Content-Type: multipart/mixed; boundary="===============6009609819179501338==" MIME-Version: 1.0 From: kernel test robot To: kbuild-all@lists.01.org Subject: Re: [PATCH 3/5] KVM: x86: Use __try_cmpxchg_user() to update guest PTE A/D bits Date: Tue, 01 Feb 2022 15:01:16 +0800 Message-ID: <202202011400.EaZmWZ48-lkp@intel.com> In-Reply-To: <20220201010838.1494405-4-seanjc@google.com> List-Id: --===============6009609819179501338== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Sean, I love your patch! Yet something to improve: [auto build test ERROR on 26291c54e111ff6ba87a164d85d4a4e134b7315c] url: https://github.com/0day-ci/linux/commits/Sean-Christopherson/x86-ua= ccess-CMPXCHG-KVM-bug-fixes/20220201-091001 base: 26291c54e111ff6ba87a164d85d4a4e134b7315c config: i386-randconfig-a002 (https://download.01.org/0day-ci/archive/20220= 201/202202011400.EaZmWZ48-lkp(a)intel.com/config) compiler: clang version 14.0.0 (https://github.com/llvm/llvm-project 6b1e84= 4b69f15bb7dffaf9365cd2b355d2eb7579) reproduce (this is a W=3D1 build): wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/= make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # https://github.com/0day-ci/linux/commit/c880d7a9df876f20dc3acdd89= 3c5c71f3cda5029 git remote add linux-review https://github.com/0day-ci/linux git fetch --no-tags linux-review Sean-Christopherson/x86-uaccess-CM= PXCHG-KVM-bug-fixes/20220201-091001 git checkout c880d7a9df876f20dc3acdd893c5c71f3cda5029 # save the config file to linux build tree mkdir build_dir COMPILER_INSTALL_PATH=3D$HOME/0day COMPILER=3Dclang make.cross W=3D= 1 O=3Dbuild_dir ARCH=3Di386 SHELL=3D/bin/bash If you fix the issue, kindly add following tag as appropriate Reported-by: kernel test robot All errors (new ones prefixed by >>): In file included from arch/x86/kvm/mmu/mmu.c:4246: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' ret =3D __try_cmpxchg_user(ptep_user, &orig_pte, pte, fa= ult); ^ arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:606:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 1: __ret =3D __try_cmpxchg_user_asm("b", "q", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ In file included from arch/x86/kvm/mmu/mmu.c:4246: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:610:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 2: __ret =3D __try_cmpxchg_user_asm("w", "r", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ In file included from arch/x86/kvm/mmu/mmu.c:4246: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:614:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 4: __ret =3D __try_cmpxchg_user_asm("l", "r", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ In file included from arch/x86/kvm/mmu/mmu.c:4250: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' ret =3D __try_cmpxchg_user(ptep_user, &orig_pte, pte, fa= ult); ^ arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:606:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 1: __ret =3D __try_cmpxchg_user_asm("b", "q", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ In file included from arch/x86/kvm/mmu/mmu.c:4250: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:610:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 2: __ret =3D __try_cmpxchg_user_asm("w", "r", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ In file included from arch/x86/kvm/mmu/mmu.c:4250: >> arch/x86/kvm/mmu/paging_tmpl.h:244:9: error: invalid output size for con= straint '+a' arch/x86/include/asm/uaccess.h:629:11: note: expanded from macro '__try_= cmpxchg_user' __ret =3D !unsafe_try_cmpxchg_user(_ptr, _oldp, _nval, _label); = \ ^ arch/x86/include/asm/uaccess.h:614:18: note: expanded from macro 'unsafe= _try_cmpxchg_user' case 4: __ret =3D __try_cmpxchg_user_asm("l", "r", = \ ^ arch/x86/include/asm/uaccess.h:467:22: note: expanded from macro '__try_= cmpxchg_user_asm' [old] "+a" (__old) \ ^ 6 errors generated. vim +244 arch/x86/kvm/mmu/paging_tmpl.h 191 = 192 static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu, 193 struct kvm_mmu *mmu, 194 struct guest_walker *walker, 195 gpa_t addr, int write_fault) 196 { 197 unsigned level, index; 198 pt_element_t pte, orig_pte; 199 pt_element_t __user *ptep_user; 200 gfn_t table_gfn; 201 int ret; 202 = 203 /* dirty/accessed bits are not supported, so no need to update them= */ 204 if (!PT_HAVE_ACCESSED_DIRTY(mmu)) 205 return 0; 206 = 207 for (level =3D walker->max_level; level >=3D walker->level; --level= ) { 208 pte =3D orig_pte =3D walker->ptes[level - 1]; 209 table_gfn =3D walker->table_gfn[level - 1]; 210 ptep_user =3D walker->ptep_user[level - 1]; 211 index =3D offset_in_page(ptep_user) / sizeof(pt_element_t); 212 if (!(pte & PT_GUEST_ACCESSED_MASK)) { 213 trace_kvm_mmu_set_accessed_bit(table_gfn, index, sizeof(pte)); 214 pte |=3D PT_GUEST_ACCESSED_MASK; 215 } 216 if (level =3D=3D walker->level && write_fault && 217 !(pte & PT_GUEST_DIRTY_MASK)) { 218 trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte)); 219 #if PTTYPE =3D=3D PTTYPE_EPT 220 if (kvm_x86_ops.nested_ops->write_log_dirty(vcpu, addr)) 221 return -EINVAL; 222 #endif 223 pte |=3D PT_GUEST_DIRTY_MASK; 224 } 225 if (pte =3D=3D orig_pte) 226 continue; 227 = 228 /* 229 * If the slot is read-only, simply do not process the accessed 230 * and dirty bits. This is the correct thing to do if the slot 231 * is ROM, and page tables in read-as-ROM/write-as-MMIO slots 232 * are only supported if the accessed and dirty bits are already 233 * set in the ROM (so that MMIO writes are never needed). 234 * 235 * Note that NPT does not allow this at all and faults, since 236 * it always wants nested page table entries for the guest 237 * page tables to be writable. And EPT works but will simply 238 * overwrite the read-only memory to set the accessed and dirty 239 * bits. 240 */ 241 if (unlikely(!walker->pte_writable[level - 1])) 242 continue; 243 = > 244 ret =3D __try_cmpxchg_user(ptep_user, &orig_pte, pte, fault); 245 if (ret) 246 return ret; 247 = 248 kvm_vcpu_mark_page_dirty(vcpu, table_gfn); 249 walker->ptes[level - 1] =3D pte; 250 } 251 return 0; 252 } 253 = --- 0-DAY CI Kernel Test Service, Intel Corporation https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org --===============6009609819179501338==--