| Message ID | 1497402776-22348-9-git-send-email-yi.y.sun@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
>>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> > @@ -179,6 +182,10 @@ static void free_socket_resources(unsigned int socket) > } > > info->feat_init = false; > + > + memset(info->cos_ref, 0, MAX_COS_REG_CNT * sizeof(unsigned int)); > + > + bitmap_zero(info->dom_set, DOMID_IDLE + 1); > } I can see the point of the latter (as you add the new structure member), but if the former is necessary, shouldn't it have been done in an earlier patch? Or should be field only be introduced here? > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, > return -ENOENT; > } > > + domain_lock(d); > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) > + { > + d->arch.psr_cos_ids[socket] = 0; > + set_bit(d->domain_id, socket_info[socket].dom_set); > + } Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? Or wait - I think it's even wrong to split the test from the set, as the lock doesn't protect dom_set[]. > +int psr_set_val(struct domain *d, unsigned int socket, > + uint64_t new_val, enum cbm_type type) > +{ > + unsigned int old_cos, array_len; > + int cos, ret; > + unsigned int *ref; > + uint32_t *val_array, val; > + struct psr_socket_info *info = get_socket_info(socket); > + enum psr_feat_type feat_type; > + > + if ( IS_ERR(info) ) > + return PTR_ERR(info); > + > + val = new_val; > + if ( new_val != val ) > + return -EINVAL; > + > + feat_type = psr_cbm_type_to_feat_type(type); > + if ( feat_type >= ARRAY_SIZE(info->features) || > + !info->features[feat_type] ) > + return -ENOENT; > + > + /* > + * Step 0: > + * old_cos means the COS ID current domain is using. By default, it is 0. > + * > + * For every COS ID, there is a reference count to record how many domains > + * are using the COS register corresponding to this COS ID. > + * - If ref[old_cos] is 0, that means this COS is not used by any domain. > + * - If ref[old_cos] is 1, that means this COS is only used by current > + * domain. > + * - If ref[old_cos] is more than 1, that mean multiple domains are using > + * this COS. > + */ > + domain_lock(d); > + if ( !test_bit(d->domain_id, info->dom_set) ) > + { > + d->arch.psr_cos_ids[socket] = 0; > + set_bit(d->domain_id, info->dom_set); > + } Same here. > + old_cos = d->arch.psr_cos_ids[socket]; > + domain_unlock(d); > + > + ASSERT(old_cos < MAX_COS_REG_CNT); > + > + ref = info->cos_ref; > + > + /* > + * Step 1: > + * Gather a value array to store all features cos_reg_val[old_cos]. > + * And, set the input new val into array according to the feature's > + * position in array. > + */ > + array_len = get_cos_num(info); > + val_array = xzalloc_array(uint32_t, array_len); > + if ( !val_array ) > + return -ENOMEM; > + > + if ( (ret = gather_val_array(val_array, array_len, info, old_cos)) != 0 ) > + goto free_array; > + > + if ( (ret = insert_val_into_array(val_array, array_len, info, > + feat_type, type, val)) != 0 ) > + goto free_array; > + > + spin_lock(&info->ref_lock); > + > + /* > + * Step 2: > + * Try to find if there is already a COS ID on which all features' values > + * are same as the array. Then, we can reuse this COS ID. > + */ > + cos = find_cos(val_array, array_len, feat_type, info); > + if ( cos == old_cos ) > + { > + ret = 0; > + goto unlock_free_array; > + } > + > + /* > + * Step 3: > + * If fail to find, we need pick an available COS ID. > + * In fact, only COS ID which ref is 1 or 0 can be picked for current > + * domain. If old_cos is not 0 and its ref==1, that means only current > + * domain is using this old_cos ID. So, this old_cos ID certainly can > + * be reused by current domain. Ref==0 means there is no any domain > + * using this COS ID. So it can be used for current domain too. > + */ > + if ( cos < 0 ) > + { > + cos = pick_avail_cos(info, val_array, array_len, old_cos, feat_type); > + if ( cos < 0 ) > + { > + ret = cos; > + goto unlock_free_array; > + } > + > + /* > + * Step 4: > + * Write the feature's MSRs according to the COS ID. > + */ > + ret = write_psr_msrs(socket, cos, val_array, array_len, feat_type); > + if ( ret ) > + goto unlock_free_array; > + } > + > + /* > + * Step 5: > + * Find the COS ID (find_cos result is '>= 0' or an available COS ID is > + * picked, then update ref according to COS ID. > + */ > + ref[cos]++; > + ASSERT(!cos || ref[cos]); > + ASSERT(!old_cos || ref[old_cos]); > + ref[old_cos]--; > + spin_unlock(&info->ref_lock); > + > + /* > + * Step 6: > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know > + * which COS the domain is using on the socket. One domain can only use > + * one COS ID at same time on each socket. > + */ > + domain_lock(d); > + d->arch.psr_cos_ids[socket] = cos; > + domain_unlock(d); > + > + /* > + * Step 7: > + * Then, set the dom_set bit which corresponds to domain_id to mark this > + * domain has been set and the COS ID of the domain is valid. > + */ > + set_bit(d->domain_id, info->dom_set); With the way things are being done above, doesn't this belong in the domain_lock()-ed region? Jan
On 17-06-28 01:14:59, Jan Beulich wrote: > >>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> > > @@ -179,6 +182,10 @@ static void free_socket_resources(unsigned int socket) > > } > > > > info->feat_init = false; > > + > > + memset(info->cos_ref, 0, MAX_COS_REG_CNT * sizeof(unsigned int)); > > + > > + bitmap_zero(info->dom_set, DOMID_IDLE + 1); > > } > > I can see the point of the latter (as you add the new structure member), but > if the former is necessary, shouldn't it have been done in an earlier patch? > Or should be field only be introduced here? > You are right. I will move the cos_ref clearing into earlier patch. > > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, > > return -ENOENT; > > } > > > > + domain_lock(d); > > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) > > + { > > + d->arch.psr_cos_ids[socket] = 0; > > + set_bit(d->domain_id, socket_info[socket].dom_set); > > + } > > Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? > Or wait - I think it's even wrong to split the test from the set, as the lock > doesn't protect dom_set[]. > Will change it to test_and_set_bit. > > +int psr_set_val(struct domain *d, unsigned int socket, > > + uint64_t new_val, enum cbm_type type) > > +{ > > + unsigned int old_cos, array_len; > > + int cos, ret; > > + unsigned int *ref; > > + uint32_t *val_array, val; > > + struct psr_socket_info *info = get_socket_info(socket); > > + enum psr_feat_type feat_type; > > + > > + if ( IS_ERR(info) ) > > + return PTR_ERR(info); > > + > > + val = new_val; > > + if ( new_val != val ) > > + return -EINVAL; > > + > > + feat_type = psr_cbm_type_to_feat_type(type); > > + if ( feat_type >= ARRAY_SIZE(info->features) || > > + !info->features[feat_type] ) > > + return -ENOENT; > > + > > + /* > > + * Step 0: > > + * old_cos means the COS ID current domain is using. By default, it is 0. > > + * > > + * For every COS ID, there is a reference count to record how many domains > > + * are using the COS register corresponding to this COS ID. > > + * - If ref[old_cos] is 0, that means this COS is not used by any domain. > > + * - If ref[old_cos] is 1, that means this COS is only used by current > > + * domain. > > + * - If ref[old_cos] is more than 1, that mean multiple domains are using > > + * this COS. > > + */ > > + domain_lock(d); > > + if ( !test_bit(d->domain_id, info->dom_set) ) > > + { > > + d->arch.psr_cos_ids[socket] = 0; > > + set_bit(d->domain_id, info->dom_set); > > + } > > Same here. > > > + old_cos = d->arch.psr_cos_ids[socket]; > > + domain_unlock(d); > > + > > + ASSERT(old_cos < MAX_COS_REG_CNT); > > + > > + ref = info->cos_ref; > > + > > + /* > > + * Step 1: > > + * Gather a value array to store all features cos_reg_val[old_cos]. > > + * And, set the input new val into array according to the feature's > > + * position in array. > > + */ > > + array_len = get_cos_num(info); > > + val_array = xzalloc_array(uint32_t, array_len); > > + if ( !val_array ) > > + return -ENOMEM; > > + > > + if ( (ret = gather_val_array(val_array, array_len, info, old_cos)) != 0 ) > > + goto free_array; > > + > > + if ( (ret = insert_val_into_array(val_array, array_len, info, > > + feat_type, type, val)) != 0 ) > > + goto free_array; > > + > > + spin_lock(&info->ref_lock); > > + > > + /* > > + * Step 2: > > + * Try to find if there is already a COS ID on which all features' values > > + * are same as the array. Then, we can reuse this COS ID. > > + */ > > + cos = find_cos(val_array, array_len, feat_type, info); > > + if ( cos == old_cos ) > > + { > > + ret = 0; > > + goto unlock_free_array; > > + } > > + > > + /* > > + * Step 3: > > + * If fail to find, we need pick an available COS ID. > > + * In fact, only COS ID which ref is 1 or 0 can be picked for current > > + * domain. If old_cos is not 0 and its ref==1, that means only current > > + * domain is using this old_cos ID. So, this old_cos ID certainly can > > + * be reused by current domain. Ref==0 means there is no any domain > > + * using this COS ID. So it can be used for current domain too. > > + */ > > + if ( cos < 0 ) > > + { > > + cos = pick_avail_cos(info, val_array, array_len, old_cos, feat_type); > > + if ( cos < 0 ) > > + { > > + ret = cos; > > + goto unlock_free_array; > > + } > > + > > + /* > > + * Step 4: > > + * Write the feature's MSRs according to the COS ID. > > + */ > > + ret = write_psr_msrs(socket, cos, val_array, array_len, feat_type); > > + if ( ret ) > > + goto unlock_free_array; > > + } > > + > > + /* > > + * Step 5: > > + * Find the COS ID (find_cos result is '>= 0' or an available COS ID is > > + * picked, then update ref according to COS ID. > > + */ > > + ref[cos]++; > > + ASSERT(!cos || ref[cos]); > > + ASSERT(!old_cos || ref[old_cos]); > > + ref[old_cos]--; > > + spin_unlock(&info->ref_lock); > > + > > + /* > > + * Step 6: > > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know > > + * which COS the domain is using on the socket. One domain can only use > > + * one COS ID at same time on each socket. > > + */ > > + domain_lock(d); > > + d->arch.psr_cos_ids[socket] = cos; > > + domain_unlock(d); > > + > > + /* > > + * Step 7: > > + * Then, set the dom_set bit which corresponds to domain_id to mark this > > + * domain has been set and the COS ID of the domain is valid. > > + */ > > + set_bit(d->domain_id, info->dom_set); > > With the way things are being done above, doesn't this belong in the > domain_lock()-ed region? > Yes, should be. Thanks! > Jan
>>> Yi Sun <yi.y.sun@linux.intel.com> 06/28/17 11:10 AM >>> >On 17-06-28 01:14:59, Jan Beulich wrote: >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> >> > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, >> > return -ENOENT; >> > } >> > >> > + domain_lock(d); >> > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) >> > + { >> > + d->arch.psr_cos_ids[socket] = 0; >> > + set_bit(d->domain_id, socket_info[socket].dom_set); >> > + } >> >> Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? >> Or wait - I think it's even wrong to split the test from the set, as the lock >> doesn't protect dom_set[]. With the last sentence here (which I had added after having written all of the rest of the reply, I'm afraid I've managed to confuse you: >>> >>Will change it to test_and_set_bit. >... >> > + /* >> > + * Step 6: >> > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know >> > + * which COS the domain is using on the socket. One domain can only use >> > + * one COS ID at same time on each socket. >> > + */ >> > + domain_lock(d); >> > + d->arch.psr_cos_ids[socket] = cos; >> > + domain_unlock(d); >> > + >> > + /* >> > + * Step 7: >> > + * Then, set the dom_set bit which corresponds to domain_id to mark this >> > + * domain has been set and the COS ID of the domain is valid. >> > + */ >> > + set_bit(d->domain_id, info->dom_set); >> >> With the way things are being done above, doesn't this belong in the >> domain_lock()-ed region? I should have deleted this, since - as said above - the lock doesn't guard against anything dom_set[]-wise. So ... >Yes, should be. Thanks! ... I think you rather shouldn't do this. Instead you may want to consider whether the other domain_lock()-ed regions couldn't be further shrunk. Jan
On 17-06-28 05:43:58, Jan Beulich wrote: > >>> Yi Sun <yi.y.sun@linux.intel.com> 06/28/17 11:10 AM >>> > >On 17-06-28 01:14:59, Jan Beulich wrote: > >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> > >> > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, > >> > return -ENOENT; > >> > } > >> > > >> > + domain_lock(d); > >> > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) > >> > + { > >> > + d->arch.psr_cos_ids[socket] = 0; > >> > + set_bit(d->domain_id, socket_info[socket].dom_set); > >> > + } > >> > >> Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? > >> Or wait - I think it's even wrong to split the test from the set, as the lock > >> doesn't protect dom_set[]. > > With the last sentence here (which I had added after having written all of the > rest of the reply, I'm afraid I've managed to confuse you: > > >>> > >>Will change it to test_and_set_bit. > >... > >> > + /* > >> > + * Step 6: > >> > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know > >> > + * which COS the domain is using on the socket. One domain can only use > >> > + * one COS ID at same time on each socket. > >> > + */ > >> > + domain_lock(d); > >> > + d->arch.psr_cos_ids[socket] = cos; > >> > + domain_unlock(d); > >> > + > >> > + /* > >> > + * Step 7: > >> > + * Then, set the dom_set bit which corresponds to domain_id to mark this > >> > + * domain has been set and the COS ID of the domain is valid. > >> > + */ > >> > + set_bit(d->domain_id, info->dom_set); > >> > >> With the way things are being done above, doesn't this belong in the > >> domain_lock()-ed region? > > I should have deleted this, since - as said above - the lock doesn't guard > against anything dom_set[]-wise. So ... > > >Yes, should be. Thanks! > > ... I think you rather shouldn't do this. Instead you may want to consider whether > the other domain_lock()-ed regions couldn't be further shrunk. > I want to confirm below two points with you: 1. remove this 'set_bit' here if above 'test_bit' is replaced to 'test_and_set_bit'. 2. For the 'be further shrunk', I think the 'domain_lock' above 'set_bit' can be removed if 'test_and_set_bit' is used.
>>> Yi Sun <yi.y.sun@linux.intel.com> 06/29/17 7:12 AM >>> >On 17-06-28 05:43:58, Jan Beulich wrote: >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/28/17 11:10 AM >>> >> >On 17-06-28 01:14:59, Jan Beulich wrote: >> >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> >> >> > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, >> >> > return -ENOENT; >> >> > } >> >> > >> >> > + domain_lock(d); >> >> > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) >> >> > + { >> >> > + d->arch.psr_cos_ids[socket] = 0; >> >> > + set_bit(d->domain_id, socket_info[socket].dom_set); >> >> > + } >> >> >> >> Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? >> >> Or wait - I think it's even wrong to split the test from the set, as the lock >> >> doesn't protect dom_set[]. >> >> With the last sentence here (which I had added after having written all of the >> rest of the reply, I'm afraid I've managed to confuse you: >> >> >>> >> >>Will change it to test_and_set_bit. >> >... >> >> > + /* >> >> > + * Step 6: >> >> > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know >> >> > + * which COS the domain is using on the socket. One domain can only use >> >> > + * one COS ID at same time on each socket. >> >> > + */ >> >> > + domain_lock(d); >> >> > + d->arch.psr_cos_ids[socket] = cos; >> >> > + domain_unlock(d); >> >> > + >> >> > + /* >> >> > + * Step 7: >> >> > + * Then, set the dom_set bit which corresponds to domain_id to mark this >> >> > + * domain has been set and the COS ID of the domain is valid. >> >> > + */ >> >> > + set_bit(d->domain_id, info->dom_set); >> >> >> >> With the way things are being done above, doesn't this belong in the >> >> domain_lock()-ed region? >> >> I should have deleted this, since - as said above - the lock doesn't guard >> against anything dom_set[]-wise. So ... >> >> >Yes, should be. Thanks! >> >> ... I think you rather shouldn't do this. Instead you may want to consider whether >> the other domain_lock()-ed regions couldn't be further shrunk. >> >I want to confirm below two points with you: >1. remove this 'set_bit' here if above 'test_bit' is replaced to >'test_and_set_bit'. I don't think so, at least not for the ones still visible in context here. I've only suggested to convert test/set pairs into test_and_set. The one at step 7 doesn't have a test next to it, so either it was redundant with some other set (in which case it should indeed be dropped), or it needs to stay as is. >2. For the 'be further shrunk', I think the 'domain_lock' above 'set_bit' can be >removed if 'test_and_set_bit' is used. I don't think it can be removed altogether, but I think it could be moved into the body of the if(). Jan
On 17-06-29 00:24:50, Jan Beulich wrote: > >>> Yi Sun <yi.y.sun@linux.intel.com> 06/29/17 7:12 AM >>> > >On 17-06-28 05:43:58, Jan Beulich wrote: > >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/28/17 11:10 AM >>> > >> >On 17-06-28 01:14:59, Jan Beulich wrote: > >> >> >>> Yi Sun <yi.y.sun@linux.intel.com> 06/14/17 3:25 AM >>> > >> >> > @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, > >> >> > return -ENOENT; > >> >> > } > >> >> > > >> >> > + domain_lock(d); > >> >> > + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) > >> >> > + { > >> >> > + d->arch.psr_cos_ids[socket] = 0; > >> >> > + set_bit(d->domain_id, socket_info[socket].dom_set); > >> >> > + } > >> >> > >> >> Any reason not to use test_and_set_bit() here? I.e. is this on any hot path? > >> >> Or wait - I think it's even wrong to split the test from the set, as the lock > >> >> doesn't protect dom_set[]. > >> > >> With the last sentence here (which I had added after having written all of the > >> rest of the reply, I'm afraid I've managed to confuse you: > >> > >> >>> > >> >>Will change it to test_and_set_bit. > >> >... > >> >> > + /* > >> >> > + * Step 6: > >> >> > + * Save the COS ID into current domain's psr_cos_ids[] so that we can know > >> >> > + * which COS the domain is using on the socket. One domain can only use > >> >> > + * one COS ID at same time on each socket. > >> >> > + */ > >> >> > + domain_lock(d); > >> >> > + d->arch.psr_cos_ids[socket] = cos; > >> >> > + domain_unlock(d); > >> >> > + > >> >> > + /* > >> >> > + * Step 7: > >> >> > + * Then, set the dom_set bit which corresponds to domain_id to mark this > >> >> > + * domain has been set and the COS ID of the domain is valid. > >> >> > + */ > >> >> > + set_bit(d->domain_id, info->dom_set); > >> >> > >> >> With the way things are being done above, doesn't this belong in the > >> >> domain_lock()-ed region? > >> > >> I should have deleted this, since - as said above - the lock doesn't guard > >> against anything dom_set[]-wise. So ... > >> > >> >Yes, should be. Thanks! > >> > >> ... I think you rather shouldn't do this. Instead you may want to consider whether > >> the other domain_lock()-ed regions couldn't be further shrunk. > >> > >I want to confirm below two points with you: > >1. remove this 'set_bit' here if above 'test_bit' is replaced to > >'test_and_set_bit'. > > I don't think so, at least not for the ones still visible in context here. I've only > suggested to convert test/set pairs into test_and_set. The one at step 7 doesn't > have a test next to it, so either it was redundant with some other set (in which > case it should indeed be dropped), or it needs to stay as is. > For the 'set_bit' at Step 7, it is redundant because the bit has been set anyway when entering 'psr_set_val' if we use 'test_and_set_bit' there. So, I think we can drop Step 7. > >2. For the 'be further shrunk', I think the 'domain_lock' above 'set_bit' can be > >removed if 'test_and_set_bit' is used. > > I don't think it can be removed altogether, but I think it could be moved into the > body of the if(). > I think we still need to keep the lock protection range in current codes. We need the lock to protect the action to get 'cos id' too. There is a scenario to explain this: if the domain's bit in dom_set has been cleared, 'psr_get_val' and 'psr_set_val' are called almost at same time, but 'psr_get_val' is a little bit eariler. It sets dom_set bit to 1 firstly. At that time, 'psr_set_val' checks the bit and finds it has been set, it goes to next instruction to get old_cos. But the 'psr_get_val' may not restore the cos id to 0 yet. So, the old_cos got in 'psr_set_val' is wrong. To avoid this, I think should use current codes to protect the whole range. psr_get_val() psr_set_val() //old bit is 0, enter statement. if (!test_and_set_bit()) { //old bit is 1, skip statement. if (!test_and_set_bit()) old_cos = d->arch.psr_cos_ids[socket]; domain_lock(); d->arch.psr_cos_ids[socket] = 0; domain_unlock(); }
diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c index 5b62c5c..124a1c6 100644 --- a/xen/arch/x86/domctl.c +++ b/xen/arch/x86/domctl.c @@ -1411,21 +1411,21 @@ long arch_do_domctl( uint32_t val32; case XEN_DOMCTL_PSR_CAT_OP_SET_L3_CBM: - ret = psr_set_l3_cbm(d, domctl->u.psr_cat_op.target, - domctl->u.psr_cat_op.data, - PSR_CBM_TYPE_L3); + ret = psr_set_val(d, domctl->u.psr_cat_op.target, + domctl->u.psr_cat_op.data, + PSR_CBM_TYPE_L3); break; case XEN_DOMCTL_PSR_CAT_OP_SET_L3_CODE: - ret = psr_set_l3_cbm(d, domctl->u.psr_cat_op.target, - domctl->u.psr_cat_op.data, - PSR_CBM_TYPE_L3_CODE); + ret = psr_set_val(d, domctl->u.psr_cat_op.target, + domctl->u.psr_cat_op.data, + PSR_CBM_TYPE_L3_CODE); break; case XEN_DOMCTL_PSR_CAT_OP_SET_L3_DATA: - ret = psr_set_l3_cbm(d, domctl->u.psr_cat_op.target, - domctl->u.psr_cat_op.data, - PSR_CBM_TYPE_L3_DATA); + ret = psr_set_val(d, domctl->u.psr_cat_op.target, + domctl->u.psr_cat_op.data, + PSR_CBM_TYPE_L3_DATA); break; case XEN_DOMCTL_PSR_CAT_OP_GET_L3_CBM: diff --git a/xen/arch/x86/psr.c b/xen/arch/x86/psr.c index 7583f4d..fb84223 100644 --- a/xen/arch/x86/psr.c +++ b/xen/arch/x86/psr.c @@ -117,6 +117,7 @@ static const struct feat_props { * ref_lock - A lock to protect cos_ref. * cos_ref - A reference count array to record how many domains are using the * COS ID. Every entry of cos_ref corresponds to one COS ID. + * dom_set - A bitmap to indicate which domain's cos id has been set. */ struct psr_socket_info { bool feat_init; @@ -124,6 +125,8 @@ struct psr_socket_info { struct feat_node *features[PSR_SOCKET_FEAT_NUM]; spinlock_t ref_lock; unsigned int cos_ref[MAX_COS_REG_CNT]; + /* Every bit corresponds to a domain. Index is domain_id. */ + DECLARE_BITMAP(dom_set, DOMID_IDLE + 1); }; struct psr_assoc { @@ -179,6 +182,10 @@ static void free_socket_resources(unsigned int socket) } info->feat_init = false; + + memset(info->cos_ref, 0, MAX_COS_REG_CNT * sizeof(unsigned int)); + + bitmap_zero(info->dom_set, DOMID_IDLE + 1); } static enum psr_feat_type psr_cbm_type_to_feat_type(enum cbm_type type) @@ -449,13 +456,25 @@ void psr_ctxt_switch_to(struct domain *d) if ( psr_cmt_enabled() ) psr_assoc_rmid(®, d->arch.psr_rmid); - /* If domain's 'psr_cos_ids' is NULL, we set default value for it. */ + /* + * If the domain is not set in 'dom_set' bitmap, that means the domain's + * cos id is not valid. So, we have to use default value (0) to set ASSOC + * register. Furthermore, if domain's 'psr_cos_ids' is NULL, we need + * default value for it too (for case that the domain's psr_cos_ids is not + * successfully allocated). + */ if ( psra->cos_mask ) - reg = psr_assoc_cos(reg, - (d->arch.psr_cos_ids ? - d->arch.psr_cos_ids[cpu_to_socket(smp_processor_id())] : - 0), - psra->cos_mask); + { + unsigned int socket = cpu_to_socket(smp_processor_id()); + struct psr_socket_info *info = socket_info + socket; + unsigned int cos = 0; + + if ( likely(test_bit(d->domain_id, info->dom_set)) && + d->arch.psr_cos_ids ) + cos = d->arch.psr_cos_ids[socket]; + + reg = psr_assoc_cos(reg, cos, psra->cos_mask); + } if ( reg != psra->val ) { @@ -537,7 +556,16 @@ int psr_get_val(struct domain *d, unsigned int socket, return -ENOENT; } + domain_lock(d); + if ( !test_bit(d->domain_id, socket_info[socket].dom_set) ) + { + d->arch.psr_cos_ids[socket] = 0; + set_bit(d->domain_id, socket_info[socket].dom_set); + } + cos = d->arch.psr_cos_ids[socket]; + domain_unlock(d); + /* * If input cos exceeds current feature's cos_max, we should return its * default value which is stored in cos 0. This case only happens @@ -561,15 +589,224 @@ int psr_get_val(struct domain *d, unsigned int socket, return -EINVAL; } -int psr_set_l3_cbm(struct domain *d, unsigned int socket, - uint64_t cbm, enum cbm_type type) +/* Set value functions */ +static unsigned int get_cos_num(const struct psr_socket_info *info) { return 0; } -/* Called with domain lock held, no extra lock needed for 'psr_cos_ids' */ +static int gather_val_array(uint32_t val[], + unsigned int array_len, + const struct psr_socket_info *info, + unsigned int old_cos) +{ + return -EINVAL; +} + +static int insert_val_into_array(uint32_t val[], + unsigned int array_len, + const struct psr_socket_info *info, + enum psr_feat_type feat_type, + enum cbm_type type, + uint32_t new_val) +{ + return -EINVAL; +} + +static int find_cos(const uint32_t val[], unsigned int array_len, + enum psr_feat_type feat_type, + const struct psr_socket_info *info) +{ + return -ENOENT; +} + +static int pick_avail_cos(const struct psr_socket_info *info, + const uint32_t val[], unsigned int array_len, + unsigned int old_cos, + enum psr_feat_type feat_type) +{ + return -ENOENT; +} + +static int write_psr_msrs(unsigned int socket, unsigned int cos, + uint32_t val[], unsigned int array_len, + enum psr_feat_type feat_type) +{ + return -ENOENT; +} + +int psr_set_val(struct domain *d, unsigned int socket, + uint64_t new_val, enum cbm_type type) +{ + unsigned int old_cos, array_len; + int cos, ret; + unsigned int *ref; + uint32_t *val_array, val; + struct psr_socket_info *info = get_socket_info(socket); + enum psr_feat_type feat_type; + + if ( IS_ERR(info) ) + return PTR_ERR(info); + + val = new_val; + if ( new_val != val ) + return -EINVAL; + + feat_type = psr_cbm_type_to_feat_type(type); + if ( feat_type >= ARRAY_SIZE(info->features) || + !info->features[feat_type] ) + return -ENOENT; + + /* + * Step 0: + * old_cos means the COS ID current domain is using. By default, it is 0. + * + * For every COS ID, there is a reference count to record how many domains + * are using the COS register corresponding to this COS ID. + * - If ref[old_cos] is 0, that means this COS is not used by any domain. + * - If ref[old_cos] is 1, that means this COS is only used by current + * domain. + * - If ref[old_cos] is more than 1, that mean multiple domains are using + * this COS. + */ + domain_lock(d); + if ( !test_bit(d->domain_id, info->dom_set) ) + { + d->arch.psr_cos_ids[socket] = 0; + set_bit(d->domain_id, info->dom_set); + } + + old_cos = d->arch.psr_cos_ids[socket]; + domain_unlock(d); + + ASSERT(old_cos < MAX_COS_REG_CNT); + + ref = info->cos_ref; + + /* + * Step 1: + * Gather a value array to store all features cos_reg_val[old_cos]. + * And, set the input new val into array according to the feature's + * position in array. + */ + array_len = get_cos_num(info); + val_array = xzalloc_array(uint32_t, array_len); + if ( !val_array ) + return -ENOMEM; + + if ( (ret = gather_val_array(val_array, array_len, info, old_cos)) != 0 ) + goto free_array; + + if ( (ret = insert_val_into_array(val_array, array_len, info, + feat_type, type, val)) != 0 ) + goto free_array; + + spin_lock(&info->ref_lock); + + /* + * Step 2: + * Try to find if there is already a COS ID on which all features' values + * are same as the array. Then, we can reuse this COS ID. + */ + cos = find_cos(val_array, array_len, feat_type, info); + if ( cos == old_cos ) + { + ret = 0; + goto unlock_free_array; + } + + /* + * Step 3: + * If fail to find, we need pick an available COS ID. + * In fact, only COS ID which ref is 1 or 0 can be picked for current + * domain. If old_cos is not 0 and its ref==1, that means only current + * domain is using this old_cos ID. So, this old_cos ID certainly can + * be reused by current domain. Ref==0 means there is no any domain + * using this COS ID. So it can be used for current domain too. + */ + if ( cos < 0 ) + { + cos = pick_avail_cos(info, val_array, array_len, old_cos, feat_type); + if ( cos < 0 ) + { + ret = cos; + goto unlock_free_array; + } + + /* + * Step 4: + * Write the feature's MSRs according to the COS ID. + */ + ret = write_psr_msrs(socket, cos, val_array, array_len, feat_type); + if ( ret ) + goto unlock_free_array; + } + + /* + * Step 5: + * Find the COS ID (find_cos result is '>= 0' or an available COS ID is + * picked, then update ref according to COS ID. + */ + ref[cos]++; + ASSERT(!cos || ref[cos]); + ASSERT(!old_cos || ref[old_cos]); + ref[old_cos]--; + spin_unlock(&info->ref_lock); + + /* + * Step 6: + * Save the COS ID into current domain's psr_cos_ids[] so that we can know + * which COS the domain is using on the socket. One domain can only use + * one COS ID at same time on each socket. + */ + domain_lock(d); + d->arch.psr_cos_ids[socket] = cos; + domain_unlock(d); + + /* + * Step 7: + * Then, set the dom_set bit which corresponds to domain_id to mark this + * domain has been set and the COS ID of the domain is valid. + */ + set_bit(d->domain_id, info->dom_set); + + goto free_array; + + unlock_free_array: + spin_unlock(&info->ref_lock); + + free_array: + xfree(val_array); + return ret; +} + static void psr_free_cos(struct domain *d) { + unsigned int socket, cos; + + ASSERT(socket_info); + + if ( !d->arch.psr_cos_ids ) + return; + + /* Domain is destroyed so its cos_ref should be decreased. */ + for ( socket = 0; socket < nr_sockets; socket++ ) + { + struct psr_socket_info *info = socket_info + socket; + + clear_bit(d->domain_id, info->dom_set); + + /* cos 0 is default one which does not need be handled. */ + cos = d->arch.psr_cos_ids[socket]; + if ( cos == 0 ) + continue; + + spin_lock(&info->ref_lock); + ASSERT(info->cos_ref[cos]); + info->cos_ref[cos]--; + spin_unlock(&info->ref_lock); + } + xfree(d->arch.psr_cos_ids); d->arch.psr_cos_ids = NULL; } diff --git a/xen/include/asm-x86/psr.h b/xen/include/asm-x86/psr.h index ac9118e..15b9a25 100644 --- a/xen/include/asm-x86/psr.h +++ b/xen/include/asm-x86/psr.h @@ -73,8 +73,8 @@ int psr_get_info(unsigned int socket, enum cbm_type type, uint32_t data[], unsigned int array_len); int psr_get_val(struct domain *d, unsigned int socket, uint32_t *val, enum cbm_type type); -int psr_set_l3_cbm(struct domain *d, unsigned int socket, - uint64_t cbm, enum cbm_type type); +int psr_set_val(struct domain *d, unsigned int socket, + uint64_t val, enum cbm_type type); void psr_domain_init(struct domain *d); void psr_domain_free(struct domain *d);
As set value flow is the most complicated one in psr, it will be divided to some patches to make things clearer. This patch implements the set value framework to show a whole picture firstly. It also changes domctl interface to make it more general. To make the set value flow be general and can support multiple features at same time, it includes below steps: 1. Get COS ID that current domain is using. 2. Gather a value array to store all features current value into it and replace the current value of the feature which is being set to the new input value. 3. Find if there is already a COS ID on which all features' values are same as the array. Then, we can reuse this COS ID. 4. If fail to find, we need pick an available COS ID. Only COS ID which ref is 0 or 1 can be picked. 5. Write the feature's MSRs according to the COS ID. 6. Update ref according to COS ID. 7. Save the COS ID into current domain's psr_cos_ids[socket] so that we can know which COS the domain is using on the socket. 8. Set dom_ids bit corresponding to the domain so that we can know the domain has been set and the COS ID of the domain is valid. So, some functions are abstracted and the callback functions will be implemented in next patches. Here is an example to understand the process. The CPU supports two featuers, e.g. L3 CAT and L2 CAT. User wants to set L3 CAT of Dom1 to 0x1ff. 1. At the initial time, the old_cos of Dom1 is 0. The COS registers values are below at this time. ------------------------------- | COS 0 | COS 1 | COS 2 | ... | ------------------------------- L3 CAT | 0x7ff | 0x7ff | 0x7ff | ... | ------------------------------- L2 CAT | 0xff | 0xff | 0xff | ... | ------------------------------- 2. Gather the value array and insert new value into it: val[0]: 0x1ff val[1]: 0xff 3. It cannot find a matching COS. 4. Pick COS 1 to store the value set. 5. Write the L3 CAT COS 1 registers. The COS registers values are changed to below now. ------------------------------- | COS 0 | COS 1 | COS 2 | ... | ------------------------------- L3 CAT | 0x7ff | 0x1ff | ... | ... | ------------------------------- L2 CAT | 0xff | 0xff | ... | ... | ------------------------------- 6. The ref[1] is increased to 1 because Dom1 is using it now. 7. Save 1 to Dom1's psr_cos_ids[socket]. 8. Set the bit in 'dom_ids[]'. Then, user wants to set L3 CAT of Dom2 to 0x1ff too. The old_cos of Dom2 is 0 too. Repeat above flow. The val array assembled is: val[0]: 0x1ff val[1]: 0xff So, it can find a matching COS, COS 1. Then, it can reuse COS 1 for Dom2. The ref[1] is increased to 2 now because both Dom1 and Dom2 are using this COS ID. Set 1 to Dom2's psr_cos_ids[socket]. There is one thing need to emphasize that we need restore domain's COS ID to 0 when socket is offline. Otherwise, a wrong COS ID will be used when the socket is online again. That may cause user see the wrong CBM shown. But it takes much time to iterate all domains to restore COS ID to 0. So, we define a 'dom_ids[]' to represents all domains, one bit corresponds to one domain. If the bit is 0 when entering 'psr_ctxt_switch_to', that means this is the first time the domain is switched to this socket or domain's COS ID has not been set since the socket is online. So, the COS ID set to ASSOC register on this socket should be default value, 0. If not, that means the domain's COS ID has been set when the socket was online. So, this COS ID is valid and we can directly use it. We restore the domain's COS ID to 0 if the bit corresponding to the domain is 0 but the domain's COS ID is not 0 when 'psr_get_val' is called. This can avoid CPU serialization if restoring action is exectued in 'psr_ctxt_switch_to'. Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com> --- v12: - remove the memebers position changes in 'psr_socket_info'. (suggested by Jan Beulich) - rename 'dom_ids' to 'dom_set'. (suggested by Jan Beulich) - call 'bitmap_zero' to clear bitmap. (suggested by Jan Beulich) - combine two if()-s in 'psr_ctxt_switch_to' and add comment. (suggested by Jan Beulich) - remove redundant check in 'psr_get_val'. (suggested by Jan Beulich) - use 'domain_lock()' to protect 'psr_cos_ids' so that the codes do not depend on 'domctl_lock'. (suggested by Jan Beulich) - adjust codes to avoid cast in 'psr_set_val'. (suggested by Jan Beulich) - adjust codes to avoid duplication of error paths in 'psr_set_val'. (suggested by Jan Beulich) - clear the domain bit anyway in 'psr_free_cos'. (suggested by Jan Beulich) - use the default cos id when getting old_cos in 'psr_set_val'. v11: - define 'dom_ids[]' and implement related flows. - restore domain cos id to 0 in 'psr_get_val'. - rename 'write_psr_msr' to 'write_psr_msrs' and change its parameters to handle value array the feature's all MSRs. - fix coding style issue. (suggested by Jan Beulich) - do not need check 'cos' in ASSERT. (suggested by Jan Beulich) - rename 'insert_val_to_array' to 'insert_val_into_array'. (suggested by Jan Beulich) - remove 'ref_lock' from parameter list in 'find_cos' and 'pick_avail_cos'. (suggested by Jan Beulich) - remove ASSERT check to 'ref_lock' in 'find_cos' and 'pick_avail_cos'. (suggested by Jan Beulich) - fix a bug for checking 'feat_type'. (suggested by Jan Beulich) - move 'free_array' label. (suggested by Jan Beulich) - modify comments and commit message. v10: - restore domain cos id to 0 when socket is offline. (suggested by Jan Beulich) - check 'psr_cat_op.data' to make sure only lower 32 bits are valid. (suggested by Jan Beulich) - remove unnecessary fixed width type of parameters and variables. (suggested by Jan Beulich) - rename 'insert_new_val_to_array' to 'insert_val_to_array'. (suggested by Jan Beulich) - input 'ref_lock' pointer into functions to check if it has been locked. (suggested by Jan Beulich) - add comment to declare the set process is protected by 'domctl_lock'. (suggested by Jan Beulich) - check 'feat_type'. (suggested by Jan Beulich) - remove 'feat_mask'. (suggested by Jan Beulich) - remove unnecessary criteria of ASSERT. (suggested by Jan Beulich) - adjust flow of 'psr_set_val' to avoid 'goto' for successful cases. (suggested by Jan Beulich) - use ASSERT to check 'socket_info' in 'psr_free_cos'. (suggested by Jan Beulich) - remove unnecessary comment in 'psr_free_cos'. (suggested by Jan Beulich) v9: - use goto style error handling in 'psr_set_val'. (suggested by Wei Liu) - use ASSERT for checking old_cos. (suggested by Wei Liu and Jan Beulich) - fix coding style issue. (suggested by Wei Liu) - rename 'assemble_val_array' to 'combine_val_array' in pervious patch. (suggested by Wei Liu) - use 'spin_is_locked' to check ref_lock. (suggested by Roger Pau) - add an input parameter 'array_len' for 'write_psr_msr'. - check 'socket_info' and 'psr_cos_ids' in this patch. (suggested by Jan Beulich) - modify patch title to indicate 'L3 CAT'. (suggested by Jan Beulich) - fix commit message words. (suggested by Jan Beulich) - change 'assemble_val_array' to 'gather_val_array'. (suggested by Jan Beulich) - change 'set_new_val_to_array' to 'insert_new_val_to_array'. (suggested by Jan Beulich) - change parameter 'm' of 'insert_new_val_to_array' to 'new_val'. (suggested by Jan Beulich) - change 'write_psr_msr' to 'write_psr_msrs'. (suggested by Jan Beulich) - correct comments. (suggested by Jan Beulich) - remove unnecessary comments. (suggested by Jan Beulich) - adjust conditions after 'find_cos' to save a level of indentation. (suggested by Jan Beulich) - add 'ASSERT(!old_cos || ref[old_cos])'. (suggested by Jan Beulich) - move ASSERT() check into locked region. (suggested by Jan Beulich) - replace parameter '*val' to 'val[]' in some functions. (suggested by Jan Beulich) - change 'write_psr_msr' parameters to prepare to only set one new value for one feature. (suggested by Jan Beulich) - changes about 'uint64_t' to 'uint32_t'. (suggested by Jan Beulich) - add explanation about context switch. (suggested by Jan Beulich) v5: - modify commit message. (suggested by Jan Beulich) - return an error for all helper functions in set flow. (suggested by Jan Beulich) - remove unnecessary cast. (suggested by Jan Beulich) - divide 'get_old_set_new' to two functions, 'assemble_val_array' and 'set_new_val_to_array'. (suggested by Jan Beulich) - modify comments. (suggested by Jan Beulich) - adjust code format. (suggested by Jan Beulich) - change 'alloc_new_cos' to 'pick_avail_cos' to make name accurate. (suggested by Jan Beulich) - check feature type when entering 'psr_set_val'. (suggested by Jan Beulich) - use ASSERT to check ref. (suggested by Jan Beulich) - rename 'dat[]' to 'data[]'. (suggested by Jan Beulich) v4: - create this patch to make codes easier to understand. (suggested by Jan Beulich) --- --- xen/arch/x86/domctl.c | 18 ++-- xen/arch/x86/psr.c | 255 ++++++++++++++++++++++++++++++++++++++++++++-- xen/include/asm-x86/psr.h | 4 +- 3 files changed, 257 insertions(+), 20 deletions(-)