linux/drivers/firewire/core-cdev.c, branch v2.6.39

firewire: core: fix card->reset_jiffies overflow

2011-01-23T11:31:01Z

On a 32-bit machine with, e.g., HZ=1000, jiffies will overflow after about 50 days, so if there are between 25 and 50 days between bus resets, the card->reset_jiffies comparisons can get wrong results. To fix this, ensure that this timestamp always uses 64 bits. Signed-off-by: Clemens Ladisch Signed-off-by: "Stefan Richter"

firewire: cdev: remove unneeded reference

2011-01-23T11:31:01Z

For outbound transactions, the IDR's and the callback's references now have exactly the same lifetime, so we do not need both of them. Signed-off-by: Clemens Ladisch Signed-off-by: "Stefan Richter"

firewire: cdev: always wait for outbound transactions to complete

2011-01-23T11:31:00Z

We must not use fw_cancel_transaction() because it cannot correctly abort still-active transactions. The only place in core-cdev where this matters is when the file is released. Instead of trying to abort the transactions, we wait for them to complete normally, i.e., until all outbound transaction resources have been removed from the IDR tree. Signed-off-by: Clemens Ladisch Signed-off-by: "Stefan Richter"

firewire: cdev: remove unneeded idr_find() from complete_transaction()

2011-01-23T11:31:00Z

Outbound transactions are never aborted with release_client_resource(), so it is not necessary for complete_transaction() to check whether the resource is still registered. Only shutdown_resource() can abort such an transaction, and this is already handled with the in_shutdown check. Signed-off-by: Clemens Ladisch Signed-off-by: "Stefan Richter"

firewire: make PHY packet header format consistent

2010-12-13T19:39:14Z

Change the header of PHY packets to be sent to include a pseudo transaction code. This makes the header consistent with that of received PHY packets, and allows at_context_queue_packet() and log_ar_at_event() to see the packet type directly instead of having to deduce it from the header length or even from the header contents. Signed-off-by: Clemens Ladisch Signed-off-by: Stefan Richter

Merge firewire branches to be released post v2.6.35

2010-08-02T08:09:04Z

Conflicts: drivers/firewire/core-card.c drivers/firewire/core-cdev.c and forgotten #include in drivers/firewire/ohci.c Signed-off-by: Stefan Richter

firewire: add isochronous multichannel reception

2010-07-29T21:09:18Z

This adds the DMA context programming and userspace ABI for multichannel reception, i.e. for listening on multiple channel numbers by means of a single DMA context. The use case is reception of more streams than there are IR DMA units offered by the link layer. This is already implemented by the older ohci1394 + ieee1394 + raw1394 stack. And as discussed recently on linux1394-devel, this feature is occasionally used in practice. The big drawbacks of this mode are that buffer layout and interrupt generation necessarily differ from single-channel reception: Headers and trailers are not stripped from packets, packets are not aligned with buffer chunks, interrupts are per buffer chunk, not per packet. These drawbacks also cause a rather hefty code footprint to support this rarely used OHCI-1394 feature. (367 lines added, among them 94 lines of added userspace ABI documentation.) This implementation enforces that a multichannel reception context may only listen to channels to which no single-channel context on the same link layer is presently listening to. OHCI-1394 would allow to overlay single-channel contexts by the multi-channel context, but this would be a departure from the present first-come-first-served policy of IR context creation. The implementation is heavily based on an earlier one by Jay Fenlason. Thanks Jay. Signed-off-by: Stefan Richter

firewire: core: small clarifications in core-cdev

2010-07-29T21:09:18Z

Make a note on the seemingly unused linux/sched.h. Rename an irritatingly named variable. Signed-off-by: Stefan Richter

firewire: core: remove unused code

2010-07-29T21:06:25Z

ioctl_create_iso_context enforces ctx->header_size >= 4. Signed-off-by: Stefan Richter

firewire: cdev: improve FW_CDEV_IOC_ALLOCATE

2010-07-23T11:36:28Z

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