Skip to contentSkip to conversation

addition of loglevel for printf in Xen HV

PATCH RESEND In review Steven Rostedt 2 participants 6 messages Oct 20, 2006 – Oct 26, 2006

lore
Share & tools
lore
#0 lore raw

PATCH RESEND 1/1 addition of loglevel for printf in Xen HV

In the current implementation, the logging in the Hypervisor is basically at two levels. The developer can print "always" or "verbose".
The always print is when the developer uses "printk/printf" in the
hypervisor. The "verbose" is when the developer uses DPRINTK, since the DPRINTK macro is only defined when the "verbose" flag is set.

Now the problem we are having, is that when we send out a distribution, we want enough information to be printed if one of our customers has a problem. But we also don't want to have so much printed that it affects the performance of the system. Or worst yet, if the system has a serial console attached, one of the DomU's could in essence cause a DoS attack on the whole system by causing the DomU to do certain errors that causes a thousand prints to the serial console. These prints are done with interrupts disabled, so the system suffers large latencies for each print.

Since it would be helpful to have prints for different environments, I've introduced with this patch a Linux like loglevel from 0-7. When the user sets the loglevel to 0, only emergency prints are preformed, and when 7 is selected, all prints are done.

Here's the list of levels that I have added:

#define XEN_EMERG   "<0>"
#define XEN_ALERT   "<1>"
#define XEN_CRIT    "<2>"
#define XEN_ERR     "<3>"
#define XEN_WARNING "<4>"
#define XEN_GUEST   "<5>"
#define XEN_INFO    "<6>"
#define XEN_DEBUG   "<7>"

Basically, these correspond to their Linux equivalents. The one difference, is that the KERN_NOTICE was replaced by XEN_GUEST. The XEN_GUEST is the level that a guest can "at will" cause this print. So a distribution can set the print log level to 4 (XEN_WARNING) and all prints that are of errors and can be used for debugging are now available, without the prints that allow a Guest to cause a DoS.

I also added the keyhandlers '0' to '7' to allow the user to change the loglevel on a running system to the corresponding key. This is similar to the way Linux uses sysrq to change its log levels.

Two things that are not implemented with this patch but should be.

1. How to default the log level on build. Should the verbose now be a number, and not a config? i.e. VERBOSE=4

2. How the user in Linux can change the log level (from dom0). Should we add a hypercall or some other method to allow the user to change the logging?

Also, I made a pass though to assign all DPRINTK a log level. I may not have done this correctly. But a loglevel of 7 is now the same a VERBOSE=y and a loglevel of 0 is almost the same of a VERBOSE=n. There was a couple of places where XEN_EMERG was used.

This patch really does need to go in. Whether in its current form, or modified greatly. The ability to give the developer a level of printing is a great asset. That is why Linux has this ability.

-- Steve

xen/arch/ia64/vmx/vlsapic.c +2 -2
diff -r ffbd9ea05f75 xen/arch/ia64/vmx/vlsapic.c
@@ -389,7 +389,7 @@ void vlsapic_reset(VCPU *vcpu)
     vcpu->arch.arch_vmx.vlapic.vcpu = vcpu;
     hvm_vioapic_add_lapic(&vcpu->arch.arch_vmx.vlapic, vcpu);
 #endif
-    DPRINTK("VLSAPIC inservice base=%p\n", &VLSAPIC_INSVC(vcpu,0) );
+    DPRINTK(XEN_INFO "VLSAPIC inservice base=%p\n", &VLSAPIC_INSVC(vcpu,0) );
 }
 
 /*
@@ -518,7 +518,7 @@ int vmx_vcpu_pend_interrupt(VCPU *vcpu, 
     int ret;
 
     if (vector & ~0xff) {
-        DPRINTK("vmx_vcpu_pend_interrupt: bad vector\n");
+        DPRINTK(XEN_GUEST "vmx_vcpu_pend_interrupt: bad vector\n");
         return -1;
     }
     local_irq_save(spsr);
xen/arch/ia64/vmx/vmx_hypercall.c +1 -1
diff -r ffbd9ea05f75 xen/arch/ia64/vmx/vmx_hypercall.c
@@ -79,7 +79,7 @@ do_hvm_op(unsigned long op, XEN_GUEST_HA
     }
 
     default:
-        DPRINTK("Bad HVM op %ld.\n", op);
+        DPRINTK(XEN_GUEST "Bad HVM op %ld.\n", op);
         rc = -ENOSYS;
     }
     return rc;
xen/arch/ia64/vmx/vmx_init.c +1 -1
diff -r ffbd9ea05f75 xen/arch/ia64/vmx/vmx_init.c
@@ -277,7 +277,7 @@ static void vmx_create_event_channels(st
 			p = get_vio(v->domain, o->vcpu_id);
 			o->arch.arch_vmx.xen_port = p->vp_eport =
 			                alloc_unbound_xen_event_channel(o, 0);
-			DPRINTK("Allocated port %d for hvm.\n",
+			DPRINTK(XEN_GUEST "Allocated port %d for hvm.\n",
 			        o->arch.arch_vmx.xen_port);
 		}
 	}
xen/arch/ia64/xen/dom0_ops.c +2 -1
diff -r ffbd9ea05f75 xen/arch/ia64/xen/dom0_ops.c
@@ -252,7 +252,8 @@ do_dom0vp_op(unsigned long cmd,
     case IA64_DOM0VP_phystomach:
         ret = ____lookup_domain_mpa(d, arg0 << PAGE_SHIFT);
         if (ret == INVALID_MFN) {
-            DPRINTK("%s:%d INVALID_MFN ret: 0x%lx\n", __func__, __LINE__, ret);
+            DPRINTK(XEN_INFO "%s:%d INVALID_MFN ret: 0x%lx\n",
+                     __func__, __LINE__, ret);
         } else {
             ret = (ret & _PFN_MASK) >> PAGE_SHIFT;//XXX pte_pfn()
         }
xen/arch/ia64/xen/domain.c +1 -1
diff -r ffbd9ea05f75 xen/arch/ia64/xen/domain.c
@@ -625,7 +625,7 @@ int shadow_mode_control(struct domain *d
 	//struct vcpu *v;
 
 	if (unlikely(d == current->domain)) {
-		DPRINTK("Don't try to do a shadow op on yourself!\n");
+		DPRINTK(XEN_GUEST "Don't try to do a shadow op on yourself!\n");
 		return -EINVAL;
 	}   
 
xen/arch/ia64/xen/irq.c +8 -4
diff -r ffbd9ea05f75 xen/arch/ia64/xen/irq.c
@@ -377,7 +377,7 @@ int pirq_guest_bind(struct vcpu *v, int 
     {
         if ( desc->action != NULL )
         {
-            DPRINTK("Cannot bind IRQ %d to guest. In use by '%s'.\n",
+            DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. In use by '%s'.\n",
                     irq, desc->action->name);
             rc = -EBUSY;
             goto out;
@@ -386,7 +386,8 @@ int pirq_guest_bind(struct vcpu *v, int 
         action = xmalloc(irq_guest_action_t);
         if ( (desc->action = (struct irqaction *)action) == NULL )
         {
-            DPRINTK("Cannot bind IRQ %d to guest. Out of memory.\n", irq);
+            DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. Out of memory.\n",
+                    irq);
             rc = -ENOMEM;
             goto out;
         }
@@ -410,7 +411,8 @@ int pirq_guest_bind(struct vcpu *v, int 
     }
     else if ( !will_share || !action->shareable )
     {
-        DPRINTK("Cannot bind IRQ %d to guest. Will not share with others.\n",
+        DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. "
+                           "Will not share with others.\n",
                 irq);
         rc = -EBUSY;
         goto out;
@@ -418,7 +420,9 @@ int pirq_guest_bind(struct vcpu *v, int 
 
     if ( action->nr_guests == IRQ_MAX_GUESTS )
     {
-        DPRINTK("Cannot bind IRQ %d to guest. Already at max share.\n", irq);
+        DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. "
+                           "Already at max share.\n",
+                            irq);
         rc = -EBUSY;
         goto out;
     }
xen/arch/ia64/xen/mm.c +26 -19

156 lines not on this page · raw message

xen/arch/ia64/xen/vcpu.c +1 -1
diff -r ffbd9ea05f75 xen/arch/ia64/xen/vcpu.c
@@ -1442,7 +1442,7 @@ again:
 			set_metaphysical_rr0();
 		}
 		if (bundle->i64[0] == 0 && bundle->i64[1] == 0) {
-			DPRINTK("%s gip 0x%lx\n", __func__, gip);
+			DPRINTK(XEN_INFO "%s gip 0x%lx\n", __func__, gip);
 			return 0;
 		}
 		return 1;
xen/arch/powerpc/mm.c +1 -1
diff -r ffbd9ea05f75 xen/arch/powerpc/mm.c
@@ -268,7 +268,7 @@ int allocate_rma(struct domain *d, unsig
 
     d->arch.rma_page = alloc_domheap_pages(d, order, 0);
     if (d->arch.rma_page == NULL) {
-        DPRINTK("Could not allocate order=%d RMA for domain %u\n",
+        DPRINTK(XEN_GUEST "Could not allocate order=%d RMA for domain %u\n",
                 order, d->domain_id);
         return -ENOMEM;
     }
xen/arch/powerpc/shadow.c +2 -2
diff -r ffbd9ea05f75 xen/arch/powerpc/shadow.c
@@ -120,14 +120,14 @@ int shadow_domctl(struct domain *d,
 {
     if ( unlikely(d == current->domain) )
     {
-        DPRINTK("Don't try to do a shadow op on yourself!\n");
+        DPRINTK(XEN_GUEST "Don't try to do a shadow op on yourself!\n");
         return -EINVAL;
     }
 
     switch ( sc->op )
     {
     case XEN_DOMCTL_SHADOW_OP_OFF:
-         DPRINTK("Shadow is mandatory!\n");
+         DPRINTK(XEN_GUEST "Shadow is mandatory!\n");
          return -EINVAL;
 
     case XEN_DOMCTL_SHADOW_OP_GET_ALLOCATION:
xen/arch/x86/domain.c +2 -1
diff -r ffbd9ea05f75 xen/arch/x86/domain.c
@@ -556,7 +556,8 @@ static void load_segments(struct vcpu *n
              put_user(regs->r11,           rsp-10) |
              put_user(regs->rcx,           rsp-11) )
         {
-            DPRINTK("Error while creating failsafe callback frame.\n");
+            DPRINTK(XEN_GUEST "Error while creating failsafe "
+                    "callback frame.\n");
             domain_crash(n->domain);
         }
 
xen/arch/x86/extable.c +1 -1
diff -r ffbd9ea05f75 xen/arch/x86/extable.c
@@ -74,7 +74,7 @@ search_pre_exception_table(struct cpu_us
     unsigned long addr = (unsigned long)regs->eip;
     unsigned long fixup = search_one_table(
         __start___pre_ex_table, __stop___pre_ex_table-1, addr);
-    DPRINTK("Pre-exception: %p -> %p\n", _p(addr), _p(fixup));
+    DPRINTK(XEN_INFO "Pre-exception: %p -> %p\n", _p(addr), _p(fixup));
 #ifdef PERF_COUNTERS
     if ( fixup )
         perfc_incrc(exception_fixed);
xen/arch/x86/hvm/hvm.c +14 -11
diff -r ffbd9ea05f75 xen/arch/x86/hvm/hvm.c
@@ -192,7 +192,8 @@ void hvm_create_event_channels(struct vc
             p = get_vio(v->domain, o->vcpu_id);
             o->arch.hvm_vcpu.xen_port = p->vp_eport =
                 alloc_unbound_xen_event_channel(o, 0);
-            DPRINTK("Allocated port %d for hvm.\n", o->arch.hvm_vcpu.xen_port);
+            DPRINTK(XEN_INFO "Allocated port %d for hvm.\n",
+                    o->arch.hvm_vcpu.xen_port);
         }
     }
 }
@@ -343,7 +344,8 @@ static void hvm_vcpu_down(void)
     struct domain *d = v->domain;
     int online_count = 0;
 
-    DPRINTK("DOM%d/VCPU%d: going offline.\n", d->domain_id, v->vcpu_id);
+    DPRINTK(XEN_GUEST "DOM%d/VCPU%d: going offline.\n",
+           d->domain_id, v->vcpu_id);
 
     /* Doesn't halt us immediately, but we'll never return to guest context. */
     set_bit(_VCPUF_down, &v->vcpu_flags);
@@ -359,7 +361,8 @@ static void hvm_vcpu_down(void)
     /* ... Shut down the domain if not. */
     if ( online_count == 0 )
     {
-        DPRINTK("DOM%d: all CPUs offline -- powering off.\n", d->domain_id);
+        DPRINTK(XEN_GUEST "DOM%d: all CPUs offline -- powering off.\n",
+                d->domain_id);
         domain_shutdown(d, SHUTDOWN_poweroff);
     }
 }
@@ -499,7 +502,7 @@ void hvm_do_hypercall(struct cpu_user_re
 
     if ( (pregs->eax >= NR_hypercalls) || !hvm_hypercall_table[pregs->eax] )
     {
-        DPRINTK("HVM vcpu %d:%d did a bad hypercall %d.\n",
+        DPRINTK(XEN_GUEST "HVM vcpu %d:%d did a bad hypercall %d.\n",
                 current->domain->domain_id, current->vcpu_id,
                 pregs->eax);
         pregs->eax = -ENOSYS;
@@ -545,7 +548,7 @@ static long do_memory_op_compat32(int cm
     }
 
     default:
-        DPRINTK("memory_op %d.\n", cmd);
+        DPRINTK(XEN_GUEST "memory_op %d.\n", cmd);
         rc = -ENOSYS;
         break;
     }
@@ -578,7 +581,7 @@ void hvm_do_hypercall(struct cpu_user_re
     pregs->rax = (uint32_t)pregs->eax; /* mask in case compat32 caller */
     if ( (pregs->rax >= NR_hypercalls) || !hvm_hypercall64_table[pregs->rax] )
     {
-        DPRINTK("HVM vcpu %d:%d did a bad hypercall %ld.\n",
+        DPRINTK(XEN_GUEST "HVM vcpu %d:%d did a bad hypercall %ld.\n",
                 current->domain->domain_id, current->vcpu_id,
                 pregs->rax);
         pregs->rax = -ENOSYS;
@@ -629,7 +632,7 @@ int hvm_bringup_ap(int vcpuid, int tramp
 
     if ( bsp->vcpu_id != 0 )
     {
-        DPRINTK("Not calling hvm_bringup_ap from BSP context.\n");
+        DPRINTK(XEN_GUEST "Not calling hvm_bringup_ap from BSP context.\n");
         domain_crash_synchronous();
     }
 
@@ -638,7 +641,7 @@ int hvm_bringup_ap(int vcpuid, int tramp
 
     if ( (ctxt = xmalloc(struct vcpu_guest_context)) == NULL )
     {
-        DPRINTK("Failed to allocate memory in hvm_bringup_ap.\n");
+        DPRINTK(XEN_GUEST "Failed to allocate memory in hvm_bringup_ap.\n");
         return -ENOMEM;
     }
 
@@ -652,13 +655,13 @@ int hvm_bringup_ap(int vcpuid, int tramp
 
     if ( rc != 0 )
     {
-        DPRINTK("AP %d bringup failed in boot_vcpu %x.\n", vcpuid, rc);
+        DPRINTK(XEN_GUEST "AP %d bringup failed in boot_vcpu %x.\n", vcpuid, rc);
         return rc;
     }
 
     if ( test_and_clear_bit(_VCPUF_down, &d->vcpu[vcpuid]->vcpu_flags) )
         vcpu_wake(d->vcpu[vcpuid]);
-    DPRINTK("AP %d bringup suceeded.\n", vcpuid);
+    DPRINTK(XEN_INFO "AP %d bringup suceeded.\n", vcpuid);
 
     xfree(ctxt);
 
@@ -717,7 +720,7 @@ long do_hvm_op(unsigned long op, XEN_GUE
 
     default:
     {
-        DPRINTK("Bad HVM op %ld.\n", op);
+        DPRINTK(XEN_GUEST "Bad HVM op %ld.\n", op);
         rc = -ENOSYS;
         break;
     }
xen/arch/x86/hvm/instrlen.c +2 -2
diff -r ffbd9ea05f75 xen/arch/x86/hvm/instrlen.c
@@ -317,7 +317,7 @@ done_prefixes:
 
         if ( modrm_mod == 3 )
         {
-            DPRINTK("Cannot parse ModRM.mod == 3.\n");
+            DPRINTK(XEN_GUEST "Cannot parse ModRM.mod == 3.\n");
             goto cannot_emulate;
         }
 
@@ -444,7 +444,7 @@ done:
     return length;
 
 cannot_emulate:
-    DPRINTK("Cannot emulate %02x at address %lx (eip %lx, mode %d)\n",
+    DPRINTK(XEN_GUEST "Cannot emulate %02x at address %lx (eip %lx, mode %d)\n",
             b, (unsigned long)_regs.eip, (unsigned long)regs->eip, mode);
     return -1;
 }
xen/arch/x86/hvm/svm/svm.c +1 -1
diff -r ffbd9ea05f75 xen/arch/x86/hvm/svm/svm.c
@@ -806,7 +806,7 @@ static void svm_final_setup_guest(struct
 
     if ( !shadow_mode_external(d) )
     {
-        DPRINTK("Can't init HVM for dom %u vcpu %u: "
+        DPRINTK(XEN_WARNING "Can't init HVM for dom %u vcpu %u: "
                 "not in shadow external mode\n", d->domain_id, v->vcpu_id);
         domain_crash(d);
     }
xen/arch/x86/hvm/vmx/vmcs.c +1 -1
diff -r ffbd9ea05f75 xen/arch/x86/hvm/vmx/vmcs.c
@@ -145,7 +145,7 @@ static struct vmcs_struct *vmx_alloc_vmc
 
     if ( (vmcs = alloc_xenheap_page()) == NULL )
     {
-        DPRINTK("Failed to allocate VMCS.\n");
+        DPRINTK(XEN_GUEST "Failed to allocate VMCS.\n");
         return NULL;
     }
 
xen/arch/x86/hvm/vmx/vmx.c +4 -4
diff -r ffbd9ea05f75 xen/arch/x86/hvm/vmx/vmx.c
@@ -71,7 +71,7 @@ static int vmx_initialize_guest_resource
 
     if ( !shadow_mode_external(d) )
     {
-        DPRINTK("Can't init HVM for dom %u vcpu %u: "
+        DPRINTK(XEN_WARNING "Can't init HVM for dom %u vcpu %u: "
                 "not in shadow external mode\n", 
                 d->domain_id, v->vcpu_id);
         domain_crash(d);
@@ -83,7 +83,7 @@ static int vmx_initialize_guest_resource
 
         if ( (rc = vmx_create_vmcs(vc)) != 0 )
         {
-            DPRINTK("Failed to create VMCS for vcpu %d: err=%d.\n",
+            DPRINTK(XEN_WARNING "Failed to create VMCS for vcpu %d: err=%d.\n",
                     vc->vcpu_id, rc);
             return 0;
         }
@@ -92,14 +92,14 @@ static int vmx_initialize_guest_resource
 
         if ( (io_bitmap_a = alloc_xenheap_pages(IO_BITMAP_ORDER)) == NULL )
         {
-            DPRINTK("Failed to allocate io bitmap b for vcpu %d.\n",
+            DPRINTK(XEN_WARNING "Failed to allocate io bitmap b for vcpu %d.\n",
                     vc->vcpu_id);
             return 0;
         }
 
         if ( (io_bitmap_b = alloc_xenheap_pages(IO_BITMAP_ORDER)) == NULL )
         {
-            DPRINTK("Failed to allocate io bitmap b for vcpu %d.\n",
+            DPRINTK(XEN_WARNING "Failed to allocate io bitmap b for vcpu %d.\n",
                     vc->vcpu_id);
             return 0;
         }
xen/arch/x86/io_apic.c +2 -1
diff -r ffbd9ea05f75 xen/arch/x86/io_apic.c
@@ -1983,7 +1983,8 @@ int ioapic_guest_read(unsigned long phys
 }
 
 #define WARN_BOGUS_WRITE(f, a...)                                       \
-    DPRINTK("\n%s: apic=%d, pin=%d, old_irq=%d, new_irq=%d\n"           \
+    DPRINTK(XEN_INFO "\n%s: "                                           \
+            "apic=%d, pin=%d, old_irq=%d, new_irq=%d\n"                 \
             "%s: old_entry=%08x, new_entry=%08x\n"                      \
             "%s: " f, __FUNCTION__, apic, pin, old_irq, new_irq,        \
             __FUNCTION__, *(u32 *)&old_rte, *(u32 *)&new_rte,           \
xen/arch/x86/irq.c +7 -4
diff -r ffbd9ea05f75 xen/arch/x86/irq.c
@@ -432,7 +432,7 @@ int pirq_guest_bind(struct vcpu *v, int 
     {
         if ( desc->action != NULL )
         {
-            DPRINTK("Cannot bind IRQ %d to guest. In use by '%s'.\n",
+            DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. In use by '%s'.\n",
                     irq, desc->action->name);
             rc = -EBUSY;
             goto out;
@@ -441,7 +441,8 @@ int pirq_guest_bind(struct vcpu *v, int 
         action = xmalloc(irq_guest_action_t);
         if ( (desc->action = (struct irqaction *)action) == NULL )
         {
-            DPRINTK("Cannot bind IRQ %d to guest. Out of memory.\n", irq);
+            DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. Out of memory.\n",
+                    irq);
             rc = -ENOMEM;
             goto out;
         }
@@ -464,7 +465,8 @@ int pirq_guest_bind(struct vcpu *v, int 
     }
     else if ( !will_share || !action->shareable )
     {
-        DPRINTK("Cannot bind IRQ %d to guest. Will not share with others.\n",
+        DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. "
+               "Will not share with others.\n",
                 irq);
         rc = -EBUSY;
         goto out;
@@ -484,7 +486,8 @@ int pirq_guest_bind(struct vcpu *v, int 
 
     if ( action->nr_guests == IRQ_MAX_GUESTS )
     {
-        DPRINTK("Cannot bind IRQ %d to guest. Already at max share.\n", irq);
+        DPRINTK(XEN_GUEST "Cannot bind IRQ %d to guest. "
+               "Already at max share.\n", irq);
         rc = -EBUSY;
         goto out;
     }
xen/arch/x86/mm.c +2 -2
diff -r ffbd9ea05f75 xen/arch/x86/mm.c
@@ -2262,7 +2262,7 @@ int do_mmu_update(
             {
                 if ( shadow_mode_refcounts(d) )
                 {
-                    DPRINTK("mmu update on shadow-refcounted domain!");
+                    DPRINTK(XEN_INFO "mmu update on shadow-refcounted domain!");
                     break;
                 }
 
@@ -2626,7 +2626,7 @@ int steal_page(
         x = y;
         if (unlikely((x & (PGC_count_mask|PGC_allocated)) !=
                      (1 | PGC_allocated)) || unlikely(_nd != _d)) { 
-            DPRINTK("gnttab_transfer: Bad page %p: ed=%p(%u), sd=%p,"
+            DPRINTK(XEN_GUEST "gnttab_transfer: Bad page %p: ed=%p(%u), sd=%p,"
                     " caf=%08x, taf=%" PRtype_info "\n", 
                     (void *) page_to_mfn(page),
                     d, d->domain_id, unpickle_domptr(_nd), x, 
xen/arch/x86/mm/shadow/common.c +1 -1
diff -r ffbd9ea05f75 xen/arch/x86/mm/shadow/common.c
@@ -3041,7 +3041,7 @@ int shadow_domctl(struct domain *d,
 
     if ( unlikely(d == current->domain) )
     {
-        DPRINTK("Don't try to do a shadow op on yourself!\n");
+        DPRINTK(XEN_GUEST "Don't try to do a shadow op on yourself!\n");
         return -EINVAL;
     }
 
xen/arch/x86/platform_hypercall.c +1 -1
diff -r ffbd9ea05f75 xen/arch/x86/platform_hypercall.c
@@ -125,7 +125,7 @@ long do_platform_op(XEN_GUEST_HANDLE(xen
         case QUIRK_IOAPIC_GOOD_REGSEL:
 #ifndef sis_apic_bug
             sis_apic_bug = (quirk_id == QUIRK_IOAPIC_BAD_REGSEL);
-            DPRINTK("Domain 0 says that IO-APIC REGSEL is %s\n",
+            DPRINTK(XEN_INFO "Domain 0 says that IO-APIC REGSEL is %s\n",
                     sis_apic_bug ? "bad" : "good");
 #else
             BUG_ON(sis_apic_bug != (quirk_id == QUIRK_IOAPIC_BAD_REGSEL));
xen/arch/x86/traps.c +15 -13
diff -r ffbd9ea05f75 xen/arch/x86/traps.c
@@ -407,7 +407,8 @@ static inline int do_trap(int trapnr, ch
 
     if ( likely((fixup = search_exception_table(regs->eip)) != 0) )
     {
-        DPRINTK("Trap %d: %p -> %p\n", trapnr, _p(regs->eip), _p(fixup));
+        DPRINTK(XEN_EMERG "Trap %d: %p -> %p\n",
+                trapnr, _p(regs->eip), _p(fixup));
         regs->eip = fixup;
         return 0;
     }
@@ -475,7 +476,7 @@ int wrmsr_hypervisor_regs(
 
         if ( idx > 0 )
         {
-            DPRINTK("Dom%d: Out of range index %u to MSR %08x\n",
+            DPRINTK(XEN_GUEST "Dom%d: Out of range index %u to MSR %08x\n",
                     d->domain_id, idx, 0x40000000);
             return 0;
         }
@@ -485,7 +486,7 @@ int wrmsr_hypervisor_regs(
         if ( !mfn_valid(mfn) ||
              !get_page_and_type(mfn_to_page(mfn), d, PGT_writable_page) )
         {
-            DPRINTK("Dom%d: Bad GMFN %lx (MFN %lx) to MSR %08x\n",
+            DPRINTK(XEN_GUEST "Dom%d: Bad GMFN %lx (MFN %lx) to MSR %08x\n",
                     d->domain_id, gmfn, mfn, 0x40000000);
             return 0;
         }
@@ -832,17 +833,18 @@ static int __spurious_page_fault(
         return 0;
 
  spurious:
-    DPRINTK("Spurious fault in domain %u:%u at addr %lx, e/c %04x\n",
+    DPRINTK(XEN_WARNING "Spurious fault in domain %u:%u "
+            "at addr %lx, e/c %04x\n",
             current->domain->domain_id, current->vcpu_id,
             addr, regs->error_code);
 #if CONFIG_PAGING_LEVELS >= 4
-    DPRINTK(" l4e = %"PRIpte"\n", l4e_get_intpte(l4e));
+    DPRINTK(XEN_WARNING " l4e = %"PRIpte"\n", l4e_get_intpte(l4e));
 #endif
 #if CONFIG_PAGING_LEVELS >= 3
-    DPRINTK(" l3e = %"PRIpte"\n", l3e_get_intpte(l3e));
-#endif
-    DPRINTK(" l2e = %"PRIpte"\n", l2e_get_intpte(l2e));
-    DPRINTK(" l1e = %"PRIpte"\n", l1e_get_intpte(l1e));
+    DPRINTK(XEN_WARNING " l3e = %"PRIpte"\n", l3e_get_intpte(l3e));
+#endif
+    DPRINTK(XEN_WARNING " l2e = %"PRIpte"\n", l2e_get_intpte(l2e));
+    DPRINTK(XEN_WARNING " l1e = %"PRIpte"\n", l1e_get_intpte(l1e));
 #ifndef NDEBUG
     show_registers(regs);
 #endif
@@ -1313,7 +1315,7 @@ static int emulate_privileged_op(struct 
         case 0: /* Write CR0 */
             if ( (*reg ^ read_cr0()) & ~X86_CR0_TS )
             {
-                DPRINTK("Attempt to change unmodifiable CR0 flags.\n");
+                DPRINTK(XEN_GUEST "Attempt to change unmodifiable CR0 flags.\n");
                 goto fail;
             }
             (void)do_fpu_taskswitch(!!(*reg & X86_CR0_TS));
@@ -1333,7 +1335,7 @@ static int emulate_privileged_op(struct 
         case 4:
             if ( *reg != (read_cr4() & ~(X86_CR4_PGE|X86_CR4_PSE)) )
             {
-                DPRINTK("Attempt to change CR4 flags.\n");
+                DPRINTK(XEN_GUEST "Attempt to change CR4 flags.\n");
                 goto fail;
             }
             break;
@@ -1381,7 +1383,7 @@ static int emulate_privileged_op(struct 
 
             if ( (rdmsr_safe(regs->ecx, l, h) != 0) ||
                  (regs->eax != l) || (regs->edx != h) )
-                DPRINTK("Domain attempted WRMSR %p from "
+                DPRINTK(XEN_GUEST "Domain attempted WRMSR %p from "
                         "%08x:%08x to %08lx:%08lx.\n",
                         _p(regs->ecx), h, l, (long)regs->edx, (long)regs->eax);
             break;
@@ -1510,7 +1512,7 @@ asmlinkage int do_general_protection(str
 
     if ( likely((fixup = search_exception_table(regs->eip)) != 0) )
     {
-        DPRINTK("GPF (%04x): %p -> %p\n",
+        DPRINTK(XEN_WARNING "GPF (%04x): %p -> %p\n",
                 regs->error_code, _p(regs->eip), _p(fixup));
         regs->eip = fixup;
         return 0;
xen/arch/x86/x86_32/seg_fixup.c +24 -24

172 lines not on this page · raw message

xen/arch/x86/x86_64/traps.c +4 -2
diff -r ffbd9ea05f75 xen/arch/x86/x86_64/traps.c
@@ -206,7 +206,8 @@ unsigned long do_iret(void)
     if ( unlikely(copy_from_user(&iret_saved, (void *)regs->rsp,
                                  sizeof(iret_saved))) )
     {
-        DPRINTK("Fault while reading IRET context from guest stack\n");
+        DPRINTK(XEN_GUEST "Fault while reading IRET context from "
+                "guest stack\n");
         domain_crash_synchronous();
     }
 
@@ -215,7 +216,8 @@ unsigned long do_iret(void)
     {
         if ( unlikely(pagetable_is_null(v->arch.guest_table_user)) )
         {
-            DPRINTK("Guest switching to user mode with no user page tables\n");
+            DPRINTK(XEN_GUEST "Guest switching to user mode with no "
+                    "user page tables\n");
             domain_crash_synchronous();
         }
         toggle_guest_mode(v);
xen/common/event_channel.c +2 -1
diff -r ffbd9ea05f75 xen/common/event_channel.c
@@ -40,7 +40,8 @@
 
 #define ERROR_EXIT(_errno)                                          \
     do {                                                            \
-        DPRINTK("EVTCHNOP failure: domain %d, error %d, line %d\n", \
+        DPRINTK(XEN_GUEST                                           \
+                "EVTCHNOP failure: domain %d, error %d, line %d\n", \
                 current->domain->domain_id, (_errno), __LINE__);    \
         rc = (_errno);                                              \
         goto out;                                                   \
xen/common/grant_table.c +28 -21

174 lines not on this page · raw message

xen/common/keyhandler.c +15 -0
diff -r ffbd9ea05f75 xen/common/keyhandler.c
@@ -247,6 +247,13 @@ extern void perfc_reset(unsigned char ke
 extern void perfc_reset(unsigned char key);
 #endif
 
+static void do_set_log(unsigned char key,  struct cpu_user_regs *regs)
+{
+    print_log_level = key - '0';
+    /* always show this :-) */
+    printk(XEN_EMERG "loglevel set to %d\n", print_log_level);
+}
+
 static void do_debug_key(unsigned char key, struct cpu_user_regs *regs)
 {
     (void)debugger_trap_fatal(0xf001, regs);
@@ -257,6 +264,7 @@ static void do_debug_key(unsigned char k
 
 void initialize_keytable(void)
 {
+    int i;
     open_softirq(KEYPRESS_SOFTIRQ, keypress_softirq);
 
     register_irq_keyhandler(
@@ -279,6 +287,13 @@ void initialize_keytable(void)
     register_keyhandler(
         'P', perfc_reset,    "reset performance counters");
 #endif
+
+    for ( i = 0 ; i <= LOG_MAX ; i++ )
+    {
+        char buf[STR_MAX];
+        sprintf(buf, "set log level to %d", i);
+        register_irq_keyhandler('0'+i, do_set_log, buf);
+    }
 
     register_irq_keyhandler('%', do_debug_key,   "Trap to xendbg");
 }
xen/common/memory.c +5 -5
diff -r ffbd9ea05f75 xen/common/memory.c
@@ -61,7 +61,7 @@ increase_reservation(
         if ( unlikely((page = alloc_domheap_pages(
             d, extent_order, memflags)) == NULL) )
         {
-            DPRINTK("Could not allocate order=%d extent: "
+            DPRINTK(XEN_GUEST "Could not allocate order=%d extent: "
                     "id=%d memflags=%x (%ld of %d)\n",
                     extent_order, d->domain_id, memflags, i, nr_extents);
             return i;
@@ -114,7 +114,7 @@ populate_physmap(
         if ( unlikely((page = alloc_domheap_pages(
             d, extent_order, memflags)) == NULL) )
         {
-            DPRINTK("Could not allocate order=%d extent: "
+            DPRINTK(XEN_GUEST "Could not allocate order=%d extent: "
                     "id=%d memflags=%x (%ld of %d)\n",
                     extent_order, d->domain_id, memflags, i, nr_extents);
             goto out;
@@ -153,7 +153,7 @@ guest_remove_page(
     mfn = gmfn_to_mfn(d, gmfn);
     if ( unlikely(!mfn_valid(mfn)) )
     {
-        DPRINTK("Domain %u page number %lx invalid\n",
+        DPRINTK(XEN_GUEST "Domain %u page number %lx invalid\n",
                 d->domain_id, gmfn);
         return 0;
     }
@@ -161,7 +161,7 @@ guest_remove_page(
     page = mfn_to_page(mfn);
     if ( unlikely(!get_page(page, d)) )
     {
-        DPRINTK("Bad page free for domain %u\n", d->domain_id);
+        DPRINTK(XEN_GUEST "Bad page free for domain %u\n", d->domain_id);
         return 0;
     }
 
@@ -174,7 +174,7 @@ guest_remove_page(
     if ( unlikely(!page_is_removable(page)) )
     {
         /* We'll make this a guest-visible error in future, so take heed! */
-        DPRINTK("Dom%d freeing in-use page %lx (pseudophys %lx):"
+        DPRINTK(XEN_GUEST "Dom%d freeing in-use page %lx (pseudophys %lx):"
                 " count=%lx type=%lx\n",
                 d->domain_id, mfn, get_gpfn_from_mfn(mfn),
                 (unsigned long)page->count_info, page->u.inuse.type_info);
xen/common/multicall.c +1 -1
diff -r ffbd9ea05f75 xen/common/multicall.c
@@ -24,7 +24,7 @@ do_multicall(
 
     if ( unlikely(__test_and_set_bit(_MCSF_in_multicall, &mcs->flags)) )
     {
-        DPRINTK("Multicall reentry is disallowed.\n");
+        DPRINTK(XEN_GUEST "Multicall reentry is disallowed.\n");
         return -EINVAL;
     }
 
xen/common/page_alloc.c +3 -2
diff -r ffbd9ea05f75 xen/common/page_alloc.c
@@ -542,7 +542,8 @@ int assign_pages(
 
     if ( unlikely(test_bit(_DOMF_dying, &d->domain_flags)) )
     {
-        DPRINTK("Cannot assign page to domain%d -- dying.\n", d->domain_id);
+        DPRINTK(XEN_GUEST "Cannot assign page to domain%d -- dying.\n",
+                d->domain_id);
         goto fail;
     }
 
@@ -550,7 +551,7 @@ int assign_pages(
     {
         if ( unlikely((d->tot_pages + (1 << order)) > d->max_pages) )
         {
-            DPRINTK("Over-allocation for domain %u: %u > %u\n",
+            DPRINTK(XEN_GUEST "Over-allocation for domain %u: %u > %u\n",
                     d->domain_id, d->tot_pages + (1 << order), d->max_pages);
             goto fail;
         }
xen/common/schedule.c +1 -1
diff -r ffbd9ea05f75 xen/common/schedule.c
@@ -468,7 +468,7 @@ long do_set_timer_op(s_time_t timeout)
          * timeout in this case can burn a lot of CPU. We therefore go for a
          * reasonable middleground of triggering a timer event in 100ms.
          */
-        DPRINTK("Warning: huge timeout set by domain %d (vcpu %d):"
+        DPRINTK(XEN_GUEST "Warning: huge timeout set by domain %d (vcpu %d):"
                 " %"PRIx64"\n",
                 v->domain->domain_id, v->vcpu_id, (uint64_t)timeout);
         set_timer(&v->timer, NOW() + MILLISECS(100));
xen/common/trace.c +1 -1
diff -r ffbd9ea05f75 xen/common/trace.c
@@ -131,7 +131,7 @@ static int tb_set_size(int size)
      */
     if ( (opt_tbuf_size != 0) || (size <= 0) )
     {
-        DPRINTK("tb_set_size from %d to %d not implemented\n",
+        DPRINTK(XEN_GUEST "tb_set_size from %d to %d not implemented\n",
                 opt_tbuf_size, size);
         return -EINVAL;
     }
xen/drivers/char/console.c +17 -0
diff -r ffbd9ea05f75 xen/drivers/char/console.c
@@ -298,6 +298,8 @@ static inline void __putstr(const char *
     }
 }
 
+int print_log_level = PRINT_LOG_DEFAULT;
+
 void printf(const char *fmt, ...)
 {
     static char   buf[1024];
@@ -306,6 +308,7 @@ void printf(const char *fmt, ...)
     va_list       args;
     char         *p, *q;
     unsigned long flags;
+    int           level = PRINT_LOG_DEFAULT;
 
     spin_lock_irqsave(&console_lock, flags);
 
@@ -314,6 +317,19 @@ void printf(const char *fmt, ...)
     va_end(args);        
 
     p = buf;
+    if ( (p[0] == '<') && (p[1] >= '0')
+         && (p[1] <= ('0' + LOG_MAX))
+         && (p[2] == '>') )
+    {
+        level = p[1] - '0';
+        p += 3;
+    }
+
+    if ( level > print_log_level )
+    {
+        goto out;
+    }
+
     while ( (q = strchr(p, '\n')) != NULL )
     {
         *q = '\0';
@@ -333,6 +349,7 @@ void printf(const char *fmt, ...)
         start_of_line = 0;
     }
 
+ out:
     spin_unlock_irqrestore(&console_lock, flags);
 }
 
xen/include/asm-ia64/mm.h +1 -1
diff -r ffbd9ea05f75 xen/include/asm-ia64/mm.h
@@ -170,7 +170,7 @@ static inline int get_page(struct page_i
 	    unlikely((nx & PGC_count_mask) == 0) ||	/* Count overflow? */
 	    unlikely((x >> 32) != _domain)) {		/* Wrong owner? */
 
-	    DPRINTK("Error pfn %lx: rd=%p, od=%p, caf=%016lx, taf=%"
+	    DPRINTK(XEN_GUEST "Error pfn %lx: rd=%p, od=%p, caf=%016lx, taf=%"
 		PRtype_info "\n", page_to_mfn(page), domain,
 		unpickle_domptr(x >> 32), x, page->u.inuse.type_info);
 	    return 0;
xen/include/asm-x86/mm.h +2 -1
diff -r ffbd9ea05f75 xen/include/asm-x86/mm.h
@@ -214,7 +214,8 @@ static inline int get_page(struct page_i
              unlikely(d != _domain) )                /* Wrong owner? */
         {
             if ( !_shadow_mode_refcounts(domain) )
-                DPRINTK("Error pfn %lx: rd=%p, od=%p, caf=%08x, taf=%" 
+                DPRINTK(XEN_GUEST
+                        "Error pfn %lx: rd=%p, od=%p, caf=%08x, taf=%"
                         PRtype_info "\n",
                         page_to_mfn(page), domain, unpickle_domptr(d),
                         x, page->u.inuse.type_info);
xen/include/xen/config.h +21 -12
diff -r ffbd9ea05f75 xen/include/xen/config.h
@@ -13,25 +13,34 @@
 #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
 
 /* Linux syslog levels. */
-#define KERN_NOTICE  ""
-#define KERN_WARNING ""
-#define KERN_DEBUG   ""
-#define KERN_INFO    ""
-#define KERN_ERR     ""
-#define KERN_CRIT    ""
-#define KERN_EMERG   ""
-#define KERN_ALERT   ""
+#define XEN_EMERG   "<0>"
+#define XEN_ALERT   "<1>"
+#define XEN_CRIT    "<2>"
+#define XEN_ERR     "<3>"
+#define XEN_WARNING "<4>"
+#define XEN_GUEST   "<5>"
+#define XEN_INFO    "<6>"
+#define XEN_DEBUG   "<7>"
+
+#define KERN_EMERG   XEN_EMERG
+#define KERN_ALERT   XEN_ALERT
+#define KERN_CRIT    XEN_CRIT
+#define KERN_ERR     XEN_ERR
+#define KERN_WARNING XEN_WARNING
+#define KERN_NOTICE  XEN_GUEST
+#define KERN_INFO    XEN_INFO
+#define KERN_DEBUG   XEN_DEBUG
+
+#define LOG_MAX 7
+
+#define PRINT_LOG_DEFAULT 5
 
 /* Linux 'checker' project. */
 #define __iomem
 #define __user
 
-#ifdef VERBOSE
 #define DPRINTK(_f, _a...) printk("(file=%s, line=%d) " _f, \
                            __FILE__ , __LINE__ , ## _a )
-#else
-#define DPRINTK(_f, _a...) ((void)0)
-#endif
 
 #ifndef __ASSEMBLY__
 #include <xen/compiler.h>
xen/include/xen/lib.h +1 -0
diff -r ffbd9ea05f75 xen/include/xen/lib.h
@@ -60,6 +60,7 @@ extern void panic(const char *format, ..
 extern void panic(const char *format, ...)
     __attribute__ ((format (printf, 1, 2)));
 extern long vm_assist(struct domain *, unsigned int, unsigned int);
+extern int print_log_level;
 
 /* vsprintf.c */
 extern int sprintf(char * buf, const char * fmt, ...)

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#1 lore raw
Quoted 3 lines · Steven Rostedt

This patch really does need to go in. Whether in its current form, or modified greatly. The ability to give the developer a level of printing is a great asset. That is why Linux has this ability.

I agree. Here are some comments:

* We don't need 8 log levels. Noone knows the difference between
EMERG/CRIT/ALERT, or at least I really doubt that they can be used consistently in a large code base. I suggest:

    XENLOG_ERR    -- Bad errors, likely fatal (to a guest or the host)
    XENLOG_WARN   -- Weird stuff happening, recoverable (maybe)
    XENLOG_INFO   -- Interesting info, not too noisy
    XENLOG_DEBUG  -- Noisy as you like

* I think your ratelimit stuff could be usefully integrated. Instead of
having a single threshold, have two: one above which everything is printed, one below which nothing is printed. In between is rate limited.

* I think the guest/not-guest printing is orthogonal to the log-level
scale. Perhaps we should have another <> specifier (<G>?) that can be used alongside a log level specifier. This would give another two threshold values (two for guest, two for non-guest) which would maybe be overkill but would be reasonable if we gave a dom0 tool to control it. Actually this could be extended to other subsystems (shadow code for example). Then we would have a control per subsystem falling back to default thresholds if not specified. That probably really is overkill!

* Likely defaults:
Non-guest prints ERR/WARN always, no INFO/DEBUG
Guest prints ERR/WARN rate-limited, no INFO/DEBUG
(Like Linux, we'd probably have slacker default controls until initial
bootstrap has completed, so you get useful info out).

What do you think?

-- Keir

#2 lore raw
Quoted 9 lines · Keir Fraser
Quoted 3 lines · Steven Rostedt

This patch really does need to go in. Whether in its current form, or modified greatly. The ability to give the developer a level of printing is a great asset. That is why Linux has this ability.

I agree. Here are some comments:

Keir, Thanks for the comments, they are very useful.

Quoted 8 lines

* We don't need 8 log levels. Noone knows the difference between
EMERG/CRIT/ALERT, or at least I really doubt that they can be used consistently in a large code base. I suggest:

    XENLOG_ERR    -- Bad errors, likely fatal (to a guest or the host)
    XENLOG_WARN   -- Weird stuff happening, recoverable (maybe)
    XENLOG_INFO   -- Interesting info, not too noisy
    XENLOG_DEBUG  -- Noisy as you like

Perfectly agree. Four is fine, especially if we implement a <G> option for guest.

Quoted 4 lines

* I think your ratelimit stuff could be usefully integrated. Instead of
having a single threshold, have two: one above which everything is printed, one below which nothing is printed. In between is rate limited.

I'm not quite sure I understand this fully.

So what you are saying is instead of having the print_log_level have a printk_upper_log_level and a printk_lower_log_level where, as you say, everything below the printk_lower_log_level is printed, and the printk_upper_log_level where everyting above is not printed, and thus everything in between is rate limited? (remember 0 is ERR and 3 is DEBUG)

Quoted 10 lines

* I think the guest/not-guest printing is orthogonal to the log-level
scale. Perhaps we should have another <> specifier (<G>?) that can be used alongside a log level specifier. This would give another two threshold values (two for guest, two for non-guest) which would maybe be overkill but would be reasonable if we gave a dom0 tool to control it. Actually this could be extended to other subsystems (shadow code for example). Then we would have a control per subsystem falling back to default thresholds if not specified. That probably really is overkill!

Quoted 8 lines

* Likely defaults:
Non-guest prints ERR/WARN always, no INFO/DEBUG
Guest prints ERR/WARN rate-limited, no INFO/DEBUG
(Like Linux, we'd probably have slacker default controls until initial
bootstrap has completed, so you get useful info out).

What do you think?

I have to admit I like the ideas you have thrown to me.

So let me put this is my own words/code so that you can see what my view of this is.

Have 4 thresholds.

printk_dom0_upper_threshold
printk_dom0_lower_threshold
printk_guest_upper_threshold
printk_guest_lower_threshold

Have 4 log levels:

XENLOG_ERR
XENLOG_WARN
XENLOG_INFO
XENLOG_DEBUG

Add a XENLOG_GUEST (defined as "<G>")

And for less typing add

XENLOG_G_ERR
XENLOG_G_WARN
XENLOG_G_INFO
XENLOG_G_DEBUG

which would just append the XENLOG_GUEST in front of the associated warning.

example;

   printk(XENLOG_G_ERR "fatal error in guest\n");

   printk(XENLOG_INFO "dom0 is running happily\n");

Then in printk itself, we would have

  upper_thresh = printk_dom0_upper_threshold;
  lower_thresh = printk_dom0_lower_threshold;
  level = some_default;
  if (strncmp("<G>",fmt,3)==0) {
	upper_thresh = printk_guest_upper_threshold;
	lower_thresh = printk_guest_lower_threshold;
	level = some_guest_default;
	fmt += 3;

}
if (strncmp("<[0-3]>",fmt,3)==0) { /* using regexpr and not C to make
this readable */
level =~ s/<(0-3)>/$1/; /* perl format :-) */
}

if (level > upper_thresh)
return; /* do nothing */

if (level >= lower_thresh)

	if (rate_limit())
		return;

print as normal.

So anything in between the upper and lower thresholds inclusive would be rate limited. Also I'm not sure of the best way dom0 can communicate threshold changes to the hypervisor. As well as if there should be a config to set the defaults.

So what are your thoughts?

-- Steve

#3 lore raw
Quoted 4 lines · Steven Rostedt

So anything in between the upper and lower thresholds inclusive would be rate limited. Also I'm not sure of the best way dom0 can communicate threshold changes to the hypervisor. As well as if there should be a config to set the defaults.

I think yet-another-little-tool in dom0 is the right way to go. We can add a sysctl to get/set the parameters. It can pretty-print the current parameters and also provide a way of changing them. I don't think there's any benefit to integration with xm -- it's not like this needs to be integrated with xend or anything like that.

Yes, we should also have a way to set them as a boot parameter. I suggest we should only enact those explicit parameters when dom0 starts to boot though (no need to stop Xen being noisy in initial boot). We could have an extra parameter (e.g., 'quiet') to override that if someone really cares.

It'd be nice to have debug-key support for changing it too, but actually that's probably not so important if we have the dom0 tool.

Everything you said I agree with, so please do go hack it up. :-)

Another thought... By pushing the rate-limiting into printk() itself, we lose the ability to print all-or-nothing for related blocks of printk (e.g., register dumps). We might want a 'same as previous line' specifier, which would print/not-print same as previous output line?

-- Keir

Quoted 1 line

So what are your thoughts?

#4 lore raw
Quoted 6 lines · Keir Fraser

I think yet-another-little-tool in dom0 is the right way to go. We can add a sysctl to get/set the parameters. It can pretty-print the current parameters and also provide a way of changing them. I don't think there's any benefit to integration with xm -- it's not like this needs to be integrated with xend or anything like that.

That should be easy enough. What's the best way to communicate between dom0 and the HV? Add a hypercall for this?

Quoted 5 lines

Yes, we should also have a way to set them as a boot parameter. I suggest we should only enact those explicit parameters when dom0 starts to boot though (no need to stop Xen being noisy in initial boot). We could have an extra parameter (e.g., 'quiet') to override that if someone really cares.

Yeah, but this can be handled after the initial work is done.

Quoted 3 lines

It'd be nice to have debug-key support for changing it too, but actually that's probably not so important if we have the dom0 tool.

My original patch included a debug key for log level. I'm sure it won't be too hard to add something to change the thresholds too. (for those that only communicate to the machine via a serial console).

Quoted 2 lines

Everything you said I agree with, so please do go hack it up. :-)

Thanks! Will start hacking!

Quoted 6 lines

Another thought... By pushing the rate-limiting into printk() itself, we lose the ability to print all-or-nothing for related blocks of printk (e.g., register dumps). We might want a 'same as previous line' specifier, which would print/not-print same as previous output line?

Hmm, I wonder if we should have a XENLOG_CRASH level which is above XENLOG_ERR. This will be when the system is actually taking a dump, and we want to show everything. Or just simply force the thresholds to max just before showing the crash print outputs. Or add a variable to know we are in the process of crashing, and have printk ignore the print rate limit if that's the case, similar to the oops_in_progress of Linux.

-- Steve

#5 lore raw
Quoted 6 lines · Steven Rostedt
Quoted 5 lines

Another thought... By pushing the rate-limiting into printk() itself, we lose the ability to print all-or-nothing for related blocks of printk (e.g., register dumps). We might want a 'same as previous line' specifier, which would print/not-print same as previous output line?

Quoted 6 lines

Hmm, I wonder if we should have a XENLOG_CRASH level which is above XENLOG_ERR. This will be when the system is actually taking a dump, and we want to show everything. Or just simply force the thresholds to max just before showing the crash print outputs. Or add a variable to know we are in the process of crashing, and have printk ignore the print rate limit if that's the case, similar to the oops_in_progress of Linux.

We dump registers for guest crashes too, and it'd be nice to have those disable-able. I think going for a first cut without worrying about this is fine. We can finesse things later. It'll be a mega-big patch as it is!

-- Keir