vmmon: use build time version check in MapCrossPage()

Previous commit breaks build against kernels older than 5.3 as there used to be one more argument of handler called by apply_to_page_range(). Use preprocessor version check so that code intended for kernel 5.8 and later is not compiled at all when building against older kernels. Also compile crosspage_set_exec() conditionally to avoid compiler warning.
vmmon: avoid using alloc_vm_area() for crosspage allocation
2020-12-09 10:42:10 +01:00 · 2020-12-09 09:17:59 +01:00 · 2020-12-09 09:17:59 +01:00 · 2020-12-09 09:17:59 +01:00 · 2020-12-09 09:17:59 +01:00 · 2020-12-09 09:17:59 +01:00
14 changed files with 203 additions and 69 deletions
@@ -136,9 +136,6 @@ ifdef VMX86_DEBUG
 CC_OPTS += -DVMX86_DEBUG
 endif

-# Add Spectre options when available
-CC_OPTS += $(call vm_check_gcc,-mindirect-branch=thunk -mindirect-branch-register,)
-
 include $(SRCROOT)/Makefile.kernel

 else
@@ -31,7 +31,7 @@ $(DRIVER)-y := $(subst $(SRCROOT)/, , $(patsubst %.c, %.o, \
 		$(SRCROOT)/bootstrap/*.c)))

 clean:
-	rm -rf $(wildcard $(DRIVER).mod.c $(DRIVER).ko .tmp_versions \
+	rm -rf $(wildcard $(DRIVER).mod.c $(DRIVER).ko .tmp_versions .cache.mk \
 		Module.symvers Modules.symvers Module.markers modules.order \
 		$(foreach dir,linux/ common/ vmcore/ bootstrap/ \
 		./,$(addprefix $(dir),.*.cmd .*.o.flags *.o)))
@@ -2307,12 +2307,23 @@ TaskSwitchToMonitor(VMCrossPage *crosspage)
   {
      uint64 raxGetsWiped, rcxGetsWiped;

+#ifdef CALL_NOSPEC
+      __asm__ __volatile__(CALL_NOSPEC
+                           : "=a" (raxGetsWiped),
+                             "=c" (rcxGetsWiped)
+                           : "0" (codePtr),
+                             "1" (crosspage),
+                             THUNK_TARGET(codePtr)
+                           : "rdx", "r8", "r9", "r10", "r11", "cc", "memory");
+#else
      __asm__ __volatile__("call *%%rax"
                           : "=a" (raxGetsWiped),
                             "=c" (rcxGetsWiped)
                           : "0" (codePtr),
                             "1" (crosspage)
                           : "rdx", "r8", "r9", "r10", "r11", "cc", "memory");
+#endif
+
   }
 #elif defined(_MSC_VER)
   /*
@@ -0,0 +1,30 @@
+#ifndef __COMPAT_POLL_H__
+#define __COMPAT_POLL_H__
+
+#include <linux/poll.h>
+
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(4, 18, 0))
+
+#ifndef __poll_t
+typedef unsigned int __poll_t;
+#endif
+
+static inline __poll_t compat_vfs_poll(struct file *file,
+				       struct poll_table_struct *pt)
+{
+	if (unlikely(!file->f_op->poll))
+		return DEFAULT_POLLMASK;
+	return file->f_op->poll(file, pt);
+}
+
+#else
+
+static inline __poll_t compat_vfs_poll(struct file *file,
+				       struct poll_table_struct *pt)
+{
+	return vfs_poll(file, pt);
+}
+
+#endif
+
+#endif /* __COMPAT_POLL_H__ */
@@ -67,6 +67,7 @@ extern "C" {
 #if defined (VMKPANIC) 
 #include "vmk_assert.h"
 #else /* !VMKPANIC */
+#include <linux/kernel.h>
 #define _ASSERT_PANIC(name) \
           Panic(_##name##Fmt "\n", __FILE__, __LINE__)
 #define _ASSERT_PANIC_BUG(bug, name) \
@@ -107,7 +108,7 @@ NORETURN void Panic_NoSave(const char *fmt, ...) PRINTF_DECL(1, 2);
 } while(0)

 #else
-NORETURN void Panic(const char *fmt, ...) PRINTF_DECL(1, 2);
+#define Panic panic
 #endif

 void LogThrottled(uint32 *count, const char *fmt, ...) PRINTF_DECL(2, 3);
@@ -35,6 +35,8 @@
 #define INCLUDE_ALLOW_VMCORE
 #include "includeCheck.h"

+#include <asm/processor-flags.h>
+
 #define X86_MAX_INSTR_LEN  15   /* Max byte length of an x86 instruction. */

 #define NUM_IDT_VECTORS 256
@@ -75,7 +77,9 @@
 #define CR3_PDB_MASK   0xfffff000
 #define CR3_IGNORE     0xFFF
 #define PAE_CR3_IGNORE 0x1F
+#ifndef CR3_PCID_MASK
 #define CR3_PCID_MASK  0xFFF
+#endif
 #define CR3_NO_FLUSH   (1ULL << 63)

 #define CR4_VME        0x00000001
@@ -24,6 +24,7 @@

 #ifndef _X86MSR_H_
 #define _X86MSR_H_
+#include <asm/msr-index.h>
 #define INCLUDE_ALLOW_USERLEVEL
 #define INCLUDE_ALLOW_VMX

@@ -137,7 +138,9 @@ MSRQuery;
 #define MSR_TSX_CTRL_RTM_DISABLE                  (1ULL << 0)
 #define MSR_TSX_CTRL_CPUID_CLEAR                  (1ULL << 1)

+#ifndef MSR_MISC_FEATURES_ENABLES
 #define MSR_MISC_FEATURES_ENABLES            0x140
+#endif

 /* Intel Core Architecture and later: use only architected counters. */
 #define IA32_MSR_PERF_CAPABILITIES                   0x345
@@ -472,7 +475,9 @@ MSRQuery;
 #define MSR_K7_HWCR_SSEDIS         0x00008000ULL // Disable SSE bit
 #define MSR_K7_HWCR_MONMWAITUSEREN 0x00000400ULL // Enable MONITOR/MWAIT CPL>0
 #define MSR_K7_HWCR_TLBFFDIS       0x00000040ULL // Disable TLB Flush Filter
+#ifndef MSR_K7_HWCR_SMMLOCK
 #define MSR_K7_HWCR_SMMLOCK        0x00000001ULL // Lock SMM environment
+#endif

 #ifndef MSR_K8_SYSCFG
 #define MSR_K8_SYSCFG        0xc0010010
@@ -633,7 +638,9 @@ MSRQuery;
 #define MSR_HYPERV_GUESTOSID_OS_WINNT_DERIVATIVE 4ULL

 /* MSR for forcing RTM abort to recover PMC3 (see PR 2333817) */
+#ifndef MSR_TSX_FORCE_ABORT
 #define MSR_TSX_FORCE_ABORT                      0x0000010f
+#endif
 #define MSR_TSX_FORCE_ABORT_RTM_BIT_INDEX        0

 /*
@@ -687,7 +694,11 @@ typedef unsigned char MTRRType;
 /*
 * MISC_FEATURES_ENABLES bits
 */
+#ifdef MSR_MISC_FEATURES_ENABLES_CPUID_FAULT
+#define MSR_MISC_FEATURES_ENABLES_CPUID_FAULTING MSR_MISC_FEATURES_ENABLES_CPUID_FAULT
+#else
 #define MSR_MISC_FEATURES_ENABLES_CPUID_FAULTING 1
+#endif



@@ -664,7 +664,7 @@ LinuxDriverReadTSC(void *data,   // OUT: TSC values
 *-----------------------------------------------------------------------------
 */

-__attribute__((always_inline)) static Bool
+__always_inline static Bool
 LinuxDriverSyncReadTSCs(uint64 *delta) // OUT: TSC max - TSC min
 {
   TSCDelta tscDelta;
@@ -1227,7 +1227,7 @@ LinuxDriver_Ioctl(struct file *filp,    // IN:
      }

   case IOCTL_VMX86_WRITE_PAGE: {
-#if VMX86_DEVEL
+#ifdef VMX86_DEVEL
         VMMReadWritePage req;

         retval = HostIF_CopyFromUser(&req, ioarg, sizeof req);
@@ -47,13 +47,16 @@
 #include <asm/asm.h>
 #include <asm/io.h>
 #include <asm/page.h>
+#include <asm/tlbflush.h>
 #include <asm/uaccess.h>
+#include <asm/irq_vectors.h>
 #include <linux/capability.h>
 #include <linux/kthread.h>
 #include <linux/wait.h>
 #include <linux/hrtimer.h>
 #include <linux/signal.h>
 #include <linux/taskstats_kern.h> // For linux/sched/signal.h without version check
+#include <linux/eventfd.h>

 #include "vmware.h"
 #include "x86apic.h"
@@ -75,6 +78,7 @@

 #include "pgtbl.h"
 #include "versioned_atomic.h"
+#include "compat_poll.h"

 #if !defined(CONFIG_HIGH_RES_TIMERS)
 #error CONFIG_HIGH_RES_TIMERS required for acceptable performance
@@ -614,6 +618,15 @@ HostIF_FastClockUnlock(int callerID) // IN
   MutexUnlock(&fastClockMutex, callerID);
 }

+#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 8, 0)
+static int crosspage_set_exec(pte_t *pte, unsigned long addr, void *data)
+{
+	struct page *p = data;
+
+	set_pte(pte, mk_pte(p, VM_PAGE_KERNEL_EXEC));
+	return 0;
+}
+#endif

 /*
 *----------------------------------------------------------------------
@@ -634,7 +647,29 @@ HostIF_FastClockUnlock(int callerID) // IN
 static void *
 MapCrossPage(struct page *p)  // IN:
 {
+#if COMPAT_LINUX_VERSION_CHECK_LT(5, 8, 0)
   return vmap(&p, 1, VM_MAP, VM_PAGE_KERNEL_EXEC);
+#else
+   void *addr;
+
+   addr = vmap(&p, 1, VM_MAP, VM_PAGE_KERNEL_EXEC);
+   if (!addr)
+	   return NULL;
+
+   /* Starting with 5.8, vmap() always sets the NX bit, but the cross
+    * page needs to be executable. */
+   if (apply_to_page_range(current->mm, (unsigned long)addr, PAGE_SIZE,
+			   crosspage_set_exec, p)) {
+      vunmap(addr);
+      return NULL;
+   }
+
+   preempt_disable();
+   __flush_tlb_all();
+   preempt_enable();
+
+   return addr;
+#endif
 }


@@ -1738,8 +1773,11 @@ HostIF_EstimateLockedPageLimit(const VMDriver* vm,                // IN
 #else
   lockedPages += global_page_state(NR_PAGETABLE);
 #endif
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 9, 0)
+   /* NR_SLAB_* converted to byte counters in 5.9 */
+   lockedPages += global_node_page_state_pages(NR_SLAB_UNRECLAIMABLE_B);
+#elif LINUX_VERSION_CODE >= KERNEL_VERSION(4, 13, 0)
   /* NR_SLAB_* moved from zone to node in 4.13. */
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 13, 0)
   lockedPages += global_node_page_state(NR_SLAB_UNRECLAIMABLE);
 #else
   lockedPages += global_page_state(NR_SLAB_UNRECLAIMABLE);
@@ -2370,6 +2408,27 @@ HostIF_VMLockIsHeld(VMDriver *vm) // IN

 #if defined(CONFIG_SMP) || defined(CONFIG_X86_UP_IOAPIC) || \
    defined(CONFIG_X86_UP_APIC) || defined(CONFIG_X86_LOCAL_APIC)
+
+#if COMPAT_LINUX_VERSION_CHECK_LT(5, 8, 0)
+static long compat_copy_from_kernel_nofault(void *dst, VA src, size_t size)
+{
+   mm_segment_t old_fs;
+   long ret;
+
+   old_fs = get_fs();
+   set_fs(KERNEL_DS);
+   ret = HostIF_CopyFromUser(dst, src, size);
+   set_fs(old_fs);
+
+   return ret;
+}
+#else
+static long compat_copy_from_kernel_nofault(void *dst, VA src, size_t size)
+{
+   return copy_from_kernel_nofault(dst, (const void *)src, size);
+}
+#endif
+
 /*
 *----------------------------------------------------------------------
 *
@@ -2389,15 +2448,11 @@ HostIF_VMLockIsHeld(VMDriver *vm) // IN
 static Bool
 isVAReadable(VA r)  // IN:
 {
-   mm_segment_t old_fs;
   uint32 dummy;
-   int ret;
+   long ret;

-   old_fs = get_fs();
-   set_fs(KERNEL_DS);
   r = APICR_TO_ADDR(r, APICR_VERSION);
-   ret = HostIF_CopyFromUser(&dummy, r, sizeof dummy);
-   set_fs(old_fs);
+   ret = compat_copy_from_kernel_nofault(&dummy, r, sizeof(dummy));

   return ret == 0;
 }
@@ -2553,11 +2608,12 @@ HostIF_SemaphoreWait(VMDriver *vm,   // IN:
                     uint64 *args)   // IN:
 {
   struct file *file;
-   mm_segment_t old_fs;
   int res;
   int waitFD = args[0];
   int timeoutms = args[2];
   uint64 value;
+   struct poll_wqueues table;
+   unsigned int mask;

   ASSERT(vcpuid < vm->numVCPUs);

@@ -2566,33 +2622,29 @@ HostIF_SemaphoreWait(VMDriver *vm,   // IN:
      return MX_WAITERROR;
   }

-   old_fs = get_fs();
-   set_fs(KERNEL_DS);
-
-   {
-      struct poll_wqueues table;
-      unsigned int mask;
-
-      poll_initwait(&table);
-      current->state = TASK_INTERRUPTIBLE;
-      mask = file->f_op->poll(file, &table.pt);
-      if (!(mask & (POLLIN | POLLERR | POLLHUP))) {
-         vm->vmhost->vcpuSemaTask[vcpuid] = current;
-         schedule_timeout(timeoutms * HZ / 1000);  // convert to Hz
-         vm->vmhost->vcpuSemaTask[vcpuid] = NULL;
-      }
-      current->state = TASK_RUNNING;
-      poll_freewait(&table);
+   poll_initwait(&table);
+   current->state = TASK_INTERRUPTIBLE;
+   mask = compat_vfs_poll(file, &table.pt);
+   if (!(mask & (POLLIN | POLLERR | POLLHUP))) {
+      vm->vmhost->vcpuSemaTask[vcpuid] = current;
+      schedule_timeout(timeoutms * HZ / 1000);  // convert to Hz
+      vm->vmhost->vcpuSemaTask[vcpuid] = NULL;
   }
+   current->state = TASK_RUNNING;
+   poll_freewait(&table);

   /*
    * Userland only writes in multiples of sizeof(uint64). This will allow
    * the code to happily deal with a pipe or an eventfd. We only care about
    * reading no bytes (EAGAIN - non blocking fd) or sizeof(uint64).
+    *
+    * Upstream Linux changed the function parameter types/ordering in 4.14.0.
    */
-
-   res = file->f_op->read(file, (char *) &value, sizeof value, &file->f_pos);
-
+#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 14, 0)
+   res = kernel_read(file, file->f_pos, (char *)&value, sizeof value);
+#else
+   res = kernel_read(file, &value, sizeof value, &file->f_pos);
+#endif
   if (res == sizeof value) {
      res = MX_WAITNORMAL;
   } else {
@@ -2601,7 +2653,6 @@ HostIF_SemaphoreWait(VMDriver *vm,   // IN:
      }
   }

-   set_fs(old_fs);
   fput(file);

   /*
@@ -2688,8 +2739,8 @@ HostIF_SemaphoreForceWakeup(VMDriver *vm,       // IN:
 int
 HostIF_SemaphoreSignal(uint64 *args)  // IN:
 {
+   struct eventfd_ctx *eventfd;
   struct file *file;
-   mm_segment_t old_fs;
   int res;
   int signalFD = args[1];
   uint64 value = 1;  // make an eventfd happy should it be there
@@ -2699,22 +2750,34 @@ HostIF_SemaphoreSignal(uint64 *args)  // IN:
      return MX_WAITERROR;
   }

-   old_fs = get_fs();
-   set_fs(KERNEL_DS);
+   /*
+    * If it's eventfd, use specific eventfd interface as kernel writes
+    * to eventfd may not be allowed in kernel 5.10 and later.
+    */
+   eventfd = eventfd_ctx_fileget(file);
+   if (!IS_ERR(eventfd)) {
+      eventfd_signal(eventfd, 1);
+      fput(file);
+      return MX_WAITNORMAL;
+   }

   /*
    * Always write sizeof(uint64) bytes. This works fine for eventfd and
    * pipes. The data written is formatted to make an eventfd happy should
    * it be present.
+    *
+    * Upstream Linux changed the function parameter types/ordering in 4.14.0.
    */
-
-   res = file->f_op->write(file, (char *) &value, sizeof value, &file->f_pos);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 14, 0)
+   res = kernel_write(file, (char *)&value, sizeof value, file->f_pos);
+#else
+   res = kernel_write(file, &value, sizeof value, &file->f_pos);
+#endif

   if (res == sizeof value) {
      res = MX_WAITNORMAL;
   }

-   set_fs(old_fs);
   fput(file);

   /*
@@ -3261,12 +3324,9 @@ static int
 HostIFFastClockThread(void *unused)  // IN:
 {
   int res;
-   mm_segment_t oldFS;
   unsigned int rate = 0;
   unsigned int prevRate = 0;

-   oldFS = get_fs();
-   set_fs(KERNEL_DS);
   allow_signal(SIGKILL);

   while ((rate = linuxState.fastClockRate) > MIN_RATE) {
@@ -3289,8 +3349,6 @@ HostIFFastClockThread(void *unused)  // IN:
   }

 out:
-   set_fs(oldFS);
-
   /*
    * Do not exit thread until we are told to do so.
    */
@@ -136,9 +136,6 @@ ifdef VMX86_DEBUG
 CC_OPTS += -DVMX86_DEBUG
 endif

-# Add Spectre options when available
-CC_OPTS += $(call vm_check_gcc,-mindirect-branch=thunk -mindirect-branch-register,)
-
 include $(SRCROOT)/Makefile.kernel

 else
@@ -32,6 +32,6 @@ $(DRIVER)-y := driver.o hub.o userif.o netif.o bridge.o procfs.o smac_compat.o \
 ####

 clean:
-	rm -rf $(DRIVER).o $(DRIVER).mod.o $(DRIVER).mod.c $(DRIVER).ko \
+	rm -rf $(DRIVER).o $(DRIVER).mod.o $(DRIVER).mod.c $(DRIVER).ko .cache.mk \
 	       .tmp_versions Module.symvers Modules.symvers Module.markers \
 	       modules.order $($(DRIVER)-y) .*.cmd .*.o.flags
@@ -66,7 +66,7 @@
 #endif

 #if LOGLEVEL >= 4
-static struct timeval vnetTime;
+static u64 vnetTime;
 #endif

 typedef struct VNetBridge VNetBridge;
@@ -693,7 +693,7 @@ VNetBridgeReceiveFromVNet(VNetJack        *this, // IN: jack

 	 netif_rx_ni(clone);
 #	 if LOGLEVEL >= 4
-	 do_gettimeofday(&vnetTime);
+	 vnetTime = ktime_get_ns();
 #	 endif
      }
   }
@@ -1490,12 +1490,11 @@ VNetBridgeReceiveFromDev(struct sk_buff *skb,         // IN: packet to receive

 #  if LOGLEVEL >= 4
   {
-      struct timeval now;
-      do_gettimeofday(&now);
+      u64 now;
+
+      now = ktime_get_ns();
      LOG(3, (KERN_DEBUG "bridge-%s: time %d\n",
-	      bridge->name,
-	      (int)((now.tv_sec * 1000000 + now.tv_usec)
-                    - (vnetTime.tv_sec * 1000000 + vnetTime.tv_usec))));
+	      bridge->name, (int)((now - vnetTime) / NSEC_PER_USEC)));
   }
 #  endif

@@ -81,10 +81,30 @@ extern unsigned int  vnet_max_qlen;
 #   define skb_frag_page(frag) (frag)->page
 #   define skb_frag_size(frag) (frag)->size
 #endif
-#if COMPAT_LINUX_VERSION_CHECK_LT(5, 4, 0)
+#if COMPAT_LINUX_VERSION_CHECK_LT(5, 4, 0) && \
+    !(defined(CONFIG_SUSE_VERSION) && CONFIG_SUSE_VERSION == 15 && \
+      defined(CONFIG_SUSE_PATCHLEVEL) && CONFIG_SUSE_PATCHLEVEL >= 2)
 #   define skb_frag_off(frag) (frag)->page_offset
 #endif

+#if COMPAT_LINUX_VERSION_CHECK_LT(5, 10, 0)
+static inline
+__wsum compat_csum_and_copy_to_user(const void *src, void __user *dst, int len)
+{
+	int err;
+	__wsum ret;
+
+	ret = csum_and_copy_to_user(src, dst, len, 0, &err);
+	return err ? 0 : ret;
+}
+#else
+static inline
+__wsum compat_csum_and_copy_to_user(const void *src, void __user *dst, int len)
+{
+	return csum_and_copy_to_user(src, dst, len);
+}
+#endif
+
 /*
 *-----------------------------------------------------------------------------
 *
@@ -546,7 +566,6 @@ VNetCsumCopyDatagram(const struct sk_buff *skb,	// IN: skb to copy
 		     char *buf)			// OUT: where to copy data
 {
   unsigned int csum;
-   int err = 0;
   int len = skb_headlen(skb) - offset;
   char *curr = buf;
   const skb_frag_t *frag;
@@ -559,10 +578,9 @@ VNetCsumCopyDatagram(const struct sk_buff *skb,	// IN: skb to copy
      return -EINVAL;
   }

-   csum = csum_and_copy_to_user(skb->data + offset, curr, len, 0, &err);
-   if (err) {
-      return err;
-   }
+   csum = compat_csum_and_copy_to_user(skb->data + offset, curr, len);
+   if (!csum)
+	   return -EFAULT;
   curr += len;

   for (frag = skb_shinfo(skb)->frags;
@@ -573,13 +591,12 @@ VNetCsumCopyDatagram(const struct sk_buff *skb,	// IN: skb to copy
 	 const void *vaddr;

 	 vaddr = kmap(skb_frag_page(frag));
-	 tmpCsum = csum_and_copy_to_user(vaddr + skb_frag_off(frag),
-					 curr, skb_frag_size(frag), 0, &err);
+	 tmpCsum = compat_csum_and_copy_to_user(vaddr + skb_frag_off(frag),
+						curr, skb_frag_size(frag));
 	 kunmap(skb_frag_page(frag));

-	 if (err) {
-	    return err;
-	 }
+	 if (!tmpCsum)
+		 return -EFAULT;
 	 csum = csum_block_add(csum, tmpCsum, curr - buf);
 	 curr += skb_frag_size(frag);
      }
@@ -35,6 +35,8 @@
 #endif
 #endif

+#include <linux/pci_ids.h>
+
 /* LSILogic 53C1030 Parallel SCSI controller
 * LSILogic SAS1068 SAS controller
 */
@@ -53,7 +55,10 @@
 *    VMware HD Audio codec
 *    VMware HD Audio controller
 */
+#ifndef PCI_VENDOR_ID_VMWARE
 #define PCI_VENDOR_ID_VMWARE                    0x15AD
+#endif
+
 #define PCI_DEVICE_ID_VMWARE_SVGA3              0x0406
 #define PCI_DEVICE_ID_VMWARE_SVGA2              0x0405
 #define PCI_DEVICE_ID_VMWARE_SVGA               0x0710
@@ -71,7 +76,11 @@
 #define PCI_DEVICE_ID_VMWARE_1394               0x0780
 #define PCI_DEVICE_ID_VMWARE_BRIDGE             0x0790
 #define PCI_DEVICE_ID_VMWARE_ROOTPORT           0x07A0
+
+#ifndef PCI_DEVICE_ID_VMWARE_VMXNET3
 #define PCI_DEVICE_ID_VMWARE_VMXNET3            0x07B0
+#endif
+
 #define PCI_DEVICE_ID_VMWARE_PVSCSI             0x07C0
 #define PCI_DEVICE_ID_VMWARE_82574              0x07D0
 #define PCI_DEVICE_ID_VMWARE_AHCI               0x07E0
Author	SHA1	Message	Date
Michal Kubecek	6ad950dfdd	vmmon: use build time version check in MapCrossPage() Previous commit breaks build against kernels older than 5.3 as there used to be one more argument of handler called by apply_to_page_range(). Use preprocessor version check so that code intended for kernel 5.8 and later is not compiled at all when building against older kernels. Also compile crosspage_set_exec() conditionally to avoid compiler warning.	2020-12-09 10:42:10 +01:00
Michal Kubecek	605977a729	vmmon: avoid using alloc_vm_area() for crosspage allocation Mainline commit 301fa9f2ddf7 ("mm: remove alloc_vm_area") in v5.10-rc1 removed alloc_vm_area() function used to allow allocating crosspage as executable after vmap() started silently setting NX flag in kernel 5.8. Similar to what upstream did in Xen code, use apply_to_page_range() to fix the crosspage flags, except we don't need to use get_vm_area() which is not exported. Thanks to virtualbox for the hint that current->mm can be used instead of init_mm (which is not exported either).	2020-12-09 09:17:59 +01:00
Michal Kubecek	46e02c3201	vmmon: avoid kernel writes to eventfd Starting with mainline commit 4d03e3cc5982 ("fs: don't allow kernel reads and writes without iter ops") in 5.10-rc1, kernel writes are only allowed for files providing ->write_iter() callback which eventfd does not. As an attempt to address this issue on kernel side by providing ->write_iter() callback to eventfd was met with a pushback and it is unlikely to be accepted any time soon, let's use a specific eventfd interface eventfd_signal() instead. However, as the comments in HostIF_SemaphoreSignal() indicate that the file descriptor may also be a pipe (without telling when), let's keep kernel_write() as a fallback in case the file descriptor is not eventfd. Signed-off-by: Michal Kubecek <mkubecek@suse.cz>	2020-12-09 09:17:59 +01:00
Michal Kubecek	bd5ee1a786	vmnet: use new compat_csum_and_copy_to_user() calling convention Part of bigger csum_and_copy rework, commit c693cc4676a0 ("saner calling conventions for csum_and_copy_..._user()") in 5.10-rc1 drops last two arguments and changes the way errors are reported. Provide a compat wrapper with new calling convention and rewrite callers in vmnet to use it. Unfortunately there doesn't seem to be a way to recognize which calling convention kernel has so version check is used.	2020-12-09 09:17:59 +01:00
Michal Kubecek	56a9cef633	vmmon: avoid set_fs() in HostIFFastClockThread() The set_fs() function was dropped on x86_64 architecture by commit 47058bb54b57 ("x86: remove address space overrides using set_fs()") in 5.10-rc1. It's not clear what purpose does set_fs(KERNEL_DS) serve in HostIFFastClockThread(); it was certainly needed for the old code using functions expecting user pointers to access /dev/rtc but that code has been dropped between Workstation 12.5.7 and 14.0.0 in September 2017. Drop the temporary set_fs(KERNEL_DS) trick as there doesn't seem to be a need for it any more.	2020-12-09 09:17:59 +01:00
Michal Kubecek	a83868918d	vmmon: avoid set_fs() in isVAReadable() The set_fs() function was dropped on x86_64 architecture by commit 47058bb54b57 ("x86: remove address space overrides using set_fs()") in 5.10-rc1. In isVAReadable() it is used to check if a kernel address can be read from without risking a pagefault if the answer is negative by using set_fs() to pretend it's a userspace address and checking with copy_from_user(). Starting with 5.8, copy_from_kernel_nofault() does can be used to copy from a kernel address with access check but without an actual pagefault.	2020-12-09 09:17:59 +01:00
Michal Kubecek	1a90e9e8ae	vmmon: avoid set_fs() in HostIF_SemaphoreWait() and HostIF_SemaphoreSignal() The set_fs() function was dropped on x86_64 architecture by commit 47058bb54b57 ("x86: remove address space overrides using set_fs()") in 5.10-rc1. In HostIF_SemaphoreWait() it is used to allow calling read() file op on a kernel pointer, in HostIF_SemaphoreSignal() to allow calling write() file op on a kernel pointer. This is only the case for pre-5.8 kernels as we need to use kernel_read() and kernel_write() helpers on 5.8 and newer anyway. Use kernel_read() and kernel_write() always, like Workstation 16.0.0 does, so that we no longer need the set_fs(KERNEL_DS) trick. Make the code of both functions look exactly like in Workstation 16.0.0 to make future backports easier.	2020-12-09 09:17:59 +01:00
Jan Andres	51b3ade634	Fix NULL pointer dereference in eventfd read call Starting with commit 12aceb89b0bc ("eventfd: convert to f_op->read_iter()") in 5.8-rc1, the "read" function pointer in eventfd's file_operations is NULL, "read_iter" is available instead. Use kernel_read() and kernel_write() instead of directly calling the function pointers to handle this correctly.	2020-09-07 00:32:20 +02:00
Jan Andres	9731b69900	Fix NX bit handling for Linux 5.8+ Do not use vmap() to map the cross page, it needs to be executable and vmap() unconditionally sets the NX bit starting with commit cca98e9f8b5e ("mm: enforce that vmap can't map pages executable") in 5.8-rc1. Emulate previous behavior of vmap() by using alloc_vm_area() and explicitly setting the PTE.	2020-09-07 00:32:20 +02:00
Michal Kubecek	eddbdcf9ff	vmmon: include <asm/irq_vectors.h> explicitly After mainline commit 13c01139b171 ("x86/headers: Remove APIC headers from <asm/smp.h>") in 5.9-rc1, APIC headers are no longer included via <asm/smp.h> so that linux/hostif.c will use incorrect fallback definitions of SPURIOUS_APIC_VECTOR and POSTED_INTR_VECTOR even if built against kernel where these are defined. Include <asm/irq_vectors.h> in linux/hostif.c explicitly to avoid that.	2020-08-20 11:11:51 +02:00
Michal Kubecek	3281356efd	vmmon: handle change of vmstat slab counters to bytes Mainline commit d42f3245c7e2 ("mm: memcg: convert vmstat slab counters to bytes") in kernel 5.9-rc1 changed vmstat accounting of slab to byte counters. Unfortunately there doesn't seem to be any obvious way to check for this change except version check.	2020-08-20 10:14:43 +02:00
Michal Kubecek	7f41ad47c8	vmnet: work around skb_frag_t build failure for SLE15-SP2 and openSUSE-15.2 The kernels used in (upcoming) SLE15 SP2 and openSUSE Leap 15.2 are based on upstream version 5.3 but have a backport of the skb_frag_t conversion to bio_vec which reached mainline in v5.4-rc1. As we cannot easily check availability compat_skb_frag_off() (which was also added in v5.4-rc1) and have to resort to version check, add an exception for SLE15-SP2 and openSUSE-15.2 kernels.	2020-05-30 02:58:53 +02:00
Michal Kubecek	89dd7f9c95	vmmon: use standard definition of MSR_TSX_FORCE_ABORT if available The MSR_TSX_FORCE_ABORT macro is defined in mainline since commit 52f64909409c ("x86: Add TSX Force Abort CPUID/MSR") in v5.1-rc1.	2020-05-30 02:58:53 +02:00
Michal Kubecek	b1932e4565	vmnet: convert VNetBridge debugging code to ktime_get_ns() The do_gettimeofday() helper was removed by commit e4b92b108c6c ("timekeeping: remove obsolete time accessors") in v5.0-rc1 and since commit c766d1472c70 ("y2038: hide timeval/timespec/itimerval/itimerspec types") in v5.6-rc3, struct timeval should no longer be used in kernel code. Convert the do_gettimeofday() relics in VNetBridge (which are only compiled with LOGLEVEL >= 4) completely to ktime based interface.	2020-05-30 02:58:29 +02:00
Michal Kubecek	62defb4cdc	vmmon: check presence of file_operations::poll() Commit 11c5ad0ec441 ("eventpoll: switch to ->poll_mask") in v4.18-rc1, made eventpoll switch from ->poll() to ->poll_mask(), resulting in a null pointer dereference as vmmon calls the callback directly without checking its availability. Rather than calling the callback directly (which would result in null pointer dereference), use vfs_poll() wrapper. As this wrapper is only available since 4.18-rc1 cycle as well, provide a copy to use when building against older kernels. Note: even if the commit revealing the problem was reverted in 4.18-rc3 (and is unlikely to go back in the same form) so that vmmon no longer crashes with 4.18 kernel, calling the callback without a check is a bad practice so let's keep the patch in place.	2020-05-30 02:57:49 +02:00
Michal Kubecek	9eaf7f92ac	vmmon: fix indirect call with retpoline build Build against kernel with retpoline support issues warning objtool: Task_Switch()+0x425: indirect call found in RETPOLINE build This is because an indirect call in TaskSwitchToMonitor() is encoded using inline assembler so that it bypasses retpoline generation. For this purpose, macro CALL_NOSPEC exists since v4.15-rc8 (and has been backported into some distribution kernels with the rest of retpoline support). Use the macro if available and fallback to the original code if not.	2020-05-30 02:57:49 +02:00
Michal Kubecek	4afafaa6d3	vmmon: fix always_inline attribute usage Function declared with __attribute__((always_inline)) should also be declared as inline, otherwise gcc issues a warning "always_inline function might not be inlinable". It's just cosmetic but getting rid of known harmless warnings makes it easier to spot actual problems. Use the __always_inline macro for LinuxDriverSyncReadTSCs() as this is how always inline functions should be declared in kernel code.	2020-05-30 02:57:49 +02:00
Michal Kubecek	7f158e4ce9	vmmon: use standard definition of MSR_K7_HWCR_SMMLOCK if available The MSR_K7_HWCR_SMMLOCK macro is defined in mainline since commit 18c71ce9c882 ("x86/CPU/AMD: Add the Secure Encrypted Virtualization CPU feature") in v4.16-rc1.	2020-05-30 02:57:49 +02:00
Theo Bertacchini	cf7e20ba3a	modules: remove .cache.mk on make clean Remove .cache.mk inside vmmon-only and vmnet-only when executing "make clean". This file can cause issues when upgrading gcc as make will still look for includes inside older gcc includes directory. File .cache.mk with cache of generated variables was created by build since kernel v4.15-rc1, commit 3298b690b21c ("kbuild: Add a cache for generated variables") until the feature was removed in v4.18-rc1, commit e08d6de4e532 ("kbuild: remove kbuild cache").	2020-05-30 02:57:49 +02:00
Michal Kubecek	75ac365cc8	vmmon: quick workaround for objtool warnings As discussed in https://bugzilla.suse.com/show_bug.cgi?id=1059674 the reason for multiple objtool warnings is the fact that vmmon module defines its own Panic() function which never returns. While it is marked as such which is used by the compiler for optimization, there is no way to find this from object file. While this seems innocuous, it might result in problems with unwinder later. The quickest way around is to replace vmmon's own Panic() with standard kernel panic() until a cleaner solution is found.	2020-05-30 02:57:49 +02:00
Michal Kubecek	afb98b4983	vmmon: use standard definition of CR3_PCID_MASK if available The CR3_PCID_MASK is defined in mainline asm/processor-flags.h since commit 6c690ee1039b ("x86/mm: Split read_cr3() into read_cr3_pa() and __read_cr3()") in v4.13-rc1.	2020-05-30 02:57:49 +02:00
Michal Kubecek	db81a381d0	vmmon: use standard definition of MSR_MISC_FEATURES_ENABLES if available The MSR_MISC_FEATURES_ENABLES macro is defined in mainline since commit ab6d9468631a ("x86/msr: Rename MISC_FEATURE_ENABLES to MISC_FEATURES_ENABLES") in v4.12-rc1. The MSR_MISC_FEATURES_ENABLES_CPUID_FAULT macro is defined in mainline since commit e9ea1e7f53b8 ("x86/arch_prctl: Add ARCH_[GET\|SET]_CPUID") in v4.12-rc1.	2020-05-30 02:57:49 +02:00
Michal Kubecek	cd660d0dbe	vmnet: use standard definition of PCI_VENDOR_ID_VMWARE_VMXNET3 if available The PCI_DEVICE_ID_VMWARE_VMXNET3 macro is defined in mainline pci_ids.h since commit b1226c7db1d9 ("vmxnet3: Move PCI Id to pci_ids.h") in v4.10-rc1.	2020-05-30 02:57:49 +02:00
Michal Kubecek	2c723cea23	vmnet: use standard definition of PCI_VENDOR_ID_VMWARE if available The PCI_VENDOR_ID_VMWARE macro is defined in mainline pci_ids.h since commit 94e57fea6202 ("PCI: Move PCI_VENDOR_ID_VMWARE to pci_ids.h") in v3.18-rc1.	2020-05-30 02:56:37 +02:00
Michal Kubecek	a8a060c095	modules: do not override CC command line options Workstation/Player 15.5.0 tries to guess the right retpoline (Spectre) related compiler command line options, resulting in objtool warnings like warning: objtool: .text.__x86_indirect_thunk_rbp+0x0: unsupported intra-function call warning: objtool: If this is a retpoline, please patch it in with alternatives and annotate e.g. in vmmon-only/common/memtrack.o (but also many others). This seems wrong; out-of-tree modules should be built with the same compiler options that were used to build the kernel they are built against. Distribution kernel devel packages usually make sure this is the way so let them do their job.	2020-05-30 02:56:37 +02:00
Michal Kubecek	60c5393399	vmmon: fix wrong VMX86_DEVEL test If VMX86_DEVEL macro is not defined, preprocessor warns when it is tested with "#if" in vmmon-only/linux/driver.c. As VMX86_DEVEL is tested with "#ifdef" everywhere else, change this one test to match the rest.	2020-05-30 02:56:37 +02:00