ) {/* Deliberately corrupt parameter regs used by this hypercall. */ -switch ( arm_hypercall_table[*nr].nr_args ) { +switch ( hypercall_args[*nr] ) {case 5 :HYPERCALL_ARG5 (regs) =0xDEADBEEF ;case 4 :HYPERCALL_ARG4 (regs) =0xDEADBEEF ;case 3 :HYPERCALL_ARG3 (regs) =0xDEADBEEF ;case 2 :HYPERCALL_ARG2 (regs) =0xDEADBEEF ;case 1 :/* Don't clobber x0/r0 -- it's the return value */ +case 0 :/* -ENOSYS case */ break ;default :BUG (); }
@@ -1521,7 +1466,10 @@ static bool check_multicall_32bit_clean(struct multicall_entry *multi) {int i; -for ( i =0 ; i < arm_hypercall_table[multi->op].nr_args; i++ ) +if ( multi->op >=ARRAY_SIZE (hypercall_args) ) +return true; + +for ( i =0 ; i < hypercall_args[multi->op]; i++ ) {if (unlikely (multi->args[i] &0xffffffff00000000ULL ) ) {
@@ -1538,28 +1486,13 @@ static bool check_multicall_32bit_clean(struct multicall_entry *multi)enum mc_disposition arch_do_multicall_call (struct mc_state *state) {struct multicall_entry *multi = &state->call; -arm_hypercall_fn_t call =NULL ; - -if ( multi->op >=ARRAY_SIZE (arm_hypercall_table) ) - { - multi->result = -ENOSYS ; -return mc_continue; - } - - call = arm_hypercall_table[multi->op].fn; -if ( call ==NULL ) - { - multi->result = -ENOSYS ; -return mc_continue; - }if (is_32bit_domain (current->domain) && !check_multicall_32bit_clean (multi) )return mc_continue; - multi->result =call (multi->args[0 ], multi->args[1 ], - multi->args[2 ], multi->args[3 ], - multi->args[4 ]); +call_handlers_arm (multi->op, multi->result, multi->args[0 ], multi->args[1 ], + multi->args[2 ], multi->args[3 ], multi->args[4 ]);return likely (!psr_mode_is_user (guest_cpu_user_regs ())) ? mc_continue : mc_preempt;
Signature · 1 line
2.26.2
PATCH 12/12 xen/x86: add hypercall performance counters for hvm, correct pv
In reply to Juergen GrossThe HVM hypercall handler is missing incrementing the per hypercall counters. Add that.
The counters for PV are handled wrong, as they are not using perf_incra() with the number of the hypercall as index, but are incrementing the total number of hypercalls only. Fix that.
--- xen/arch/x86/hvm/hypercall.c | 2 ++ xen/arch/x86/pv/hypercall.c | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-)
xen/arch/x86/hvm/hypercall.c +2 -0
@@ -214,6 +214,8 @@ int hvm_hypercall(struct cpu_user_regs *regs)ioreq_signal_mapcache_invalidate (); } +perfc_incra (hypercalls, eax); +return curr->hcall_preempted ? HVM_HCALL_preempted : HVM_HCALL_completed; }
xen/arch/x86/pv/hypercall.c +1 -1
@@ -106,7 +106,7 @@ _pv_hypercall(struct cpu_user_regs *regs, bool compat)if ( curr->hcall_preempted ) regs->rip -=2 ; -perfc_incr (hypercalls); +perfc_incra (hypercalls, eax); }enum mc_disposition pv_do_multicall_call (struct mc_state *state)
Signature · 1 line
2.26.2
Re: [PATCH 12/12] xen/x86: add hypercall performance counters for hvm, correct pv
Quoted 6 lines · Juergen Gross
The HVM hypercall handler is missing incrementing the per hypercall counters. Add that.
The counters for PV are handled wrong, as they are not using
perf_incra()with the number of the hypercall as index, but are incrementing the total number of hypercalls only. Fix that.
Why do you say "total number"? Isn't it that all accounting goes into set_trap_table's slot, effectively making that slot a "total number" despite not being labeled that way?
Also this fix renders largely redundant the calls_to_multicall counter. Could I talk you into deleting that at the same time? (As to the "not fully redundant": I consider it suspicious that this counter gets incremented at the bottom of the function, not at the top.)
Finally I take it that with the Kconfig setting being under DEBUG, we don't consider security supported builds with PERF_COUNTERS enabled. Otherwise as a prereq I would think perfc_incra() would need teaching of array_index_nospec().
In any event, preferably with at least the description slightly adjusted,
Jan
Re: [PATCH 12/12] xen/x86: add hypercall performance counters for hvm, correct pv
Quoted 11 lines · Jan Beulich
Quoted 6 lines · Juergen Gross
The HVM hypercall handler is missing incrementing the per hypercall counters. Add that.
The counters for PV are handled wrong, as they are not using
perf_incra()with the number of the hypercall as index, but are incrementing the total number of hypercalls only. Fix that.Why do you say "total number"? Isn't it that all accounting goes into set_trap_table's slot, effectively making that slot a "total number" despite not being labeled that way?
Oh, right. Will correct.
Quoted 4 lines
Also this fix renders largely redundant the
calls_to_multicallcounter. Could I talk you into deleting that at the same time? (As to the "not fully redundant": I consider it suspicious that this counter gets incremented at the bottom of the function, not at the top.)
I think I'll just another patch doing that.
Quoted 4 lines
Finally I take it that with the
Kconfigsetting being under DEBUG, we don't consider security supported builds withPERF_COUNTERSenabled. Otherwise as a prereq I would thinkperfc_incra()would need teaching ofarray_index_nospec().
I agree.
Quoted 3 lines
In any event, preferably with at least the description slightly adjusted,
Reviewed-by Jan Beulich <[email protected]>
Thanks.
Juergen