1
0
Fork 0
alistair23-linux/drivers/firewire/core-cdev.c

1802 lines
46 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Char device for device raw access
*
* Copyright (C) 2005-2007 Kristian Hoegsberg <krh@bitplanet.net>
*/
#include <linux/bug.h>
#include <linux/compat.h>
#include <linux/delay.h>
#include <linux/device.h>
#include <linux/dma-mapping.h>
#include <linux/errno.h>
#include <linux/firewire.h>
#include <linux/firewire-cdev.h>
#include <linux/idr.h>
#include <linux/irqflags.h>
#include <linux/jiffies.h>
#include <linux/kernel.h>
#include <linux/kref.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/poll.h>
#include <linux/sched.h> /* required for linux/wait.h */
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 02:04:11 -06:00
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/string.h>
#include <linux/time.h>
#include <linux/uaccess.h>
#include <linux/vmalloc.h>
#include <linux/wait.h>
#include <linux/workqueue.h>
#include "core.h"
/*
* ABI version history is documented in linux/firewire-cdev.h.
*/
#define FW_CDEV_KERNEL_VERSION 5
firewire: cdev: improve FW_CDEV_IOC_ALLOCATE In both the ieee1394 stack and the firewire stack, the core treats kernelspace drivers better than userspace drivers when it comes to CSR address range allocation: The former may request a register to be placed automatically at a free spot anywhere inside a specified address range. The latter may only request a register at a fixed offset. Hence, userspace drivers which do not require a fixed offset potentially need to implement a retry loop with incremented offset in each retry until the kernel does not fail allocation with EBUSY. This awkward procedure is not fundamentally necessary as the core already provides a superior allocation API to kernelspace drivers. Therefore change the ioctl() ABI by addition of a region_end member in the existing struct fw_cdev_allocate. Userspace and kernelspace APIs work the same way now. There is a small cost to pay by clients though: If client source code is required to compile with older kernel headers too, then any use of the new member fw_cdev_allocate.region_end needs to be enclosed by #ifdef/#endif directives. However, any client program that seriously wants to use address range allocations will require a kernel of cdev ABI version >= 4 at runtime and a linux/firewire-cdev.h header of >= 4 anyway. This is because v4 brings FW_CDEV_EVENT_REQUEST2. The only client program in which build-time compatibility with struct fw_cdev_allocate as found in older kernel headers makes sense is libraw1394. (libraw1394 uses the older broken FW_CDEV_EVENT_REQUEST to implement a makeshift, incorrect transaction responder that does at least work somewhat in many simple scenarios, relying on guesswork by libraw1394 and by libraw1394 based applications. Plus, address range allocation and transaction responder is only one of many features that libraw1394 needs to provide, and these other features need to work with kernel and kernel-headers as old as possible. Any new linux/firewire-cdev.h based client that implements a transaction responder should never attempt to do it like libraw1394; instead it should make a header and kernel of v4 or later a hard requirement.) While we are at it, update the struct fw_cdev_allocate documentation to better reflect the recent fw_cdev_event_request2 ABI addition. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23 05:05:39 -06:00
#define FW_CDEV_VERSION_EVENT_REQUEST2 4
#define FW_CDEV_VERSION_ALLOCATE_REGION_END 4
#define FW_CDEV_VERSION_AUTO_FLUSH_ISO_OVERFLOW 5
struct client {
u32 version;
struct fw_device *device;
spinlock_t lock;
bool in_shutdown;
struct idr resource_idr;
struct list_head event_list;
wait_queue_head_t wait;
wait_queue_head_t tx_flush_wait;
u64 bus_reset_closure;
struct fw_iso_context *iso_context;
u64 iso_closure;
struct fw_iso_buffer buffer;
unsigned long vm_start;
bool buffer_is_mapped;
struct list_head phy_receiver_link;
u64 phy_receiver_closure;
struct list_head link;
struct kref kref;
};
static inline void client_get(struct client *client)
{
kref_get(&client->kref);
}
static void client_release(struct kref *kref)
{
struct client *client = container_of(kref, struct client, kref);
fw_device_put(client->device);
kfree(client);
}
static void client_put(struct client *client)
{
kref_put(&client->kref, client_release);
}
struct client_resource;
typedef void (*client_resource_release_fn_t)(struct client *,
struct client_resource *);
struct client_resource {
client_resource_release_fn_t release;
int handle;
};
struct address_handler_resource {
struct client_resource resource;
struct fw_address_handler handler;
__u64 closure;
struct client *client;
};
struct outbound_transaction_resource {
struct client_resource resource;
struct fw_transaction transaction;
};
struct inbound_transaction_resource {
struct client_resource resource;
firewire: cdev: fix responses to nodes at different card My box has two firewire cards in it: card0 and card1. My application opens /dev/fw0 (card 0) and allocates an address space. The core makes the address space available on both cards. Along comes the remote device, which sends a READ_QUADLET_REQUEST to card1. The request gets passed up to my application, which calls ioctl_send_response(). ioctl_send_response() then calls fw_send_response() with card0, because that's the card it's bound to. Card0's driver drops the response, because it isn't part of a transaction that it has outstanding. So in core-cdev: handle_request(), we need to stash the card of the inbound request in the struct inbound_transaction_resource and use that card to send the response to. The hard part will be refcounting the card correctly so it can't get deallocated while we hold a pointer to it. Here's a trivial patch, which does not do the card refcounting, but at least demonstrates what the problem is. Note that we can't depend on the fact that the core-cdev:client structure holds a card open, because in this case the card it holds open is not the card the request came in on. ..and there's no way for the core to tell cdev "this card is gone, kill any inbound transactions on it", while cdev holds the transaction open until userspace issues a SEND_RESPONSE ioctl, which may be a very, very long time. But when it does, it calls fw_send_response(), which will dereference the card... So how unhappy are we about userspace potentially holding a fw_card open forever? Signed-off-by: Jay Fenlason <fenlason@redhat.com> Reference counting to be addressed in a separate change. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (whitespace)
2010-05-18 12:02:45 -06:00
struct fw_card *card;
struct fw_request *request;
void *data;
size_t length;
};
struct descriptor_resource {
struct client_resource resource;
struct fw_descriptor descriptor;
u32 data[0];
};
struct iso_resource {
struct client_resource resource;
struct client *client;
/* Schedule work and access todo only with client->lock held. */
struct delayed_work work;
enum {ISO_RES_ALLOC, ISO_RES_REALLOC, ISO_RES_DEALLOC,
ISO_RES_ALLOC_ONCE, ISO_RES_DEALLOC_ONCE,} todo;
int generation;
u64 channels;
s32 bandwidth;
struct iso_resource_event *e_alloc, *e_dealloc;
};
static void release_iso_resource(struct client *, struct client_resource *);
static void schedule_iso_resource(struct iso_resource *r, unsigned long delay)
{
client_get(r->client);
if (!queue_delayed_work(fw_workqueue, &r->work, delay))
client_put(r->client);
}
static void schedule_if_iso_resource(struct client_resource *resource)
{
if (resource->release == release_iso_resource)
schedule_iso_resource(container_of(resource,
struct iso_resource, resource), 0);
}
/*
* dequeue_event() just kfree()'s the event, so the event has to be
* the first field in a struct XYZ_event.
*/
struct event {
struct { void *data; size_t size; } v[2];
struct list_head link;
};
struct bus_reset_event {
struct event event;
struct fw_cdev_event_bus_reset reset;
};
struct outbound_transaction_event {
struct event event;
struct client *client;
struct outbound_transaction_resource r;
struct fw_cdev_event_response response;
};
struct inbound_transaction_event {
struct event event;
firewire: cdev: fix ABI for FCP and address range mapping, add fw_cdev_event_request2 The problem: A target-like userspace driver, e.g. AV/C target or SBP-2/3 target, needs to be able to act as responder and requester. In the latter role, it needs to send requests to nods from which it received requests. This is currently impossible because fw_cdev_event_request lacks information about sender node ID. Reported-by: Jay Fenlason <fenlason@redhat.com> Libffado + libraw1394 + firewire-core is currently unable to drive two or more audio devices on the same bus. Reported-by: Arnold Krille <arnold@arnoldarts.de> This is because libffado requires destination node ID of FCP requests and sender node ID of FCP responses to match. It even prohibits libffado from working with a bus on which libraw1394 opens a /dev/fw* as default ioctl device that does not correspond with the audio device. This is because libraw1394 does not receive the sender node ID from the kernel. Moreover, fw_cdev_event_request makes it impossible to tell unicast and broadcast write requests apart. The fix: Add a replacement of struct fw_cdev_event_request request, boringly called struct fw_cdev_event_request2. The new event will be sent to a userspace client instead of the old one if the client claims compatibility with <linux/firewire-cdev.h> ABI version 4 or later. libraw1394 needs to be extended to make use of the new event, in order to properly support libffado and other FCP or address range mapping users who require correct sender node IDs. Further notes: While we are at it, change back the range of possible values of fw_cdev_event_request.tcode to 0x0...0xb like in ABI version <= 3. The preceding change "firewire: expose extended tcode of incoming lock requests to (userspace) drivers" expanded it to 0x0...0x17 which could catch sloppily coded clients by surprise. The extended range of codes is only used in the new fw_cdev_event_request2.tcode. Jay and I also suggested an alternative approach to fix the ABI for incoming requests: Add an FW_CDEV_IOC_GET_REQUEST_INFO ioctl which can be called after reception of an fw_cdev_event_request, before issuing of the closing FW_CDEV_IOC_SEND_RESPONSE ioctl. The new ioctl would reveal the vital information about a request that fw_cdev_event_request lacks. Jay showed an implementation of this approach. The former event approach adds 27 LOC of rather trivial code to core-cdev.c, the ioctl approach 34 LOC, some of which is nontrivial. The ioctl approach would certainly also add more LOC to userspace programs which require the expanded information on inbound requests. This approach is probably only on the lighter-weight side in case of clients that want to be compatible with kernels that lack the new capability, like libraw1394. However, the code to be added to such libraw1394-like clients in case of the event approach is a straight- forward additional switch () case in its event handler. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:53:55 -06:00
union {
struct fw_cdev_event_request request;
struct fw_cdev_event_request2 request2;
} req;
};
struct iso_interrupt_event {
struct event event;
struct fw_cdev_event_iso_interrupt interrupt;
};
struct iso_interrupt_mc_event {
struct event event;
struct fw_cdev_event_iso_interrupt_mc interrupt;
};
struct iso_resource_event {
struct event event;
struct fw_cdev_event_iso_resource iso_resource;
};
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
struct outbound_phy_packet_event {
struct event event;
struct client *client;
struct fw_packet p;
struct fw_cdev_event_phy_packet phy_packet;
};
struct inbound_phy_packet_event {
struct event event;
struct fw_cdev_event_phy_packet phy_packet;
};
#ifdef CONFIG_COMPAT
static void __user *u64_to_uptr(u64 value)
{
if (in_compat_syscall())
return compat_ptr(value);
else
return (void __user *)(unsigned long)value;
}
static u64 uptr_to_u64(void __user *ptr)
{
if (in_compat_syscall())
return ptr_to_compat(ptr);
else
return (u64)(unsigned long)ptr;
}
#else
static inline void __user *u64_to_uptr(u64 value)
{
return (void __user *)(unsigned long)value;
}
static inline u64 uptr_to_u64(void __user *ptr)
{
return (u64)(unsigned long)ptr;
}
#endif /* CONFIG_COMPAT */
static int fw_device_op_open(struct inode *inode, struct file *file)
{
struct fw_device *device;
struct client *client;
device = fw_device_get_by_devt(inode->i_rdev);
if (device == NULL)
return -ENODEV;
if (fw_device_is_shutdown(device)) {
fw_device_put(device);
return -ENODEV;
}
client = kzalloc(sizeof(*client), GFP_KERNEL);
if (client == NULL) {
fw_device_put(device);
return -ENOMEM;
}
client->device = device;
spin_lock_init(&client->lock);
idr_init(&client->resource_idr);
INIT_LIST_HEAD(&client->event_list);
init_waitqueue_head(&client->wait);
init_waitqueue_head(&client->tx_flush_wait);
INIT_LIST_HEAD(&client->phy_receiver_link);
firewire: cdev: prevent race between first get_info ioctl and bus reset event queuing Between open(2) of a /dev/fw* and the first FW_CDEV_IOC_GET_INFO ioctl(2) on it, the kernel already queues FW_CDEV_EVENT_BUS_RESET events to be read(2) by the client. The get_info ioctl is practically always issued right away after open, hence this condition only occurs if the client opens during a bus reset, especially during a rapid series of bus resets. The problem with this condition is twofold: - These bus reset events carry the (as yet undocumented) @closure value of 0. But it is not the kernel's place to choose closures; they are privat to the client. E.g., this 0 value forced from the kernel makes it unsafe for clients to dereference it as a pointer to a closure object without NULL pointer check. - It is impossible for clients to determine the relative order of bus reset events from get_info ioctl(2) versus those from read(2), except in one way: By comparison of closure values. Again, such a procedure imposes complexity on clients and reduces freedom in use of the bus reset closure. So, change the ABI to suppress queuing of bus reset events before the first FW_CDEV_IOC_GET_INFO ioctl was issued by the client. Note, this ABI change cannot be version-controlled. The kernel cannot distinguish old from new clients before the first FW_CDEV_IOC_GET_INFO ioctl. We will try to back-merge this change into currently maintained stable/ longterm series, and we only document the new behaviour. The old behavior is now considered a kernel bug, which it basically is. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Cc: <stable@kernel.org>
2011-07-09 08:43:22 -06:00
INIT_LIST_HEAD(&client->link);
kref_init(&client->kref);
file->private_data = client;
return nonseekable_open(inode, file);
}
static void queue_event(struct client *client, struct event *event,
void *data0, size_t size0, void *data1, size_t size1)
{
unsigned long flags;
event->v[0].data = data0;
event->v[0].size = size0;
event->v[1].data = data1;
event->v[1].size = size1;
spin_lock_irqsave(&client->lock, flags);
if (client->in_shutdown)
kfree(event);
else
list_add_tail(&event->link, &client->event_list);
spin_unlock_irqrestore(&client->lock, flags);
wake_up_interruptible(&client->wait);
}
static int dequeue_event(struct client *client,
char __user *buffer, size_t count)
{
struct event *event;
size_t size, total;
int i, ret;
ret = wait_event_interruptible(client->wait,
!list_empty(&client->event_list) ||
fw_device_is_shutdown(client->device));
if (ret < 0)
return ret;
if (list_empty(&client->event_list) &&
fw_device_is_shutdown(client->device))
return -ENODEV;
spin_lock_irq(&client->lock);
event = list_first_entry(&client->event_list, struct event, link);
list_del(&event->link);
spin_unlock_irq(&client->lock);
total = 0;
for (i = 0; i < ARRAY_SIZE(event->v) && total < count; i++) {
size = min(event->v[i].size, count - total);
if (copy_to_user(buffer + total, event->v[i].data, size)) {
ret = -EFAULT;
goto out;
}
total += size;
}
ret = total;
out:
kfree(event);
return ret;
}
static ssize_t fw_device_op_read(struct file *file, char __user *buffer,
size_t count, loff_t *offset)
{
struct client *client = file->private_data;
return dequeue_event(client, buffer, count);
}
static void fill_bus_reset_event(struct fw_cdev_event_bus_reset *event,
struct client *client)
{
struct fw_card *card = client->device->card;
spin_lock_irq(&card->lock);
event->closure = client->bus_reset_closure;
event->type = FW_CDEV_EVENT_BUS_RESET;
event->generation = client->device->generation;
event->node_id = client->device->node_id;
event->local_node_id = card->local_node->node_id;
event->bm_node_id = card->bm_node_id;
event->irm_node_id = card->irm_node->node_id;
event->root_node_id = card->root_node->node_id;
spin_unlock_irq(&card->lock);
}
static void for_each_client(struct fw_device *device,
void (*callback)(struct client *client))
{
struct client *c;
mutex_lock(&device->client_list_mutex);
list_for_each_entry(c, &device->client_list, link)
callback(c);
mutex_unlock(&device->client_list_mutex);
}
static int schedule_reallocations(int id, void *p, void *data)
{
schedule_if_iso_resource(p);
return 0;
}
static void queue_bus_reset_event(struct client *client)
{
struct bus_reset_event *e;
e = kzalloc(sizeof(*e), GFP_KERNEL);
if (e == NULL)
return;
fill_bus_reset_event(&e->reset, client);
queue_event(client, &e->event,
&e->reset, sizeof(e->reset), NULL, 0);
spin_lock_irq(&client->lock);
idr_for_each(&client->resource_idr, schedule_reallocations, client);
spin_unlock_irq(&client->lock);
}
void fw_device_cdev_update(struct fw_device *device)
{
for_each_client(device, queue_bus_reset_event);
}
static void wake_up_client(struct client *client)
{
wake_up_interruptible(&client->wait);
}
void fw_device_cdev_remove(struct fw_device *device)
{
for_each_client(device, wake_up_client);
}
union ioctl_arg {
struct fw_cdev_get_info get_info;
struct fw_cdev_send_request send_request;
struct fw_cdev_allocate allocate;
struct fw_cdev_deallocate deallocate;
struct fw_cdev_send_response send_response;
struct fw_cdev_initiate_bus_reset initiate_bus_reset;
struct fw_cdev_add_descriptor add_descriptor;
struct fw_cdev_remove_descriptor remove_descriptor;
struct fw_cdev_create_iso_context create_iso_context;
struct fw_cdev_queue_iso queue_iso;
struct fw_cdev_start_iso start_iso;
struct fw_cdev_stop_iso stop_iso;
struct fw_cdev_get_cycle_timer get_cycle_timer;
struct fw_cdev_allocate_iso_resource allocate_iso_resource;
struct fw_cdev_send_stream_packet send_stream_packet;
struct fw_cdev_get_cycle_timer2 get_cycle_timer2;
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
struct fw_cdev_send_phy_packet send_phy_packet;
struct fw_cdev_receive_phy_packets receive_phy_packets;
struct fw_cdev_set_iso_channels set_iso_channels;
struct fw_cdev_flush_iso flush_iso;
};
static int ioctl_get_info(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_get_info *a = &arg->get_info;
struct fw_cdev_event_bus_reset bus_reset;
unsigned long ret = 0;
client->version = a->version;
a->version = FW_CDEV_KERNEL_VERSION;
a->card = client->device->card->index;
down_read(&fw_device_rwsem);
if (a->rom != 0) {
size_t want = a->rom_length;
size_t have = client->device->config_rom_length * 4;
ret = copy_to_user(u64_to_uptr(a->rom),
client->device->config_rom, min(want, have));
}
a->rom_length = client->device->config_rom_length * 4;
up_read(&fw_device_rwsem);
if (ret != 0)
return -EFAULT;
firewire: cdev: prevent race between first get_info ioctl and bus reset event queuing Between open(2) of a /dev/fw* and the first FW_CDEV_IOC_GET_INFO ioctl(2) on it, the kernel already queues FW_CDEV_EVENT_BUS_RESET events to be read(2) by the client. The get_info ioctl is practically always issued right away after open, hence this condition only occurs if the client opens during a bus reset, especially during a rapid series of bus resets. The problem with this condition is twofold: - These bus reset events carry the (as yet undocumented) @closure value of 0. But it is not the kernel's place to choose closures; they are privat to the client. E.g., this 0 value forced from the kernel makes it unsafe for clients to dereference it as a pointer to a closure object without NULL pointer check. - It is impossible for clients to determine the relative order of bus reset events from get_info ioctl(2) versus those from read(2), except in one way: By comparison of closure values. Again, such a procedure imposes complexity on clients and reduces freedom in use of the bus reset closure. So, change the ABI to suppress queuing of bus reset events before the first FW_CDEV_IOC_GET_INFO ioctl was issued by the client. Note, this ABI change cannot be version-controlled. The kernel cannot distinguish old from new clients before the first FW_CDEV_IOC_GET_INFO ioctl. We will try to back-merge this change into currently maintained stable/ longterm series, and we only document the new behaviour. The old behavior is now considered a kernel bug, which it basically is. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Cc: <stable@kernel.org>
2011-07-09 08:43:22 -06:00
mutex_lock(&client->device->client_list_mutex);
client->bus_reset_closure = a->bus_reset_closure;
if (a->bus_reset != 0) {
fill_bus_reset_event(&bus_reset, client);
/* unaligned size of bus_reset is 36 bytes */
ret = copy_to_user(u64_to_uptr(a->bus_reset), &bus_reset, 36);
}
firewire: cdev: prevent race between first get_info ioctl and bus reset event queuing Between open(2) of a /dev/fw* and the first FW_CDEV_IOC_GET_INFO ioctl(2) on it, the kernel already queues FW_CDEV_EVENT_BUS_RESET events to be read(2) by the client. The get_info ioctl is practically always issued right away after open, hence this condition only occurs if the client opens during a bus reset, especially during a rapid series of bus resets. The problem with this condition is twofold: - These bus reset events carry the (as yet undocumented) @closure value of 0. But it is not the kernel's place to choose closures; they are privat to the client. E.g., this 0 value forced from the kernel makes it unsafe for clients to dereference it as a pointer to a closure object without NULL pointer check. - It is impossible for clients to determine the relative order of bus reset events from get_info ioctl(2) versus those from read(2), except in one way: By comparison of closure values. Again, such a procedure imposes complexity on clients and reduces freedom in use of the bus reset closure. So, change the ABI to suppress queuing of bus reset events before the first FW_CDEV_IOC_GET_INFO ioctl was issued by the client. Note, this ABI change cannot be version-controlled. The kernel cannot distinguish old from new clients before the first FW_CDEV_IOC_GET_INFO ioctl. We will try to back-merge this change into currently maintained stable/ longterm series, and we only document the new behaviour. The old behavior is now considered a kernel bug, which it basically is. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Cc: <stable@kernel.org>
2011-07-09 08:43:22 -06:00
if (ret == 0 && list_empty(&client->link))
list_add_tail(&client->link, &client->device->client_list);
firewire: cdev: prevent race between first get_info ioctl and bus reset event queuing Between open(2) of a /dev/fw* and the first FW_CDEV_IOC_GET_INFO ioctl(2) on it, the kernel already queues FW_CDEV_EVENT_BUS_RESET events to be read(2) by the client. The get_info ioctl is practically always issued right away after open, hence this condition only occurs if the client opens during a bus reset, especially during a rapid series of bus resets. The problem with this condition is twofold: - These bus reset events carry the (as yet undocumented) @closure value of 0. But it is not the kernel's place to choose closures; they are privat to the client. E.g., this 0 value forced from the kernel makes it unsafe for clients to dereference it as a pointer to a closure object without NULL pointer check. - It is impossible for clients to determine the relative order of bus reset events from get_info ioctl(2) versus those from read(2), except in one way: By comparison of closure values. Again, such a procedure imposes complexity on clients and reduces freedom in use of the bus reset closure. So, change the ABI to suppress queuing of bus reset events before the first FW_CDEV_IOC_GET_INFO ioctl was issued by the client. Note, this ABI change cannot be version-controlled. The kernel cannot distinguish old from new clients before the first FW_CDEV_IOC_GET_INFO ioctl. We will try to back-merge this change into currently maintained stable/ longterm series, and we only document the new behaviour. The old behavior is now considered a kernel bug, which it basically is. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Cc: <stable@kernel.org>
2011-07-09 08:43:22 -06:00
mutex_unlock(&client->device->client_list_mutex);
return ret ? -EFAULT : 0;
}
static int add_client_resource(struct client *client,
struct client_resource *resource, gfp_t gfp_mask)
{
mm, page_alloc: distinguish between being unable to sleep, unwilling to sleep and avoiding waking kswapd __GFP_WAIT has been used to identify atomic context in callers that hold spinlocks or are in interrupts. They are expected to be high priority and have access one of two watermarks lower than "min" which can be referred to as the "atomic reserve". __GFP_HIGH users get access to the first lower watermark and can be called the "high priority reserve". Over time, callers had a requirement to not block when fallback options were available. Some have abused __GFP_WAIT leading to a situation where an optimisitic allocation with a fallback option can access atomic reserves. This patch uses __GFP_ATOMIC to identify callers that are truely atomic, cannot sleep and have no alternative. High priority users continue to use __GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify callers that want to wake kswapd for background reclaim. __GFP_WAIT is redefined as a caller that is willing to enter direct reclaim and wake kswapd for background reclaim. This patch then converts a number of sites o __GFP_ATOMIC is used by callers that are high priority and have memory pools for those requests. GFP_ATOMIC uses this flag. o Callers that have a limited mempool to guarantee forward progress clear __GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall into this category where kswapd will still be woken but atomic reserves are not used as there is a one-entry mempool to guarantee progress. o Callers that are checking if they are non-blocking should use the helper gfpflags_allow_blocking() where possible. This is because checking for __GFP_WAIT as was done historically now can trigger false positives. Some exceptions like dm-crypt.c exist where the code intent is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to flag manipulations. o Callers that built their own GFP flags instead of starting with GFP_KERNEL and friends now also need to specify __GFP_KSWAPD_RECLAIM. The first key hazard to watch out for is callers that removed __GFP_WAIT and was depending on access to atomic reserves for inconspicuous reasons. In some cases it may be appropriate for them to use __GFP_HIGH. The second key hazard is callers that assembled their own combination of GFP flags instead of starting with something like GFP_KERNEL. They may now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless if it's missed in most cases as other activity will wake kswapd. Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Vitaly Wool <vitalywool@gmail.com> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 17:28:21 -07:00
bool preload = gfpflags_allow_blocking(gfp_mask);
unsigned long flags;
int ret;
if (preload)
idr_preload(gfp_mask);
spin_lock_irqsave(&client->lock, flags);
if (client->in_shutdown)
ret = -ECANCELED;
else
ret = idr_alloc(&client->resource_idr, resource, 0, 0,
GFP_NOWAIT);
if (ret >= 0) {
resource->handle = ret;
client_get(client);
schedule_if_iso_resource(resource);
}
spin_unlock_irqrestore(&client->lock, flags);
if (preload)
idr_preload_end();
return ret < 0 ? ret : 0;
}
static int release_client_resource(struct client *client, u32 handle,
client_resource_release_fn_t release,
struct client_resource **return_resource)
{
struct client_resource *resource;
spin_lock_irq(&client->lock);
if (client->in_shutdown)
resource = NULL;
else
resource = idr_find(&client->resource_idr, handle);
if (resource && resource->release == release)
idr_remove(&client->resource_idr, handle);
spin_unlock_irq(&client->lock);
if (!(resource && resource->release == release))
return -EINVAL;
if (return_resource)
*return_resource = resource;
else
resource->release(client, resource);
client_put(client);
return 0;
}
static void release_transaction(struct client *client,
struct client_resource *resource)
{
}
static void complete_transaction(struct fw_card *card, int rcode,
void *payload, size_t length, void *data)
{
struct outbound_transaction_event *e = data;
struct fw_cdev_event_response *rsp = &e->response;
struct client *client = e->client;
unsigned long flags;
if (length < rsp->length)
rsp->length = length;
if (rcode == RCODE_COMPLETE)
memcpy(rsp->data, payload, rsp->length);
spin_lock_irqsave(&client->lock, flags);
idr_remove(&client->resource_idr, e->r.resource.handle);
if (client->in_shutdown)
wake_up(&client->tx_flush_wait);
spin_unlock_irqrestore(&client->lock, flags);
rsp->type = FW_CDEV_EVENT_RESPONSE;
rsp->rcode = rcode;
/*
* In the case that sizeof(*rsp) doesn't align with the position of the
* data, and the read is short, preserve an extra copy of the data
* to stay compatible with a pre-2.6.27 bug. Since the bug is harmless
* for short reads and some apps depended on it, this is both safe
* and prudent for compatibility.
*/
if (rsp->length <= sizeof(*rsp) - offsetof(typeof(*rsp), data))
queue_event(client, &e->event, rsp, sizeof(*rsp),
rsp->data, rsp->length);
else
queue_event(client, &e->event, rsp, sizeof(*rsp) + rsp->length,
NULL, 0);
/* Drop the idr's reference */
client_put(client);
}
static int init_request(struct client *client,
struct fw_cdev_send_request *request,
int destination_id, int speed)
{
struct outbound_transaction_event *e;
int ret;
if (request->tcode != TCODE_STREAM_DATA &&
(request->length > 4096 || request->length > 512 << speed))
return -EIO;
if (request->tcode == TCODE_WRITE_QUADLET_REQUEST &&
request->length < 4)
return -EINVAL;
e = kmalloc(sizeof(*e) + request->length, GFP_KERNEL);
if (e == NULL)
return -ENOMEM;
e->client = client;
e->response.length = request->length;
e->response.closure = request->closure;
if (request->data &&
copy_from_user(e->response.data,
u64_to_uptr(request->data), request->length)) {
ret = -EFAULT;
goto failed;
}
e->r.resource.release = release_transaction;
ret = add_client_resource(client, &e->r.resource, GFP_KERNEL);
if (ret < 0)
goto failed;
fw_send_request(client->device->card, &e->r.transaction,
request->tcode, destination_id, request->generation,
speed, request->offset, e->response.data,
request->length, complete_transaction, e);
return 0;
failed:
kfree(e);
return ret;
}
static int ioctl_send_request(struct client *client, union ioctl_arg *arg)
{
switch (arg->send_request.tcode) {
case TCODE_WRITE_QUADLET_REQUEST:
case TCODE_WRITE_BLOCK_REQUEST:
case TCODE_READ_QUADLET_REQUEST:
case TCODE_READ_BLOCK_REQUEST:
case TCODE_LOCK_MASK_SWAP:
case TCODE_LOCK_COMPARE_SWAP:
case TCODE_LOCK_FETCH_ADD:
case TCODE_LOCK_LITTLE_ADD:
case TCODE_LOCK_BOUNDED_ADD:
case TCODE_LOCK_WRAP_ADD:
case TCODE_LOCK_VENDOR_DEPENDENT:
break;
default:
return -EINVAL;
}
return init_request(client, &arg->send_request, client->device->node_id,
client->device->max_speed);
}
static inline bool is_fcp_request(struct fw_request *request)
{
return request == NULL;
}
static void release_request(struct client *client,
struct client_resource *resource)
{
struct inbound_transaction_resource *r = container_of(resource,
struct inbound_transaction_resource, resource);
if (is_fcp_request(r->request))
kfree(r->data);
else
firewire: cdev: fix responses to nodes at different card My box has two firewire cards in it: card0 and card1. My application opens /dev/fw0 (card 0) and allocates an address space. The core makes the address space available on both cards. Along comes the remote device, which sends a READ_QUADLET_REQUEST to card1. The request gets passed up to my application, which calls ioctl_send_response(). ioctl_send_response() then calls fw_send_response() with card0, because that's the card it's bound to. Card0's driver drops the response, because it isn't part of a transaction that it has outstanding. So in core-cdev: handle_request(), we need to stash the card of the inbound request in the struct inbound_transaction_resource and use that card to send the response to. The hard part will be refcounting the card correctly so it can't get deallocated while we hold a pointer to it. Here's a trivial patch, which does not do the card refcounting, but at least demonstrates what the problem is. Note that we can't depend on the fact that the core-cdev:client structure holds a card open, because in this case the card it holds open is not the card the request came in on. ..and there's no way for the core to tell cdev "this card is gone, kill any inbound transactions on it", while cdev holds the transaction open until userspace issues a SEND_RESPONSE ioctl, which may be a very, very long time. But when it does, it calls fw_send_response(), which will dereference the card... So how unhappy are we about userspace potentially holding a fw_card open forever? Signed-off-by: Jay Fenlason <fenlason@redhat.com> Reference counting to be addressed in a separate change. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (whitespace)
2010-05-18 12:02:45 -06:00
fw_send_response(r->card, r->request, RCODE_CONFLICT_ERROR);
firewire: cdev: count references of cards during inbound transactions If a request comes in to an address range managed by a userspace driver i.e. <linux/firewire-cdev.h> client, the card instance of request and response may differ from the card instance of the client device. Therefore we need to take a reference of the card until the response was sent. I thought about putting the reference counting into core-transaction.c, but the various high-level drivers besides cdev clients (firewire-net, firewire-sbp2, firedtv) use the card pointer in their fw_address_handler address_callback method only to look up devices of which they already hold the necessary references. So this seems to be a specific firewire-cdev issue which is better addressed locally. We do not need the reference - in case of FCP_REQUEST or FCP_RESPONSE requests because then the firewire-core will send the split transaction response for us already in the context of the request handler, - if it is the same card as the client device's because we hold a card reference indirectly via teh client->device reference. To keep things simple, we take the reference nevertheless. Jay Fenlason wrote: > there's no way for the core to tell cdev "this card is gone, > kill any inbound transactions on it", while cdev holds the transaction > open until userspace issues a SEND_RESPONSE ioctl, which may be a very, > very long time. But when it does, it calls fw_send_response(), which > will dereference the card... > > So how unhappy are we about userspace potentially holding a fw_card > open forever? While termination of inbound transcations at card removal could be implemented, it is IMO not worth the effort. Currently, the effect of holding a reference of a card that has been removed is to block the process that called the pci_remove of the card. This is - either a user process ran by root. Root can find and kill processes that have /dev/fw* open, if desired. - a kernel thread (which one?) in case of hot removal of a PCCard or ExpressCard. The latter case could be a problem indeed. firewire-core's card shutdown and card release should probably be improved not to block in shutdown, just to defer freeing of memory until release. This is not a new problem though; the same already always happens with the client->device->card without the need of inbound transactions or other special conditions involved, other than the client not closing the file. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:52:27 -06:00
fw_card_put(r->card);
kfree(r);
}
static void handle_request(struct fw_card *card, struct fw_request *request,
int tcode, int destination, int source,
int generation, unsigned long long offset,
void *payload, size_t length, void *callback_data)
{
struct address_handler_resource *handler = callback_data;
struct inbound_transaction_resource *r;
struct inbound_transaction_event *e;
firewire: cdev: fix ABI for FCP and address range mapping, add fw_cdev_event_request2 The problem: A target-like userspace driver, e.g. AV/C target or SBP-2/3 target, needs to be able to act as responder and requester. In the latter role, it needs to send requests to nods from which it received requests. This is currently impossible because fw_cdev_event_request lacks information about sender node ID. Reported-by: Jay Fenlason <fenlason@redhat.com> Libffado + libraw1394 + firewire-core is currently unable to drive two or more audio devices on the same bus. Reported-by: Arnold Krille <arnold@arnoldarts.de> This is because libffado requires destination node ID of FCP requests and sender node ID of FCP responses to match. It even prohibits libffado from working with a bus on which libraw1394 opens a /dev/fw* as default ioctl device that does not correspond with the audio device. This is because libraw1394 does not receive the sender node ID from the kernel. Moreover, fw_cdev_event_request makes it impossible to tell unicast and broadcast write requests apart. The fix: Add a replacement of struct fw_cdev_event_request request, boringly called struct fw_cdev_event_request2. The new event will be sent to a userspace client instead of the old one if the client claims compatibility with <linux/firewire-cdev.h> ABI version 4 or later. libraw1394 needs to be extended to make use of the new event, in order to properly support libffado and other FCP or address range mapping users who require correct sender node IDs. Further notes: While we are at it, change back the range of possible values of fw_cdev_event_request.tcode to 0x0...0xb like in ABI version <= 3. The preceding change "firewire: expose extended tcode of incoming lock requests to (userspace) drivers" expanded it to 0x0...0x17 which could catch sloppily coded clients by surprise. The extended range of codes is only used in the new fw_cdev_event_request2.tcode. Jay and I also suggested an alternative approach to fix the ABI for incoming requests: Add an FW_CDEV_IOC_GET_REQUEST_INFO ioctl which can be called after reception of an fw_cdev_event_request, before issuing of the closing FW_CDEV_IOC_SEND_RESPONSE ioctl. The new ioctl would reveal the vital information about a request that fw_cdev_event_request lacks. Jay showed an implementation of this approach. The former event approach adds 27 LOC of rather trivial code to core-cdev.c, the ioctl approach 34 LOC, some of which is nontrivial. The ioctl approach would certainly also add more LOC to userspace programs which require the expanded information on inbound requests. This approach is probably only on the lighter-weight side in case of clients that want to be compatible with kernels that lack the new capability, like libraw1394. However, the code to be added to such libraw1394-like clients in case of the event approach is a straight- forward additional switch () case in its event handler. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:53:55 -06:00
size_t event_size0;
void *fcp_frame = NULL;
int ret;
firewire: cdev: count references of cards during inbound transactions If a request comes in to an address range managed by a userspace driver i.e. <linux/firewire-cdev.h> client, the card instance of request and response may differ from the card instance of the client device. Therefore we need to take a reference of the card until the response was sent. I thought about putting the reference counting into core-transaction.c, but the various high-level drivers besides cdev clients (firewire-net, firewire-sbp2, firedtv) use the card pointer in their fw_address_handler address_callback method only to look up devices of which they already hold the necessary references. So this seems to be a specific firewire-cdev issue which is better addressed locally. We do not need the reference - in case of FCP_REQUEST or FCP_RESPONSE requests because then the firewire-core will send the split transaction response for us already in the context of the request handler, - if it is the same card as the client device's because we hold a card reference indirectly via teh client->device reference. To keep things simple, we take the reference nevertheless. Jay Fenlason wrote: > there's no way for the core to tell cdev "this card is gone, > kill any inbound transactions on it", while cdev holds the transaction > open until userspace issues a SEND_RESPONSE ioctl, which may be a very, > very long time. But when it does, it calls fw_send_response(), which > will dereference the card... > > So how unhappy are we about userspace potentially holding a fw_card > open forever? While termination of inbound transcations at card removal could be implemented, it is IMO not worth the effort. Currently, the effect of holding a reference of a card that has been removed is to block the process that called the pci_remove of the card. This is - either a user process ran by root. Root can find and kill processes that have /dev/fw* open, if desired. - a kernel thread (which one?) in case of hot removal of a PCCard or ExpressCard. The latter case could be a problem indeed. firewire-core's card shutdown and card release should probably be improved not to block in shutdown, just to defer freeing of memory until release. This is not a new problem though; the same already always happens with the client->device->card without the need of inbound transactions or other special conditions involved, other than the client not closing the file. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:52:27 -06:00
/* card may be different from handler->client->device->card */
fw_card_get(card);
r = kmalloc(sizeof(*r), GFP_ATOMIC);
e = kmalloc(sizeof(*e), GFP_ATOMIC);
if (r == NULL || e == NULL)
goto failed;
firewire: cdev: fix responses to nodes at different card My box has two firewire cards in it: card0 and card1. My application opens /dev/fw0 (card 0) and allocates an address space. The core makes the address space available on both cards. Along comes the remote device, which sends a READ_QUADLET_REQUEST to card1. The request gets passed up to my application, which calls ioctl_send_response(). ioctl_send_response() then calls fw_send_response() with card0, because that's the card it's bound to. Card0's driver drops the response, because it isn't part of a transaction that it has outstanding. So in core-cdev: handle_request(), we need to stash the card of the inbound request in the struct inbound_transaction_resource and use that card to send the response to. The hard part will be refcounting the card correctly so it can't get deallocated while we hold a pointer to it. Here's a trivial patch, which does not do the card refcounting, but at least demonstrates what the problem is. Note that we can't depend on the fact that the core-cdev:client structure holds a card open, because in this case the card it holds open is not the card the request came in on. ..and there's no way for the core to tell cdev "this card is gone, kill any inbound transactions on it", while cdev holds the transaction open until userspace issues a SEND_RESPONSE ioctl, which may be a very, very long time. But when it does, it calls fw_send_response(), which will dereference the card... So how unhappy are we about userspace potentially holding a fw_card open forever? Signed-off-by: Jay Fenlason <fenlason@redhat.com> Reference counting to be addressed in a separate change. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (whitespace)
2010-05-18 12:02:45 -06:00
r->card = card;
r->request = request;
r->data = payload;
r->length = length;
if (is_fcp_request(request)) {
/*
* FIXME: Let core-transaction.c manage a
* single reference-counted copy?
*/
fcp_frame = kmemdup(payload, length, GFP_ATOMIC);
if (fcp_frame == NULL)
goto failed;
r->data = fcp_frame;
}
r->resource.release = release_request;
ret = add_client_resource(handler->client, &r->resource, GFP_ATOMIC);
if (ret < 0)
goto failed;
firewire: cdev: fix ABI for FCP and address range mapping, add fw_cdev_event_request2 The problem: A target-like userspace driver, e.g. AV/C target or SBP-2/3 target, needs to be able to act as responder and requester. In the latter role, it needs to send requests to nods from which it received requests. This is currently impossible because fw_cdev_event_request lacks information about sender node ID. Reported-by: Jay Fenlason <fenlason@redhat.com> Libffado + libraw1394 + firewire-core is currently unable to drive two or more audio devices on the same bus. Reported-by: Arnold Krille <arnold@arnoldarts.de> This is because libffado requires destination node ID of FCP requests and sender node ID of FCP responses to match. It even prohibits libffado from working with a bus on which libraw1394 opens a /dev/fw* as default ioctl device that does not correspond with the audio device. This is because libraw1394 does not receive the sender node ID from the kernel. Moreover, fw_cdev_event_request makes it impossible to tell unicast and broadcast write requests apart. The fix: Add a replacement of struct fw_cdev_event_request request, boringly called struct fw_cdev_event_request2. The new event will be sent to a userspace client instead of the old one if the client claims compatibility with <linux/firewire-cdev.h> ABI version 4 or later. libraw1394 needs to be extended to make use of the new event, in order to properly support libffado and other FCP or address range mapping users who require correct sender node IDs. Further notes: While we are at it, change back the range of possible values of fw_cdev_event_request.tcode to 0x0...0xb like in ABI version <= 3. The preceding change "firewire: expose extended tcode of incoming lock requests to (userspace) drivers" expanded it to 0x0...0x17 which could catch sloppily coded clients by surprise. The extended range of codes is only used in the new fw_cdev_event_request2.tcode. Jay and I also suggested an alternative approach to fix the ABI for incoming requests: Add an FW_CDEV_IOC_GET_REQUEST_INFO ioctl which can be called after reception of an fw_cdev_event_request, before issuing of the closing FW_CDEV_IOC_SEND_RESPONSE ioctl. The new ioctl would reveal the vital information about a request that fw_cdev_event_request lacks. Jay showed an implementation of this approach. The former event approach adds 27 LOC of rather trivial code to core-cdev.c, the ioctl approach 34 LOC, some of which is nontrivial. The ioctl approach would certainly also add more LOC to userspace programs which require the expanded information on inbound requests. This approach is probably only on the lighter-weight side in case of clients that want to be compatible with kernels that lack the new capability, like libraw1394. However, the code to be added to such libraw1394-like clients in case of the event approach is a straight- forward additional switch () case in its event handler. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:53:55 -06:00
if (handler->client->version < FW_CDEV_VERSION_EVENT_REQUEST2) {
struct fw_cdev_event_request *req = &e->req.request;
if (tcode & 0x10)
tcode = TCODE_LOCK_REQUEST;
req->type = FW_CDEV_EVENT_REQUEST;
req->tcode = tcode;
req->offset = offset;
req->length = length;
req->handle = r->resource.handle;
req->closure = handler->closure;
event_size0 = sizeof(*req);
} else {
struct fw_cdev_event_request2 *req = &e->req.request2;
req->type = FW_CDEV_EVENT_REQUEST2;
req->tcode = tcode;
req->offset = offset;
req->source_node_id = source;
req->destination_node_id = destination;
req->card = card->index;
req->generation = generation;
req->length = length;
req->handle = r->resource.handle;
req->closure = handler->closure;
event_size0 = sizeof(*req);
}
queue_event(handler->client, &e->event,
firewire: cdev: fix ABI for FCP and address range mapping, add fw_cdev_event_request2 The problem: A target-like userspace driver, e.g. AV/C target or SBP-2/3 target, needs to be able to act as responder and requester. In the latter role, it needs to send requests to nods from which it received requests. This is currently impossible because fw_cdev_event_request lacks information about sender node ID. Reported-by: Jay Fenlason <fenlason@redhat.com> Libffado + libraw1394 + firewire-core is currently unable to drive two or more audio devices on the same bus. Reported-by: Arnold Krille <arnold@arnoldarts.de> This is because libffado requires destination node ID of FCP requests and sender node ID of FCP responses to match. It even prohibits libffado from working with a bus on which libraw1394 opens a /dev/fw* as default ioctl device that does not correspond with the audio device. This is because libraw1394 does not receive the sender node ID from the kernel. Moreover, fw_cdev_event_request makes it impossible to tell unicast and broadcast write requests apart. The fix: Add a replacement of struct fw_cdev_event_request request, boringly called struct fw_cdev_event_request2. The new event will be sent to a userspace client instead of the old one if the client claims compatibility with <linux/firewire-cdev.h> ABI version 4 or later. libraw1394 needs to be extended to make use of the new event, in order to properly support libffado and other FCP or address range mapping users who require correct sender node IDs. Further notes: While we are at it, change back the range of possible values of fw_cdev_event_request.tcode to 0x0...0xb like in ABI version <= 3. The preceding change "firewire: expose extended tcode of incoming lock requests to (userspace) drivers" expanded it to 0x0...0x17 which could catch sloppily coded clients by surprise. The extended range of codes is only used in the new fw_cdev_event_request2.tcode. Jay and I also suggested an alternative approach to fix the ABI for incoming requests: Add an FW_CDEV_IOC_GET_REQUEST_INFO ioctl which can be called after reception of an fw_cdev_event_request, before issuing of the closing FW_CDEV_IOC_SEND_RESPONSE ioctl. The new ioctl would reveal the vital information about a request that fw_cdev_event_request lacks. Jay showed an implementation of this approach. The former event approach adds 27 LOC of rather trivial code to core-cdev.c, the ioctl approach 34 LOC, some of which is nontrivial. The ioctl approach would certainly also add more LOC to userspace programs which require the expanded information on inbound requests. This approach is probably only on the lighter-weight side in case of clients that want to be compatible with kernels that lack the new capability, like libraw1394. However, the code to be added to such libraw1394-like clients in case of the event approach is a straight- forward additional switch () case in its event handler. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:53:55 -06:00
&e->req, event_size0, r->data, length);
return;
failed:
kfree(r);
kfree(e);
kfree(fcp_frame);
if (!is_fcp_request(request))
fw_send_response(card, request, RCODE_CONFLICT_ERROR);
firewire: cdev: count references of cards during inbound transactions If a request comes in to an address range managed by a userspace driver i.e. <linux/firewire-cdev.h> client, the card instance of request and response may differ from the card instance of the client device. Therefore we need to take a reference of the card until the response was sent. I thought about putting the reference counting into core-transaction.c, but the various high-level drivers besides cdev clients (firewire-net, firewire-sbp2, firedtv) use the card pointer in their fw_address_handler address_callback method only to look up devices of which they already hold the necessary references. So this seems to be a specific firewire-cdev issue which is better addressed locally. We do not need the reference - in case of FCP_REQUEST or FCP_RESPONSE requests because then the firewire-core will send the split transaction response for us already in the context of the request handler, - if it is the same card as the client device's because we hold a card reference indirectly via teh client->device reference. To keep things simple, we take the reference nevertheless. Jay Fenlason wrote: > there's no way for the core to tell cdev "this card is gone, > kill any inbound transactions on it", while cdev holds the transaction > open until userspace issues a SEND_RESPONSE ioctl, which may be a very, > very long time. But when it does, it calls fw_send_response(), which > will dereference the card... > > So how unhappy are we about userspace potentially holding a fw_card > open forever? While termination of inbound transcations at card removal could be implemented, it is IMO not worth the effort. Currently, the effect of holding a reference of a card that has been removed is to block the process that called the pci_remove of the card. This is - either a user process ran by root. Root can find and kill processes that have /dev/fw* open, if desired. - a kernel thread (which one?) in case of hot removal of a PCCard or ExpressCard. The latter case could be a problem indeed. firewire-core's card shutdown and card release should probably be improved not to block in shutdown, just to defer freeing of memory until release. This is not a new problem though; the same already always happens with the client->device->card without the need of inbound transactions or other special conditions involved, other than the client not closing the file. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:52:27 -06:00
fw_card_put(card);
}
static void release_address_handler(struct client *client,
struct client_resource *resource)
{
struct address_handler_resource *r =
container_of(resource, struct address_handler_resource, resource);
fw_core_remove_address_handler(&r->handler);
kfree(r);
}
static int ioctl_allocate(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_allocate *a = &arg->allocate;
struct address_handler_resource *r;
struct fw_address_region region;
int ret;
r = kmalloc(sizeof(*r), GFP_KERNEL);
if (r == NULL)
return -ENOMEM;
region.start = a->offset;
firewire: cdev: improve FW_CDEV_IOC_ALLOCATE In both the ieee1394 stack and the firewire stack, the core treats kernelspace drivers better than userspace drivers when it comes to CSR address range allocation: The former may request a register to be placed automatically at a free spot anywhere inside a specified address range. The latter may only request a register at a fixed offset. Hence, userspace drivers which do not require a fixed offset potentially need to implement a retry loop with incremented offset in each retry until the kernel does not fail allocation with EBUSY. This awkward procedure is not fundamentally necessary as the core already provides a superior allocation API to kernelspace drivers. Therefore change the ioctl() ABI by addition of a region_end member in the existing struct fw_cdev_allocate. Userspace and kernelspace APIs work the same way now. There is a small cost to pay by clients though: If client source code is required to compile with older kernel headers too, then any use of the new member fw_cdev_allocate.region_end needs to be enclosed by #ifdef/#endif directives. However, any client program that seriously wants to use address range allocations will require a kernel of cdev ABI version >= 4 at runtime and a linux/firewire-cdev.h header of >= 4 anyway. This is because v4 brings FW_CDEV_EVENT_REQUEST2. The only client program in which build-time compatibility with struct fw_cdev_allocate as found in older kernel headers makes sense is libraw1394. (libraw1394 uses the older broken FW_CDEV_EVENT_REQUEST to implement a makeshift, incorrect transaction responder that does at least work somewhat in many simple scenarios, relying on guesswork by libraw1394 and by libraw1394 based applications. Plus, address range allocation and transaction responder is only one of many features that libraw1394 needs to provide, and these other features need to work with kernel and kernel-headers as old as possible. Any new linux/firewire-cdev.h based client that implements a transaction responder should never attempt to do it like libraw1394; instead it should make a header and kernel of v4 or later a hard requirement.) While we are at it, update the struct fw_cdev_allocate documentation to better reflect the recent fw_cdev_event_request2 ABI addition. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23 05:05:39 -06:00
if (client->version < FW_CDEV_VERSION_ALLOCATE_REGION_END)
region.end = a->offset + a->length;
else
region.end = a->region_end;
r->handler.length = a->length;
r->handler.address_callback = handle_request;
r->handler.callback_data = r;
r->closure = a->closure;
r->client = client;
ret = fw_core_add_address_handler(&r->handler, &region);
if (ret < 0) {
kfree(r);
return ret;
}
firewire: cdev: improve FW_CDEV_IOC_ALLOCATE In both the ieee1394 stack and the firewire stack, the core treats kernelspace drivers better than userspace drivers when it comes to CSR address range allocation: The former may request a register to be placed automatically at a free spot anywhere inside a specified address range. The latter may only request a register at a fixed offset. Hence, userspace drivers which do not require a fixed offset potentially need to implement a retry loop with incremented offset in each retry until the kernel does not fail allocation with EBUSY. This awkward procedure is not fundamentally necessary as the core already provides a superior allocation API to kernelspace drivers. Therefore change the ioctl() ABI by addition of a region_end member in the existing struct fw_cdev_allocate. Userspace and kernelspace APIs work the same way now. There is a small cost to pay by clients though: If client source code is required to compile with older kernel headers too, then any use of the new member fw_cdev_allocate.region_end needs to be enclosed by #ifdef/#endif directives. However, any client program that seriously wants to use address range allocations will require a kernel of cdev ABI version >= 4 at runtime and a linux/firewire-cdev.h header of >= 4 anyway. This is because v4 brings FW_CDEV_EVENT_REQUEST2. The only client program in which build-time compatibility with struct fw_cdev_allocate as found in older kernel headers makes sense is libraw1394. (libraw1394 uses the older broken FW_CDEV_EVENT_REQUEST to implement a makeshift, incorrect transaction responder that does at least work somewhat in many simple scenarios, relying on guesswork by libraw1394 and by libraw1394 based applications. Plus, address range allocation and transaction responder is only one of many features that libraw1394 needs to provide, and these other features need to work with kernel and kernel-headers as old as possible. Any new linux/firewire-cdev.h based client that implements a transaction responder should never attempt to do it like libraw1394; instead it should make a header and kernel of v4 or later a hard requirement.) While we are at it, update the struct fw_cdev_allocate documentation to better reflect the recent fw_cdev_event_request2 ABI addition. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-23 05:05:39 -06:00
a->offset = r->handler.offset;
r->resource.release = release_address_handler;
ret = add_client_resource(client, &r->resource, GFP_KERNEL);
if (ret < 0) {
release_address_handler(client, &r->resource);
return ret;
}
a->handle = r->resource.handle;
return 0;
}
static int ioctl_deallocate(struct client *client, union ioctl_arg *arg)
{
return release_client_resource(client, arg->deallocate.handle,
release_address_handler, NULL);
}
static int ioctl_send_response(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_send_response *a = &arg->send_response;
struct client_resource *resource;
struct inbound_transaction_resource *r;
int ret = 0;
if (release_client_resource(client, a->handle,
release_request, &resource) < 0)
return -EINVAL;
r = container_of(resource, struct inbound_transaction_resource,
resource);
if (is_fcp_request(r->request))
goto out;
if (a->length != fw_get_response_length(r->request)) {
ret = -EINVAL;
kfree(r->request);
goto out;
}
if (copy_from_user(r->data, u64_to_uptr(a->data), a->length)) {
ret = -EFAULT;
kfree(r->request);
goto out;
}
firewire: cdev: fix responses to nodes at different card My box has two firewire cards in it: card0 and card1. My application opens /dev/fw0 (card 0) and allocates an address space. The core makes the address space available on both cards. Along comes the remote device, which sends a READ_QUADLET_REQUEST to card1. The request gets passed up to my application, which calls ioctl_send_response(). ioctl_send_response() then calls fw_send_response() with card0, because that's the card it's bound to. Card0's driver drops the response, because it isn't part of a transaction that it has outstanding. So in core-cdev: handle_request(), we need to stash the card of the inbound request in the struct inbound_transaction_resource and use that card to send the response to. The hard part will be refcounting the card correctly so it can't get deallocated while we hold a pointer to it. Here's a trivial patch, which does not do the card refcounting, but at least demonstrates what the problem is. Note that we can't depend on the fact that the core-cdev:client structure holds a card open, because in this case the card it holds open is not the card the request came in on. ..and there's no way for the core to tell cdev "this card is gone, kill any inbound transactions on it", while cdev holds the transaction open until userspace issues a SEND_RESPONSE ioctl, which may be a very, very long time. But when it does, it calls fw_send_response(), which will dereference the card... So how unhappy are we about userspace potentially holding a fw_card open forever? Signed-off-by: Jay Fenlason <fenlason@redhat.com> Reference counting to be addressed in a separate change. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (whitespace)
2010-05-18 12:02:45 -06:00
fw_send_response(r->card, r->request, a->rcode);
out:
firewire: cdev: count references of cards during inbound transactions If a request comes in to an address range managed by a userspace driver i.e. <linux/firewire-cdev.h> client, the card instance of request and response may differ from the card instance of the client device. Therefore we need to take a reference of the card until the response was sent. I thought about putting the reference counting into core-transaction.c, but the various high-level drivers besides cdev clients (firewire-net, firewire-sbp2, firedtv) use the card pointer in their fw_address_handler address_callback method only to look up devices of which they already hold the necessary references. So this seems to be a specific firewire-cdev issue which is better addressed locally. We do not need the reference - in case of FCP_REQUEST or FCP_RESPONSE requests because then the firewire-core will send the split transaction response for us already in the context of the request handler, - if it is the same card as the client device's because we hold a card reference indirectly via teh client->device reference. To keep things simple, we take the reference nevertheless. Jay Fenlason wrote: > there's no way for the core to tell cdev "this card is gone, > kill any inbound transactions on it", while cdev holds the transaction > open until userspace issues a SEND_RESPONSE ioctl, which may be a very, > very long time. But when it does, it calls fw_send_response(), which > will dereference the card... > > So how unhappy are we about userspace potentially holding a fw_card > open forever? While termination of inbound transcations at card removal could be implemented, it is IMO not worth the effort. Currently, the effect of holding a reference of a card that has been removed is to block the process that called the pci_remove of the card. This is - either a user process ran by root. Root can find and kill processes that have /dev/fw* open, if desired. - a kernel thread (which one?) in case of hot removal of a PCCard or ExpressCard. The latter case could be a problem indeed. firewire-core's card shutdown and card release should probably be improved not to block in shutdown, just to defer freeing of memory until release. This is not a new problem though; the same already always happens with the client->device->card without the need of inbound transactions or other special conditions involved, other than the client not closing the file. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-06-20 14:52:27 -06:00
fw_card_put(r->card);
kfree(r);
return ret;
}
static int ioctl_initiate_bus_reset(struct client *client, union ioctl_arg *arg)
{
fw_schedule_bus_reset(client->device->card, true,
arg->initiate_bus_reset.type == FW_CDEV_SHORT_RESET);
return 0;
}
static void release_descriptor(struct client *client,
struct client_resource *resource)
{
struct descriptor_resource *r =
container_of(resource, struct descriptor_resource, resource);
fw_core_remove_descriptor(&r->descriptor);
kfree(r);
}
static int ioctl_add_descriptor(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_add_descriptor *a = &arg->add_descriptor;
struct descriptor_resource *r;
int ret;
/* Access policy: Allow this ioctl only on local nodes' device files. */
if (!client->device->is_local)
return -ENOSYS;
if (a->length > 256)
return -EINVAL;
r = kmalloc(sizeof(*r) + a->length * 4, GFP_KERNEL);
if (r == NULL)
return -ENOMEM;
if (copy_from_user(r->data, u64_to_uptr(a->data), a->length * 4)) {
ret = -EFAULT;
goto failed;
}
r->descriptor.length = a->length;
r->descriptor.immediate = a->immediate;
r->descriptor.key = a->key;
r->descriptor.data = r->data;
ret = fw_core_add_descriptor(&r->descriptor);
if (ret < 0)
goto failed;
r->resource.release = release_descriptor;
ret = add_client_resource(client, &r->resource, GFP_KERNEL);
if (ret < 0) {
fw_core_remove_descriptor(&r->descriptor);
goto failed;
}
a->handle = r->resource.handle;
return 0;
failed:
kfree(r);
return ret;
}
static int ioctl_remove_descriptor(struct client *client, union ioctl_arg *arg)
{
return release_client_resource(client, arg->remove_descriptor.handle,
release_descriptor, NULL);
}
static void iso_callback(struct fw_iso_context *context, u32 cycle,
size_t header_length, void *header, void *data)
{
struct client *client = data;
struct iso_interrupt_event *e;
e = kmalloc(sizeof(*e) + header_length, GFP_ATOMIC);
if (e == NULL)
return;
e->interrupt.type = FW_CDEV_EVENT_ISO_INTERRUPT;
e->interrupt.closure = client->iso_closure;
e->interrupt.cycle = cycle;
e->interrupt.header_length = header_length;
memcpy(e->interrupt.header, header, header_length);
queue_event(client, &e->event, &e->interrupt,
sizeof(e->interrupt) + header_length, NULL, 0);
}
static void iso_mc_callback(struct fw_iso_context *context,
dma_addr_t completed, void *data)
{
struct client *client = data;
struct iso_interrupt_mc_event *e;
e = kmalloc(sizeof(*e), GFP_ATOMIC);
if (e == NULL)
return;
e->interrupt.type = FW_CDEV_EVENT_ISO_INTERRUPT_MULTICHANNEL;
e->interrupt.closure = client->iso_closure;
e->interrupt.completed = fw_iso_buffer_lookup(&client->buffer,
completed);
queue_event(client, &e->event, &e->interrupt,
sizeof(e->interrupt), NULL, 0);
}
static enum dma_data_direction iso_dma_direction(struct fw_iso_context *context)
{
if (context->type == FW_ISO_CONTEXT_TRANSMIT)
return DMA_TO_DEVICE;
else
return DMA_FROM_DEVICE;
}
static int ioctl_create_iso_context(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_create_iso_context *a = &arg->create_iso_context;
struct fw_iso_context *context;
fw_iso_callback_t cb;
int ret;
BUILD_BUG_ON(FW_CDEV_ISO_CONTEXT_TRANSMIT != FW_ISO_CONTEXT_TRANSMIT ||
FW_CDEV_ISO_CONTEXT_RECEIVE != FW_ISO_CONTEXT_RECEIVE ||
FW_CDEV_ISO_CONTEXT_RECEIVE_MULTICHANNEL !=
FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL);
switch (a->type) {
case FW_ISO_CONTEXT_TRANSMIT:
if (a->speed > SCODE_3200 || a->channel > 63)
return -EINVAL;
cb = iso_callback;
break;
case FW_ISO_CONTEXT_RECEIVE:
if (a->header_size < 4 || (a->header_size & 3) ||
a->channel > 63)
return -EINVAL;
cb = iso_callback;
break;
case FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL:
cb = (fw_iso_callback_t)iso_mc_callback;
break;
default:
return -EINVAL;
}
context = fw_iso_context_create(client->device->card, a->type,
a->channel, a->speed, a->header_size, cb, client);
if (IS_ERR(context))
return PTR_ERR(context);
if (client->version < FW_CDEV_VERSION_AUTO_FLUSH_ISO_OVERFLOW)
context->drop_overflow_headers = true;
/* We only support one context at this time. */
spin_lock_irq(&client->lock);
if (client->iso_context != NULL) {
spin_unlock_irq(&client->lock);
fw_iso_context_destroy(context);
return -EBUSY;
}
if (!client->buffer_is_mapped) {
ret = fw_iso_buffer_map_dma(&client->buffer,
client->device->card,
iso_dma_direction(context));
if (ret < 0) {
spin_unlock_irq(&client->lock);
fw_iso_context_destroy(context);
return ret;
}
client->buffer_is_mapped = true;
}
client->iso_closure = a->closure;
client->iso_context = context;
spin_unlock_irq(&client->lock);
a->handle = 0;
return 0;
}
static int ioctl_set_iso_channels(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_set_iso_channels *a = &arg->set_iso_channels;
struct fw_iso_context *ctx = client->iso_context;
if (ctx == NULL || a->handle != 0)
return -EINVAL;
return fw_iso_context_set_channels(ctx, &a->channels);
}
/* Macros for decoding the iso packet control header. */
#define GET_PAYLOAD_LENGTH(v) ((v) & 0xffff)
#define GET_INTERRUPT(v) (((v) >> 16) & 0x01)
#define GET_SKIP(v) (((v) >> 17) & 0x01)
#define GET_TAG(v) (((v) >> 18) & 0x03)
#define GET_SY(v) (((v) >> 20) & 0x0f)
#define GET_HEADER_LENGTH(v) (((v) >> 24) & 0xff)
static int ioctl_queue_iso(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_queue_iso *a = &arg->queue_iso;
struct fw_cdev_iso_packet __user *p, *end, *next;
struct fw_iso_context *ctx = client->iso_context;
unsigned long payload, buffer_end, transmit_header_bytes = 0;
u32 control;
int count;
struct {
struct fw_iso_packet packet;
u8 header[256];
} u;
if (ctx == NULL || a->handle != 0)
return -EINVAL;
/*
* If the user passes a non-NULL data pointer, has mmap()'ed
* the iso buffer, and the pointer points inside the buffer,
* we setup the payload pointers accordingly. Otherwise we
* set them both to 0, which will still let packets with
* payload_length == 0 through. In other words, if no packets
* use the indirect payload, the iso buffer need not be mapped
* and the a->data pointer is ignored.
*/
payload = (unsigned long)a->data - client->vm_start;
buffer_end = client->buffer.page_count << PAGE_SHIFT;
if (a->data == 0 || client->buffer.pages == NULL ||
payload >= buffer_end) {
payload = 0;
buffer_end = 0;
}
if (ctx->type == FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL && payload & 3)
return -EINVAL;
p = (struct fw_cdev_iso_packet __user *)u64_to_uptr(a->packets);
Remove 'type' argument from access_ok() function Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument of the user address range verification function since we got rid of the old racy i386-only code to walk page tables by hand. It existed because the original 80386 would not honor the write protect bit when in kernel mode, so you had to do COW by hand before doing any user access. But we haven't supported that in a long time, and these days the 'type' argument is a purely historical artifact. A discussion about extending 'user_access_begin()' to do the range checking resulted this patch, because there is no way we're going to move the old VERIFY_xyz interface to that model. And it's best done at the end of the merge window when I've done most of my merges, so let's just get this done once and for all. This patch was mostly done with a sed-script, with manual fix-ups for the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form. There were a couple of notable cases: - csky still had the old "verify_area()" name as an alias. - the iter_iov code had magical hardcoded knowledge of the actual values of VERIFY_{READ,WRITE} (not that they mattered, since nothing really used it) - microblaze used the type argument for a debug printout but other than those oddities this should be a total no-op patch. I tried to fix up all architectures, did fairly extensive grepping for access_ok() uses, and the changes are trivial, but I may have missed something. Any missed conversion should be trivially fixable, though. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-03 19:57:57 -07:00
if (!access_ok(p, a->size))
return -EFAULT;
end = (void __user *)p + a->size;
count = 0;
while (p < end) {
if (get_user(control, &p->control))
return -EFAULT;
u.packet.payload_length = GET_PAYLOAD_LENGTH(control);
u.packet.interrupt = GET_INTERRUPT(control);
u.packet.skip = GET_SKIP(control);
u.packet.tag = GET_TAG(control);
u.packet.sy = GET_SY(control);
u.packet.header_length = GET_HEADER_LENGTH(control);
switch (ctx->type) {
case FW_ISO_CONTEXT_TRANSMIT:
if (u.packet.header_length & 3)
return -EINVAL;
transmit_header_bytes = u.packet.header_length;
break;
case FW_ISO_CONTEXT_RECEIVE:
if (u.packet.header_length == 0 ||
u.packet.header_length % ctx->header_size != 0)
return -EINVAL;
break;
case FW_ISO_CONTEXT_RECEIVE_MULTICHANNEL:
if (u.packet.payload_length == 0 ||
u.packet.payload_length & 3)
return -EINVAL;
break;
}
next = (struct fw_cdev_iso_packet __user *)
&p->header[transmit_header_bytes / 4];
if (next > end)
return -EINVAL;
if (__copy_from_user
(u.packet.header, p->header, transmit_header_bytes))
return -EFAULT;
if (u.packet.skip && ctx->type == FW_ISO_CONTEXT_TRANSMIT &&
u.packet.header_length + u.packet.payload_length > 0)
return -EINVAL;
if (payload + u.packet.payload_length > buffer_end)
return -EINVAL;
if (fw_iso_context_queue(ctx, &u.packet,
&client->buffer, payload))
break;
p = next;
payload += u.packet.payload_length;
count++;
}
fw_iso_context_queue_flush(ctx);
a->size -= uptr_to_u64(p) - a->packets;
a->packets = uptr_to_u64(p);
a->data = client->vm_start + payload;
return count;
}
static int ioctl_start_iso(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_start_iso *a = &arg->start_iso;
BUILD_BUG_ON(
FW_CDEV_ISO_CONTEXT_MATCH_TAG0 != FW_ISO_CONTEXT_MATCH_TAG0 ||
FW_CDEV_ISO_CONTEXT_MATCH_TAG1 != FW_ISO_CONTEXT_MATCH_TAG1 ||
FW_CDEV_ISO_CONTEXT_MATCH_TAG2 != FW_ISO_CONTEXT_MATCH_TAG2 ||
FW_CDEV_ISO_CONTEXT_MATCH_TAG3 != FW_ISO_CONTEXT_MATCH_TAG3 ||
FW_CDEV_ISO_CONTEXT_MATCH_ALL_TAGS != FW_ISO_CONTEXT_MATCH_ALL_TAGS);
if (client->iso_context == NULL || a->handle != 0)
return -EINVAL;
if (client->iso_context->type == FW_ISO_CONTEXT_RECEIVE &&
(a->tags == 0 || a->tags > 15 || a->sync > 15))
return -EINVAL;
return fw_iso_context_start(client->iso_context,
a->cycle, a->sync, a->tags);
}
static int ioctl_stop_iso(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_stop_iso *a = &arg->stop_iso;
if (client->iso_context == NULL || a->handle != 0)
return -EINVAL;
return fw_iso_context_stop(client->iso_context);
}
static int ioctl_flush_iso(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_flush_iso *a = &arg->flush_iso;
if (client->iso_context == NULL || a->handle != 0)
return -EINVAL;
return fw_iso_context_flush_completions(client->iso_context);
}
static int ioctl_get_cycle_timer2(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_get_cycle_timer2 *a = &arg->get_cycle_timer2;
struct fw_card *card = client->device->card;
struct timespec64 ts = {0, 0};
u32 cycle_time;
int ret = 0;
local_irq_disable();
cycle_time = card->driver->read_csr(card, CSR_CYCLE_TIME);
switch (a->clk_id) {
case CLOCK_REALTIME: ktime_get_real_ts64(&ts); break;
case CLOCK_MONOTONIC: ktime_get_ts64(&ts); break;
case CLOCK_MONOTONIC_RAW: ktime_get_raw_ts64(&ts); break;
default:
ret = -EINVAL;
}
local_irq_enable();
a->tv_sec = ts.tv_sec;
a->tv_nsec = ts.tv_nsec;
a->cycle_timer = cycle_time;
return ret;
}
static int ioctl_get_cycle_timer(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_get_cycle_timer *a = &arg->get_cycle_timer;
struct fw_cdev_get_cycle_timer2 ct2;
ct2.clk_id = CLOCK_REALTIME;
ioctl_get_cycle_timer2(client, (union ioctl_arg *)&ct2);
a->local_time = ct2.tv_sec * USEC_PER_SEC + ct2.tv_nsec / NSEC_PER_USEC;
a->cycle_timer = ct2.cycle_timer;
return 0;
}
static void iso_resource_work(struct work_struct *work)
{
struct iso_resource_event *e;
struct iso_resource *r =
container_of(work, struct iso_resource, work.work);
struct client *client = r->client;
int generation, channel, bandwidth, todo;
bool skip, free, success;
spin_lock_irq(&client->lock);
generation = client->device->generation;
todo = r->todo;
/* Allow 1000ms grace period for other reallocations. */
if (todo == ISO_RES_ALLOC &&
time_before64(get_jiffies_64(),
client->device->card->reset_jiffies + HZ)) {
schedule_iso_resource(r, DIV_ROUND_UP(HZ, 3));
skip = true;
} else {
/* We could be called twice within the same generation. */
skip = todo == ISO_RES_REALLOC &&
r->generation == generation;
}
free = todo == ISO_RES_DEALLOC ||
todo == ISO_RES_ALLOC_ONCE ||
todo == ISO_RES_DEALLOC_ONCE;
r->generation = generation;
spin_unlock_irq(&client->lock);
if (skip)
goto out;
bandwidth = r->bandwidth;
fw_iso_resource_manage(client->device->card, generation,
r->channels, &channel, &bandwidth,
todo == ISO_RES_ALLOC ||
todo == ISO_RES_REALLOC ||
todo == ISO_RES_ALLOC_ONCE);
/*
* Is this generation outdated already? As long as this resource sticks
* in the idr, it will be scheduled again for a newer generation or at
* shutdown.
*/
if (channel == -EAGAIN &&
(todo == ISO_RES_ALLOC || todo == ISO_RES_REALLOC))
goto out;
success = channel >= 0 || bandwidth > 0;
spin_lock_irq(&client->lock);
/*
* Transit from allocation to reallocation, except if the client
* requested deallocation in the meantime.
*/
if (r->todo == ISO_RES_ALLOC)
r->todo = ISO_RES_REALLOC;
/*
* Allocation or reallocation failure? Pull this resource out of the
* idr and prepare for deletion, unless the client is shutting down.
*/
if (r->todo == ISO_RES_REALLOC && !success &&
!client->in_shutdown &&
idr_remove(&client->resource_idr, r->resource.handle)) {
client_put(client);
free = true;
}
spin_unlock_irq(&client->lock);
if (todo == ISO_RES_ALLOC && channel >= 0)
r->channels = 1ULL << channel;
if (todo == ISO_RES_REALLOC && success)
goto out;
if (todo == ISO_RES_ALLOC || todo == ISO_RES_ALLOC_ONCE) {
e = r->e_alloc;
r->e_alloc = NULL;
} else {
e = r->e_dealloc;
r->e_dealloc = NULL;
}
e->iso_resource.handle = r->resource.handle;
e->iso_resource.channel = channel;
e->iso_resource.bandwidth = bandwidth;
queue_event(client, &e->event,
&e->iso_resource, sizeof(e->iso_resource), NULL, 0);
if (free) {
cancel_delayed_work(&r->work);
kfree(r->e_alloc);
kfree(r->e_dealloc);
kfree(r);
}
out:
client_put(client);
}
static void release_iso_resource(struct client *client,
struct client_resource *resource)
{
struct iso_resource *r =
container_of(resource, struct iso_resource, resource);
spin_lock_irq(&client->lock);
r->todo = ISO_RES_DEALLOC;
schedule_iso_resource(r, 0);
spin_unlock_irq(&client->lock);
}
static int init_iso_resource(struct client *client,
struct fw_cdev_allocate_iso_resource *request, int todo)
{
struct iso_resource_event *e1, *e2;
struct iso_resource *r;
int ret;
if ((request->channels == 0 && request->bandwidth == 0) ||
request->bandwidth > BANDWIDTH_AVAILABLE_INITIAL)
return -EINVAL;
r = kmalloc(sizeof(*r), GFP_KERNEL);
e1 = kmalloc(sizeof(*e1), GFP_KERNEL);
e2 = kmalloc(sizeof(*e2), GFP_KERNEL);
if (r == NULL || e1 == NULL || e2 == NULL) {
ret = -ENOMEM;
goto fail;
}
INIT_DELAYED_WORK(&r->work, iso_resource_work);
r->client = client;
r->todo = todo;
r->generation = -1;
r->channels = request->channels;
r->bandwidth = request->bandwidth;
r->e_alloc = e1;
r->e_dealloc = e2;
e1->iso_resource.closure = request->closure;
e1->iso_resource.type = FW_CDEV_EVENT_ISO_RESOURCE_ALLOCATED;
e2->iso_resource.closure = request->closure;
e2->iso_resource.type = FW_CDEV_EVENT_ISO_RESOURCE_DEALLOCATED;
if (todo == ISO_RES_ALLOC) {
r->resource.release = release_iso_resource;
ret = add_client_resource(client, &r->resource, GFP_KERNEL);
if (ret < 0)
goto fail;
} else {
r->resource.release = NULL;
r->resource.handle = -1;
schedule_iso_resource(r, 0);
}
request->handle = r->resource.handle;
return 0;
fail:
kfree(r);
kfree(e1);
kfree(e2);
return ret;
}
static int ioctl_allocate_iso_resource(struct client *client,
union ioctl_arg *arg)
{
return init_iso_resource(client,
&arg->allocate_iso_resource, ISO_RES_ALLOC);
}
static int ioctl_deallocate_iso_resource(struct client *client,
union ioctl_arg *arg)
{
return release_client_resource(client,
arg->deallocate.handle, release_iso_resource, NULL);
}
static int ioctl_allocate_iso_resource_once(struct client *client,
union ioctl_arg *arg)
{
return init_iso_resource(client,
&arg->allocate_iso_resource, ISO_RES_ALLOC_ONCE);
}
static int ioctl_deallocate_iso_resource_once(struct client *client,
union ioctl_arg *arg)
{
return init_iso_resource(client,
&arg->allocate_iso_resource, ISO_RES_DEALLOC_ONCE);
}
/*
* Returns a speed code: Maximum speed to or from this device,
* limited by the device's link speed, the local node's link speed,
* and all PHY port speeds between the two links.
*/
static int ioctl_get_speed(struct client *client, union ioctl_arg *arg)
{
return client->device->max_speed;
}
static int ioctl_send_broadcast_request(struct client *client,
union ioctl_arg *arg)
{
struct fw_cdev_send_request *a = &arg->send_request;
switch (a->tcode) {
case TCODE_WRITE_QUADLET_REQUEST:
case TCODE_WRITE_BLOCK_REQUEST:
break;
default:
return -EINVAL;
}
/* Security policy: Only allow accesses to Units Space. */
if (a->offset < CSR_REGISTER_BASE + CSR_CONFIG_ROM_END)
return -EACCES;
return init_request(client, a, LOCAL_BUS | 0x3f, SCODE_100);
}
static int ioctl_send_stream_packet(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_send_stream_packet *a = &arg->send_stream_packet;
struct fw_cdev_send_request request;
int dest;
if (a->speed > client->device->card->link_speed ||
a->length > 1024 << a->speed)
return -EIO;
if (a->tag > 3 || a->channel > 63 || a->sy > 15)
return -EINVAL;
dest = fw_stream_packet_destination_id(a->tag, a->channel, a->sy);
request.tcode = TCODE_STREAM_DATA;
request.length = a->length;
request.closure = a->closure;
request.data = a->data;
request.generation = a->generation;
return init_request(client, &request, dest, a->speed);
}
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
static void outbound_phy_packet_callback(struct fw_packet *packet,
struct fw_card *card, int status)
{
struct outbound_phy_packet_event *e =
container_of(packet, struct outbound_phy_packet_event, p);
switch (status) {
/* expected: */
case ACK_COMPLETE: e->phy_packet.rcode = RCODE_COMPLETE; break;
/* should never happen with PHY packets: */
case ACK_PENDING: e->phy_packet.rcode = RCODE_COMPLETE; break;
case ACK_BUSY_X:
case ACK_BUSY_A:
case ACK_BUSY_B: e->phy_packet.rcode = RCODE_BUSY; break;
case ACK_DATA_ERROR: e->phy_packet.rcode = RCODE_DATA_ERROR; break;
case ACK_TYPE_ERROR: e->phy_packet.rcode = RCODE_TYPE_ERROR; break;
/* stale generation; cancelled; on certain controllers: no ack */
default: e->phy_packet.rcode = status; break;
}
e->phy_packet.data[0] = packet->timestamp;
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
queue_event(e->client, &e->event, &e->phy_packet,
sizeof(e->phy_packet) + e->phy_packet.length, NULL, 0);
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
client_put(e->client);
}
static int ioctl_send_phy_packet(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_send_phy_packet *a = &arg->send_phy_packet;
struct fw_card *card = client->device->card;
struct outbound_phy_packet_event *e;
/* Access policy: Allow this ioctl only on local nodes' device files. */
if (!client->device->is_local)
return -ENOSYS;
e = kzalloc(sizeof(*e) + 4, GFP_KERNEL);
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
if (e == NULL)
return -ENOMEM;
client_get(client);
e->client = client;
e->p.speed = SCODE_100;
e->p.generation = a->generation;
e->p.header[0] = TCODE_LINK_INTERNAL << 4;
e->p.header[1] = a->data[0];
e->p.header[2] = a->data[1];
e->p.header_length = 12;
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
e->p.callback = outbound_phy_packet_callback;
e->phy_packet.closure = a->closure;
e->phy_packet.type = FW_CDEV_EVENT_PHY_PACKET_SENT;
if (is_ping_packet(a->data))
e->phy_packet.length = 4;
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
card->driver->send_request(card, &e->p);
return 0;
}
static int ioctl_receive_phy_packets(struct client *client, union ioctl_arg *arg)
{
struct fw_cdev_receive_phy_packets *a = &arg->receive_phy_packets;
struct fw_card *card = client->device->card;
/* Access policy: Allow this ioctl only on local nodes' device files. */
if (!client->device->is_local)
return -ENOSYS;
spin_lock_irq(&card->lock);
list_move_tail(&client->phy_receiver_link, &card->phy_receiver_list);
client->phy_receiver_closure = a->closure;
spin_unlock_irq(&card->lock);
return 0;
}
void fw_cdev_handle_phy_packet(struct fw_card *card, struct fw_packet *p)
{
struct client *client;
struct inbound_phy_packet_event *e;
unsigned long flags;
spin_lock_irqsave(&card->lock, flags);
list_for_each_entry(client, &card->phy_receiver_list, phy_receiver_link) {
e = kmalloc(sizeof(*e) + 8, GFP_ATOMIC);
if (e == NULL)
break;
e->phy_packet.closure = client->phy_receiver_closure;
e->phy_packet.type = FW_CDEV_EVENT_PHY_PACKET_RECEIVED;
e->phy_packet.rcode = RCODE_COMPLETE;
e->phy_packet.length = 8;
e->phy_packet.data[0] = p->header[1];
e->phy_packet.data[1] = p->header[2];
queue_event(client, &e->event,
&e->phy_packet, sizeof(e->phy_packet) + 8, NULL, 0);
}
spin_unlock_irqrestore(&card->lock, flags);
}
static int (* const ioctl_handlers[])(struct client *, union ioctl_arg *) = {
[0x00] = ioctl_get_info,
[0x01] = ioctl_send_request,
[0x02] = ioctl_allocate,
[0x03] = ioctl_deallocate,
[0x04] = ioctl_send_response,
[0x05] = ioctl_initiate_bus_reset,
[0x06] = ioctl_add_descriptor,
[0x07] = ioctl_remove_descriptor,
[0x08] = ioctl_create_iso_context,
[0x09] = ioctl_queue_iso,
[0x0a] = ioctl_start_iso,
[0x0b] = ioctl_stop_iso,
[0x0c] = ioctl_get_cycle_timer,
[0x0d] = ioctl_allocate_iso_resource,
[0x0e] = ioctl_deallocate_iso_resource,
[0x0f] = ioctl_allocate_iso_resource_once,
[0x10] = ioctl_deallocate_iso_resource_once,
[0x11] = ioctl_get_speed,
[0x12] = ioctl_send_broadcast_request,
[0x13] = ioctl_send_stream_packet,
[0x14] = ioctl_get_cycle_timer2,
firewire: cdev: add PHY packet transmission Add an FW_CDEV_IOC_SEND_PHY_PACKET ioctl() for /dev/fw* which can be used to implement bus management related functionality in userspace. This is also half of the functionality (the transmit part) that is needed to support a userspace implementation of a VersaPHY transaction layer. Safety considerations: - PHY packets are generally broadcasts and may have interesting effects on PHYs and the bus, e.g. make asynchronous arbitration impossible due to too low gap count. Hence some kind of elevated privileges should be required of a process to be able to send PHY packets. This implementation assumes that a process that is allowed to open the /dev/fw* of a local node does have this privilege. There was an inconclusive discussion about introducing POSIX capabilities as a means to check for user privileges for these kinds of operations. - The kernel does not check integrity of the supplied packet data. That would be far too much code, considering the many kinds of PHY packets. A process which got the privilege to send these packets is trusted to do it correctly. Just like with the other "send packet" ioctls, a non-blocking API is chosen; i.e. the ioctl may return even before AT DMA started. After transmission, an event for poll()/read() is enqueued. Most users are going to need a blocking API, but a blocking userspace wrapper is easy to implement, and the second of the two existing libraw1394 calls raw1394_phy_packet_write() and raw1394_start_phy_packet_write() can be better supported that way. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2010-07-16 14:25:14 -06:00
[0x15] = ioctl_send_phy_packet,
[0x16] = ioctl_receive_phy_packets,
[0x17] = ioctl_set_iso_channels,
[0x18] = ioctl_flush_iso,
};
static int dispatch_ioctl(struct client *client,
unsigned int cmd, void __user *arg)
{
union ioctl_arg buffer;
int ret;
if (fw_device_is_shutdown(client->device))
return -ENODEV;
if (_IOC_TYPE(cmd) != '#' ||
_IOC_NR(cmd) >= ARRAY_SIZE(ioctl_handlers) ||
_IOC_SIZE(cmd) > sizeof(buffer))
2011-07-09 08:42:26 -06:00
return -ENOTTY;
firewire: cdev: prevent kernel stack leaking into ioctl arguments Found by the UC-KLEE tool: A user could supply less input to firewire-cdev ioctls than write- or write/read-type ioctl handlers expect. The handlers used data from uninitialized kernel stack then. This could partially leak back to the user if the kernel subsequently generated fw_cdev_event_'s (to be read from the firewire-cdev fd) which notably would contain the _u64 closure field which many of the ioctl argument structures contain. The fact that the handlers would act on random garbage input is a lesser issue since all handlers must check their input anyway. The fix simply always null-initializes the entire ioctl argument buffer regardless of the actual length of expected user input. That is, a runtime overhead of memset(..., 40) is added to each firewirew-cdev ioctl() call. [Comment from Clemens Ladisch: This part of the stack is most likely to be already in the cache.] Remarks: - There was never any leak from kernel stack to the ioctl output buffer itself. IOW, it was not possible to read kernel stack by a read-type or write/read-type ioctl alone; the leak could at most happen in combination with read()ing subsequent event data. - The actual expected minimum user input of each ioctl from include/uapi/linux/firewire-cdev.h is, in bytes: [0x00] = 32, [0x05] = 4, [0x0a] = 16, [0x0f] = 20, [0x14] = 16, [0x01] = 36, [0x06] = 20, [0x0b] = 4, [0x10] = 20, [0x15] = 20, [0x02] = 20, [0x07] = 4, [0x0c] = 0, [0x11] = 0, [0x16] = 8, [0x03] = 4, [0x08] = 24, [0x0d] = 20, [0x12] = 36, [0x17] = 12, [0x04] = 20, [0x09] = 24, [0x0e] = 4, [0x13] = 40, [0x18] = 4. Reported-by: David Ramos <daramos@stanford.edu> Cc: <stable@vger.kernel.org> Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
2014-11-11 09:16:44 -07:00
memset(&buffer, 0, sizeof(buffer));
if (_IOC_DIR(cmd) & _IOC_WRITE)
if (copy_from_user(&buffer, arg, _IOC_SIZE(cmd)))
return -EFAULT;
ret = ioctl_handlers[_IOC_NR(cmd)](client, &buffer);
if (ret < 0)
return ret;
if (_IOC_DIR(cmd) & _IOC_READ)
if (copy_to_user(arg, &buffer, _IOC_SIZE(cmd)))
return -EFAULT;
return ret;
}
static long fw_device_op_ioctl(struct file *file,
unsigned int cmd, unsigned long arg)
{
return dispatch_ioctl(file->private_data, cmd, (void __user *)arg);
}
#ifdef CONFIG_COMPAT
static long fw_device_op_compat_ioctl(struct file *file,
unsigned int cmd, unsigned long arg)
{
return dispatch_ioctl(file->private_data, cmd, compat_ptr(arg));
}
#endif
static int fw_device_op_mmap(struct file *file, struct vm_area_struct *vma)
{
struct client *client = file->private_data;
unsigned long size;
int page_count, ret;
if (fw_device_is_shutdown(client->device))
return -ENODEV;
/* FIXME: We could support multiple buffers, but we don't. */
if (client->buffer.pages != NULL)
return -EBUSY;
if (!(vma->vm_flags & VM_SHARED))
return -EINVAL;
if (vma->vm_start & ~PAGE_MASK)
return -EINVAL;
client->vm_start = vma->vm_start;
size = vma->vm_end - vma->vm_start;
page_count = size >> PAGE_SHIFT;
if (size & ~PAGE_MASK)
return -EINVAL;
ret = fw_iso_buffer_alloc(&client->buffer, page_count);
if (ret < 0)
return ret;
spin_lock_irq(&client->lock);
if (client->iso_context) {
ret = fw_iso_buffer_map_dma(&client->buffer,
client->device->card,
iso_dma_direction(client->iso_context));
client->buffer_is_mapped = (ret == 0);
}
spin_unlock_irq(&client->lock);
if (ret < 0)
goto fail;
ret = fw_iso_buffer_map_vma(&client->buffer, vma);
if (ret < 0)
goto fail;
return 0;
fail:
fw_iso_buffer_destroy(&client->buffer, client->device->card);
return ret;
}
static int is_outbound_transaction_resource(int id, void *p, void *data)
{
struct client_resource *resource = p;
return resource->release == release_transaction;
}
static int has_outbound_transactions(struct client *client)
{
int ret;
spin_lock_irq(&client->lock);
ret = idr_for_each(&client->resource_idr,
is_outbound_transaction_resource, NULL);
spin_unlock_irq(&client->lock);
return ret;
}
static int shutdown_resource(int id, void *p, void *data)
{
struct client_resource *resource = p;
struct client *client = data;
resource->release(client, resource);
client_put(client);
return 0;
}
static int fw_device_op_release(struct inode *inode, struct file *file)
{
struct client *client = file->private_data;
struct event *event, *next_event;
spin_lock_irq(&client->device->card->lock);
list_del(&client->phy_receiver_link);
spin_unlock_irq(&client->device->card->lock);
mutex_lock(&client->device->client_list_mutex);
list_del(&client->link);
mutex_unlock(&client->device->client_list_mutex);
if (client->iso_context)
fw_iso_context_destroy(client->iso_context);
if (client->buffer.pages)
fw_iso_buffer_destroy(&client->buffer, client->device->card);
/* Freeze client->resource_idr and client->event_list */
spin_lock_irq(&client->lock);
client->in_shutdown = true;
spin_unlock_irq(&client->lock);
wait_event(client->tx_flush_wait, !has_outbound_transactions(client));
idr_for_each(&client->resource_idr, shutdown_resource, client);
idr_destroy(&client->resource_idr);
list_for_each_entry_safe(event, next_event, &client->event_list, link)
kfree(event);
client_put(client);
return 0;
}
static __poll_t fw_device_op_poll(struct file *file, poll_table * pt)
{
struct client *client = file->private_data;
__poll_t mask = 0;
poll_wait(file, &client->wait, pt);
if (fw_device_is_shutdown(client->device))
mask |= EPOLLHUP | EPOLLERR;
if (!list_empty(&client->event_list))
mask |= EPOLLIN | EPOLLRDNORM;
return mask;
}
const struct file_operations fw_device_ops = {
.owner = THIS_MODULE,
.llseek = no_llseek,
.open = fw_device_op_open,
.read = fw_device_op_read,
.unlocked_ioctl = fw_device_op_ioctl,
.mmap = fw_device_op_mmap,
.release = fw_device_op_release,
.poll = fw_device_op_poll,
#ifdef CONFIG_COMPAT
.compat_ioctl = fw_device_op_compat_ioctl,
#endif
};