diff --git a/xen/arch/x86/hvm/stdvga.c b/xen/arch/x86/hvm/stdvga.c
index f5a4b1cfe9..3a84a7c35e 100644
--- a/xen/arch/x86/hvm/stdvga.c
+++ b/xen/arch/x86/hvm/stdvga.c
@@ -104,7 +104,7 @@ static void vram_put(struct hvm_hw_stdvga *s, void *p)
 static int stdvga_outb(uint64_t addr, uint8_t val)
 {
     struct hvm_hw_stdvga *s = &current->domain->arch.hvm.stdvga;
-    int rc = 1, prev_stdvga = s->stdvga;
+    int rc = 1;
 
     switch ( addr )
     {
@@ -133,19 +133,6 @@ static int stdvga_outb(uint64_t addr, uint8_t val)
         break;
     }
 
-    /* When in standard vga mode, emulate here all writes to the vram buffer
-     * so we can immediately satisfy reads without waiting for qemu. */
-    s->stdvga = (s->sr[7] == 0x00);
-
-    if ( !prev_stdvga && s->stdvga )
-    {
-        gdprintk(XENLOG_INFO, "entering stdvga mode\n");
-    }
-    else if ( prev_stdvga && !s->stdvga )
-    {
-        gdprintk(XENLOG_INFO, "leaving stdvga mode\n");
-    }
-
     return rc;
 }
 
@@ -426,7 +413,6 @@ static int cf_check stdvga_mem_write(
     const struct hvm_io_handler *handler, uint64_t addr, uint32_t size,
     uint64_t data)
 {
-    struct hvm_hw_stdvga *s = &current->domain->arch.hvm.stdvga;
     ioreq_t p = {
         .type = IOREQ_TYPE_COPY,
         .addr = addr,
@@ -437,8 +423,7 @@ static int cf_check stdvga_mem_write(
     };
     struct ioreq_server *srv;
 
-    if ( true || !s->stdvga )
-        goto done;
+    goto done;
 
     /* Intercept mmio write */
     switch ( size )
@@ -499,19 +484,14 @@ static bool cf_check stdvga_mem_accept(
 
     spin_lock(&s->lock);
 
-    if ( p->dir == IOREQ_WRITE && (p->data_is_ptr || p->count != 1) )
+    if ( p->dir != IOREQ_WRITE || p->data_is_ptr || p->count != 1 )
     {
         /*
-         * We cannot return X86EMUL_UNHANDLEABLE on anything other then the
-         * first cycle of an I/O. So, since we cannot guarantee to always be
-         * able to send buffered writes, we have to reject any multi-cycle
-         * or "indirect" I/O.
+         * Only accept single direct writes, as that's the only thing we can
+         * accelerate using buffered ioreq handling.
          */
         goto reject;
     }
-    else if ( p->dir == IOREQ_READ &&
-              (true || !s->stdvga) )
-        goto reject;
 
     /* s->lock intentionally held */
     return 1;
diff --git a/xen/arch/x86/include/asm/hvm/io.h b/xen/arch/x86/include/asm/hvm/io.h
index d92f87f83b..d964f8efdf 100644
--- a/xen/arch/x86/include/asm/hvm/io.h
+++ b/xen/arch/x86/include/asm/hvm/io.h
@@ -126,7 +126,6 @@ struct hvm_hw_stdvga {
     uint8_t sr[8];
     uint8_t gr_index;
     uint8_t gr[9];
-    bool_t stdvga;
     uint32_t latch;
     struct page_info *vram_page[64];  /* shadow of 0xa0000-0xaffff */
     spinlock_t lock;