1 This file details changes in 2.6 which affect PCMCIA card driver authors:
2 * No dev_node_t (as of 2.6.35)
3 There is no more need to fill out a "dev_node_t" structure.
5 * New IRQ request rules (as of 2.6.35)
6 Instead of the old pcmcia_request_irq() interface, drivers may now
8 - calling request_irq/free_irq directly. Use the IRQ from *p_dev->irq.
9 - use pcmcia_request_irq(p_dev, handler_t); the PCMCIA core will
10 clean up automatically on calls to pcmcia_disable_device() or
12 - drivers still not capable of IRQF_SHARED (or not telling us so) may
13 use the deprecated pcmcia_request_exclusive_irq() for the time
14 being; they might receive a shared IRQ nonetheless.
16 * no cs_error / CS_CHECK / CONFIG_PCMCIA_DEBUG (as of 2.6.33)
17 Instead of the cs_error() callback or the CS_CHECK() macro, please use
18 Linux-style checking of return values, and -- if necessary -- debug
19 messages using "dev_dbg()" or "pr_debug()".
21 * New CIS tuple access (as of 2.6.33)
22 Instead of pcmcia_get_{first,next}_tuple(), pcmcia_get_tuple_data() and
23 pcmcia_parse_tuple(), a driver shall use "pcmcia_get_tuple()" if it is
24 only interested in one (raw) tuple, or "pcmcia_loop_tuple()" if it is
25 interested in all tuples of one type. To decode the MAC from CISTPL_FUNCE,
26 a new helper "pcmcia_get_mac_from_cis()" was added.
28 * New configuration loop helper (as of 2.6.28)
29 By calling pcmcia_loop_config(), a driver can iterate over all available
30 configuration options. During a driver's probe() phase, one doesn't need
31 to use pcmcia_get_{first,next}_tuple, pcmcia_get_tuple_data and
32 pcmcia_parse_tuple directly in most if not all cases.
34 * New release helper (as of 2.6.17)
35 Instead of calling pcmcia_release_{configuration,io,irq,win}, all that's
36 necessary now is calling pcmcia_disable_device. As there is no valid
37 reason left to call pcmcia_release_io and pcmcia_release_irq, the
38 exports for them were removed.
40 * Unify detach and REMOVAL event code, as well as attach and INSERTION
42 void (*remove) (struct pcmcia_device *dev);
43 int (*probe) (struct pcmcia_device *dev);
45 * Move suspend, resume and reset out of event handler (as of 2.6.16)
46 int (*suspend) (struct pcmcia_device *dev);
47 int (*resume) (struct pcmcia_device *dev);
48 should be initialized in struct pcmcia_driver, and handle
49 (SUSPEND == RESET_PHYSICAL) and (RESUME == CARD_RESET) events
51 * event handler initialization in struct pcmcia_driver (as of 2.6.13)
52 The event handler is notified of all events, and must be initialized
53 as the event() callback in the driver's struct pcmcia_driver.
55 * pcmcia/version.h should not be used (as of 2.6.13)
56 This file will be removed eventually.
58 * in-kernel device<->driver matching (as of 2.6.13)
59 PCMCIA devices and their correct drivers can now be matched in
60 kernelspace. See 'devicetable.txt' for details.
62 * Device model integration (as of 2.6.11)
63 A struct pcmcia_device is registered with the device model core,
64 and can be used (e.g. for SET_NETDEV_DEV) by using
65 handle_to_dev(client_handle_t * handle).
67 * Convert internal I/O port addresses to unsigned int (as of 2.6.11)
68 ioaddr_t should be replaced by unsigned int in PCMCIA card drivers.
70 * irq_mask and irq_list parameters (as of 2.6.11)
71 The irq_mask and irq_list parameters should no longer be used in
72 PCMCIA card drivers. Instead, it is the job of the PCMCIA core to
73 determine which IRQ should be used. Therefore, link->irq.IRQInfo2
76 * client->PendingEvents is gone (as of 2.6.11)
77 client->PendingEvents is no longer available.
79 * client->Attributes are gone (as of 2.6.11)
80 client->Attributes is unused, therefore it is removed from all
83 * core functions no longer available (as of 2.6.11)
84 The following functions have been removed from the kernel source
85 because they are unused by all in-kernel drivers, and no external
86 driver was reported to rely on them:
87 pcmcia_get_first_region()
88 pcmcia_get_next_region()
89 pcmcia_modify_window()
90 pcmcia_set_event_mask()
91 pcmcia_get_first_window()
92 pcmcia_get_next_window()
94 * device list iteration upon module removal (as of 2.6.10)
95 It is no longer necessary to iterate on the driver's internal
96 client list and call the ->detach() function upon module removal.
98 * Resource management. (as of 2.6.8)
99 Although the PCMCIA subsystem will allocate resources for cards,
100 it no longer marks these resources busy. This means that driver
101 authors are now responsible for claiming your resources as per
102 other drivers in Linux. You should use request_region() to mark
103 your IO regions in-use, and request_mem_region() to mark your
104 memory regions in-use. The name argument should be a pointer to
105 your driver name. Eg, for pcnet_cs, name should point to the
108 * CardServices is gone
109 CardServices() in 2.4 is just a big switch statement to call various
110 services. In 2.6, all of those entry points are exported and called
111 directly (except for pcmcia_report_error(), just use cs_error() instead).
113 * struct pcmcia_driver
114 You need to use struct pcmcia_driver and pcmcia_{un,}register_driver
115 instead of {un,}register_pccard_driver