Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

3 files changed, 352 insertions, 0 deletions

diff --git a/Documentation/driver-api/virtio/index.rst b/Documentation/driver-api/virtio/index.rst
new file mode 100644
index 000000000..528b14b29
--- /dev/null
+++ b/Documentation/driver-api/virtio/index.rst
@@ -0,0 +1,11 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+======
+Virtio
+======
+
+.. toctree::
+   :maxdepth: 1
+
+   virtio
+   writing_virtio_drivers
diff --git a/Documentation/driver-api/virtio/virtio.rst b/Documentation/driver-api/virtio/virtio.rst
new file mode 100644
index 000000000..7947b4ca6
--- /dev/null
+++ b/Documentation/driver-api/virtio/virtio.rst
@@ -0,0 +1,145 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+.. _virtio:
+
+===============
+Virtio on Linux
+===============
+
+Introduction
+============
+
+Virtio is an open standard that defines a protocol for communication
+between drivers and devices of different types, see Chapter 5 ("Device
+Types") of the virtio spec (`[1]`_). Originally developed as a standard
+for paravirtualized devices implemented by a hypervisor, it can be used
+to interface any compliant device (real or emulated) with a driver.
+
+For illustrative purposes, this document will focus on the common case
+of a Linux kernel running in a virtual machine and using paravirtualized
+devices provided by the hypervisor, which exposes them as virtio devices
+via standard mechanisms such as PCI.
+
+
+Device - Driver communication: virtqueues
+=========================================
+
+Although the virtio devices are really an abstraction layer in the
+hypervisor, they're exposed to the guest as if they are physical devices
+using a specific transport method -- PCI, MMIO or CCW -- that is
+orthogonal to the device itself. The virtio spec defines these transport
+methods in detail, including device discovery, capabilities and
+interrupt handling.
+
+The communication between the driver in the guest OS and the device in
+the hypervisor is done through shared memory (that's what makes virtio
+devices so efficient) using specialized data structures called
+virtqueues, which are actually ring buffers [#f1]_ of buffer descriptors
+similar to the ones used in a network device:
+
+.. kernel-doc:: include/uapi/linux/virtio_ring.h
+    :identifiers: struct vring_desc
+
+All the buffers the descriptors point to are allocated by the guest and
+used by the host either for reading or for writing but not for both.
+
+Refer to Chapter 2.5 ("Virtqueues") of the virtio spec (`[1]`_) for the
+reference definitions of virtqueues and "Virtqueues and virtio ring: How
+the data travels" blog post (`[2]`_) for an illustrated overview of how
+the host device and the guest driver communicate.
+
+The :c:type:`vring_virtqueue` struct models a virtqueue, including the
+ring buffers and management data. Embedded in this struct is the
+:c:type:`virtqueue` struct, which is the data structure that's
+ultimately used by virtio drivers:
+
+.. kernel-doc:: include/linux/virtio.h
+    :identifiers: struct virtqueue
+
+The callback function pointed by this struct is triggered when the
+device has consumed the buffers provided by the driver. More
+specifically, the trigger will be an interrupt issued by the hypervisor
+(see vring_interrupt()). Interrupt request handlers are registered for
+a virtqueue during the virtqueue setup process (transport-specific).
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: vring_interrupt
+
+
+Device discovery and probing
+============================
+
+In the kernel, the virtio core contains the virtio bus driver and
+transport-specific drivers like `virtio-pci` and `virtio-mmio`. Then
+there are individual virtio drivers for specific device types that are
+registered to the virtio bus driver.
+
+How a virtio device is found and configured by the kernel depends on how
+the hypervisor defines it. Taking the `QEMU virtio-console
+<https://gitlab.com/qemu-project/qemu/-/blob/master/hw/char/virtio-console.c>`__
+device as an example. When using PCI as a transport method, the device
+will present itself on the PCI bus with vendor 0x1af4 (Red Hat, Inc.)
+and device id 0x1003 (virtio console), as defined in the spec, so the
+kernel will detect it as it would do with any other PCI device.
+
+During the PCI enumeration process, if a device is found to match the
+virtio-pci driver (according to the virtio-pci device table, any PCI
+device with vendor id = 0x1af4)::
+
+	/* Qumranet donated their vendor ID for devices 0x1000 thru 0x10FF. */
+	static const struct pci_device_id virtio_pci_id_table[] = {
+		{ PCI_DEVICE(PCI_VENDOR_ID_REDHAT_QUMRANET, PCI_ANY_ID) },
+		{ 0 }
+	};
+
+then the virtio-pci driver is probed and, if the probing goes well, the
+device is registered to the virtio bus::
+
+	static int virtio_pci_probe(struct pci_dev *pci_dev,
+				    const struct pci_device_id *id)
+	{
+		...
+
+		if (force_legacy) {
+			rc = virtio_pci_legacy_probe(vp_dev);
+			/* Also try modern mode if we can't map BAR0 (no IO space). */
+			if (rc == -ENODEV || rc == -ENOMEM)
+				rc = virtio_pci_modern_probe(vp_dev);
+			if (rc)
+				goto err_probe;
+		} else {
+			rc = virtio_pci_modern_probe(vp_dev);
+			if (rc == -ENODEV)
+				rc = virtio_pci_legacy_probe(vp_dev);
+			if (rc)
+				goto err_probe;
+		}
+
+		...
+
+		rc = register_virtio_device(&vp_dev->vdev);
+
+When the device is registered to the virtio bus the kernel will look
+for a driver in the bus that can handle the device and call that
+driver's ``probe`` method.
+
+At this point, the virtqueues will be allocated and configured by
+calling the appropriate ``virtio_find`` helper function, such as
+virtio_find_single_vq() or virtio_find_vqs(), which will end up calling
+a transport-specific ``find_vqs`` method.
+
+
+References
+==========
+
+_`[1]` Virtio Spec v1.2:
+https://docs.oasis-open.org/virtio/virtio/v1.2/virtio-v1.2.html
+
+.. Check for later versions of the spec as well.
+
+_`[2]` Virtqueues and virtio ring: How the data travels
+https://www.redhat.com/en/blog/virtqueues-and-virtio-ring-how-data-travels
+
+.. rubric:: Footnotes
+
+.. [#f1] that's why they may be also referred to as virtrings.
diff --git a/Documentation/driver-api/virtio/writing_virtio_drivers.rst b/Documentation/driver-api/virtio/writing_virtio_drivers.rst
new file mode 100644
index 000000000..e5de6f5d0
--- /dev/null
+++ b/Documentation/driver-api/virtio/writing_virtio_drivers.rst
@@ -0,0 +1,196 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+.. _writing_virtio_drivers:
+
+======================
+Writing Virtio Drivers
+======================
+
+Introduction
+============
+
+This document serves as a basic guideline for driver programmers that
+need to hack a new virtio driver or understand the essentials of the
+existing ones. See :ref:`Virtio on Linux <virtio>` for a general
+overview of virtio.
+
+
+Driver boilerplate
+==================
+
+As a bare minimum, a virtio driver needs to register in the virtio bus
+and configure the virtqueues for the device according to its spec, the
+configuration of the virtqueues in the driver side must match the
+virtqueue definitions in the device. A basic driver skeleton could look
+like this::
+
+	#include <linux/virtio.h>
+	#include <linux/virtio_ids.h>
+	#include <linux/virtio_config.h>
+	#include <linux/module.h>
+
+	/* device private data (one per device) */
+	struct virtio_dummy_dev {
+		struct virtqueue *vq;
+	};
+
+	static void virtio_dummy_recv_cb(struct virtqueue *vq)
+	{
+		struct virtio_dummy_dev *dev = vq->vdev->priv;
+		char *buf;
+		unsigned int len;
+
+		while ((buf = virtqueue_get_buf(dev->vq, &len)) != NULL) {
+			/* process the received data */
+		}
+	}
+
+	static int virtio_dummy_probe(struct virtio_device *vdev)
+	{
+		struct virtio_dummy_dev *dev = NULL;
+
+		/* initialize device data */
+		dev = kzalloc(sizeof(struct virtio_dummy_dev), GFP_KERNEL);
+		if (!dev)
+			return -ENOMEM;
+
+		/* the device has a single virtqueue */
+		dev->vq = virtio_find_single_vq(vdev, virtio_dummy_recv_cb, "input");
+		if (IS_ERR(dev->vq)) {
+			kfree(dev);
+			return PTR_ERR(dev->vq);
+
+		}
+		vdev->priv = dev;
+
+		/* from this point on, the device can notify and get callbacks */
+		virtio_device_ready(vdev);
+
+		return 0;
+	}
+
+	static void virtio_dummy_remove(struct virtio_device *vdev)
+	{
+		struct virtio_dummy_dev *dev = vdev->priv;
+
+		/*
+		 * disable vq interrupts: equivalent to
+		 * vdev->config->reset(vdev)
+		 */
+		virtio_reset_device(vdev);
+
+		/* detach unused buffers */
+		while ((buf = virtqueue_detach_unused_buf(dev->vq)) != NULL) {
+			kfree(buf);
+		}
+
+		/* remove virtqueues */
+		vdev->config->del_vqs(vdev);
+
+		kfree(dev);
+	}
+
+	static const struct virtio_device_id id_table[] = {
+		{ VIRTIO_ID_DUMMY, VIRTIO_DEV_ANY_ID },
+		{ 0 },
+	};
+
+	static struct virtio_driver virtio_dummy_driver = {
+		.driver.name =  KBUILD_MODNAME,
+		.id_table =     id_table,
+		.probe =        virtio_dummy_probe,
+		.remove =       virtio_dummy_remove,
+	};
+
+	module_virtio_driver(virtio_dummy_driver);
+	MODULE_DEVICE_TABLE(virtio, id_table);
+	MODULE_DESCRIPTION("Dummy virtio driver");
+	MODULE_LICENSE("GPL");
+
+The device id ``VIRTIO_ID_DUMMY`` here is a placeholder, virtio drivers
+should be added only for devices that are defined in the spec, see
+include/uapi/linux/virtio_ids.h. Device ids need to be at least reserved
+in the virtio spec before being added to that file.
+
+If your driver doesn't have to do anything special in its ``init`` and
+``exit`` methods, you can use the module_virtio_driver() helper to
+reduce the amount of boilerplate code.
+
+The ``probe`` method does the minimum driver setup in this case
+(memory allocation for the device data) and initializes the
+virtqueue. virtio_device_ready() is used to enable the virtqueue and to
+notify the device that the driver is ready to manage the device
+("DRIVER_OK"). The virtqueues are anyway enabled automatically by the
+core after ``probe`` returns.
+
+.. kernel-doc:: include/linux/virtio_config.h
+    :identifiers: virtio_device_ready
+
+In any case, the virtqueues need to be enabled before adding buffers to
+them.
+
+Sending and receiving data
+==========================
+
+The virtio_dummy_recv_cb() callback in the code above will be triggered
+when the device notifies the driver after it finishes processing a
+descriptor or descriptor chain, either for reading or writing. However,
+that's only the second half of the virtio device-driver communication
+process, as the communication is always started by the driver regardless
+of the direction of the data transfer.
+
+To configure a buffer transfer from the driver to the device, first you
+have to add the buffers -- packed as `scatterlists` -- to the
+appropriate virtqueue using any of the virtqueue_add_inbuf(),
+virtqueue_add_outbuf() or virtqueue_add_sgs(), depending on whether you
+need to add one input `scatterlist` (for the device to fill in), one
+output `scatterlist` (for the device to consume) or multiple
+`scatterlists`, respectively. Then, once the virtqueue is set up, a call
+to virtqueue_kick() sends a notification that will be serviced by the
+hypervisor that implements the device::
+
+	struct scatterlist sg[1];
+	sg_init_one(sg, buffer, BUFLEN);
+	virtqueue_add_inbuf(dev->vq, sg, 1, buffer, GFP_ATOMIC);
+	virtqueue_kick(dev->vq);
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_add_inbuf
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_add_outbuf
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_add_sgs
+
+Then, after the device has read or written the buffers prepared by the
+driver and notifies it back, the driver can call virtqueue_get_buf() to
+read the data produced by the device (if the virtqueue was set up with
+input buffers) or simply to reclaim the buffers if they were already
+consumed by the device:
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_get_buf_ctx
+
+The virtqueue callbacks can be disabled and re-enabled using the
+virtqueue_disable_cb() and the family of virtqueue_enable_cb() functions
+respectively. See drivers/virtio/virtio_ring.c for more details:
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_disable_cb
+
+.. kernel-doc:: drivers/virtio/virtio_ring.c
+    :identifiers: virtqueue_enable_cb
+
+But note that some spurious callbacks can still be triggered under
+certain scenarios. The way to disable callbacks reliably is to reset the
+device or the virtqueue (virtio_reset_device()).
+
+
+References
+==========
+
+_`[1]` Virtio Spec v1.2:
+https://docs.oasis-open.org/virtio/virtio/v1.2/virtio-v1.2.html
+
+Check for later versions of the spec as well.