------------------------------------------------------------------------ r1334 | scjody | 2005-09-27 21:52:21 +0200 (Tue, 27 Sep 2005) | 9 lines Changed paths: M /trunk/sbp2.c The following patch updates the hot-unplug and module removal fixes that were already committed to linux1394's repo to the version that was submitted to mainline. In particular, it changes how the scsi_device pointer is obtained, and it slightly optimizes and cleans up the abort and reset EH callbacks. Signed-off-by: Stefan Richter Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1333 | scjody | 2005-09-27 18:14:01 +0200 (Tue, 27 Sep 2005) | 5 lines Changed paths: M /trunk/ieee1394_core.h Make whitespace consistent with Linus' tree. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1332 | scjody | 2005-09-27 16:40:03 +0200 (Tue, 27 Sep 2005) | 13 lines Changed paths: M /trunk/nodemgr.c ieee1394: reorder activities after bus reset (fixes device detection) ieee1394/nodemgr: Units were not detected if the local IRM performed a bus reset. ("The root node is not cycle master capable; selecting a new root node and resetting...", often seen with iPods and other SBP-2 devices). Rearrange the order of IRM duties and node scanning. TODO: Audit the ROM caching and parsing code for underlying issues. Signed-off-by: Stefan Richter Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1331 | scjody | 2005-09-27 16:39:59 +0200 (Tue, 27 Sep 2005) | 10 lines Changed paths: M /trunk/nodemgr.c ieee1394: skip unnecessary pause when scanning config ROMs ieee1394/nodemgr: Skip a superfluous pause that occured when the config ROM of a node was scanned unsuccessfully. This also occurs if a node without link wrongly enables its "link active" self ID flag. A GWCTech 6-port hub does this. Signed-off-by: Stefan Richter Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1330 | scjody | 2005-09-27 16:39:55 +0200 (Tue, 27 Sep 2005) | 24 lines Changed paths: M /trunk/ieee1394_core.c Kill skb->list tree 2fd51ff3b7097eb3ffc41ea3a1d8b3ba04715b4c parent 6869c4d8e066e21623c812c448a05f1ed931c9c6 author David S. Miller Wed, 10 Aug 2005 09:25:21 -0700 committer David S. Miller Tue, 30 Aug 2005 05:31:14 -0700 [NET]: Kill skb->list Remove the "list" member of struct sk_buff, as it is entirely redundant. All SKB list removal callers know which list the SKB is on, so storing this in sk_buff does nothing other than taking up some space. Two tricky bits were SCTP, which I took care of, and two ATM drivers which Francois Romieu fixed up. Signed-off-by: David S. Miller Signed-off-by: Francois Romieu Signed-off-by: Jody McIntyre drivers/ieee1394/ieee1394_core.c | 4 ------------------------------------------------------------------------ r1329 | scjody | 2005-09-27 16:39:51 +0200 (Tue, 27 Sep 2005) | 23 lines Changed paths: M /trunk/ohci1394.c late spinlock initialization in ieee1394/ohci tree 986c5aa31761cbc186b4de7b410d539633f5ceb1 parent a46206e74e1897bf34d6b58f0991a0d6f3797e27 author Al Viro Thu, 25 Aug 2005 23:13:14 +0100 committer Linus Torvalds Sat, 27 Aug 2005 06:30:30 -0700 [PATCH] late spinlock initialization in ieee1394/ohci spinlock used in irq handler should be initialized before registering irq, even if we know that our device has interrupts disabled; handler is registered shared and taking spinlock is done unconditionally. As it is, we can and do get oopsen on boot for some configuration, depending on irq routing - I've got a reproducer. Signed-off-by: Al Viro Signed-off-by: Ben Collins Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre drivers/ieee1394/ohci1394.c | 8 +++++++- 1 files changed, 7 insertions(+), 1 deletion(-) ------------------------------------------------------------------------ r1328 | scjody | 2005-09-27 16:39:47 +0200 (Tue, 27 Sep 2005) | 8 lines Changed paths: M /trunk/eth1394.c M /trunk/eth1394.h Make eth1394 match Linus eth1394: We decided to work around the rawiso problem a different way. This patch makes SVN match the changes sent to akpm + Linus. Signed-off-by: Ben Collins Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1327 | scjody | 2005-09-27 16:39:44 +0200 (Tue, 27 Sep 2005) | 24 lines Changed paths: M /trunk/pcilynx.c I2C: Drop debug eeprom dump code in pcilynx tree 15746d1a5c66fa476487a7f95d84115f86ac3f81 parent 4e0c64cfc1126a2946487d4771168458bb9ba9be author Jean Delvare Thu, 25 Aug 2005 16:43:29 +0200 committer Greg Kroah-Hartman Mon, 05 Sep 2005 23:26:52 -0700 [PATCH] I2C: Drop debug eeprom dump code in pcilynx The pcilynx driver includes code to dump the contents of an i2c eeprom for debugging purposes. The same can be done from userspace using the i2cdump tool (part of the lm_sensors project) instead, in a more efficient and flexible way. Thus I would suggest that this functionality be simply dropped from the pcilynx driver. Signed-off-by: Jean Delvare Signed-off-by: Greg Kroah-Hartman Signed-off-by: Jody McIntyre drivers/ieee1394/pcilynx.c | 20 -------------------- 1 files changed, 20 deletions(-) ------------------------------------------------------------------------ r1326 | scjody | 2005-09-27 16:39:40 +0200 (Tue, 27 Sep 2005) | 51 lines Changed paths: M /trunk/nodemgr.c provide MODALIAS= enviroment variable to autoload ieee1394 modules via udev tree 42dca9f6fb039abcde575b377ffd29ce4e29b1ca parent 736c7b808f38f3bb72941345e11e236ec65dec3d author Olaf Hering Wed, 07 Sep 2005 05:18:18 -0700 committer Linus Torvalds Thu, 08 Sep 2005 06:57:41 -0700 [PATCH] provide MODALIAS= enviroment variable to autoload ieee1394 modules via udev https://bugzilla.novell.com/show_bug.cgi?id=103746 Old 2.6.13 hotplug enviroment for 'plug in firewire disk' event: ==> debug.01139.ieee1394.add.8211 <== set -- ieee1394 UDEV_LOG='7' ACTION='add' DEVPATH='/class/ieee1394/00010410100036e0-0' SUBSYSTEM='ieee1394' SEQNUM='1139' PHYSDEVPATH='/devices/pci0001:10/0001:10:0d.0/0001:11:0a.0/fw-host0/00010410100036e0/00010410100036e0-0' PHYSDEVBUS='ieee1394' VENDOR_ID='000000' MODEL_ID='001010' GUID='00010410100036e0' SPECIFIER_ID='00609e' VERSION='010483' UDEVD_EVENT='1' Module spb2 is not loaded. grep sbp2 /lib/modules/2.6.13-20050901172817-default/modules.alias alias ieee1394:ven*mo*sp0000609Ever00010483* sbp2 printf 'ieee1394:ven%08Xmo%08Xsp%08Xver%08X\n' '0x000000' '0x001010' '0x00609e' '0x010483' ieee1394:ven00000000mo00001010sp0000609Ever00010483 modprobe -v ieee1394:ven00000000mo00001010sp0000609Ever00010483 insmod /lib/modules/2.6.13-20050901172817-default/kernel/drivers/ieee1394/sbp2.ko Providing a MODALIAS= enviroment variable with the content above will fix it. Signed-off-by: Olaf Hering Cc: Ben Collins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre drivers/ieee1394/nodemgr.c | 8 ++++++++ 1 files changed, 8 insertions(+) ------------------------------------------------------------------------ r1325 | scjody | 2005-09-27 16:39:37 +0200 (Tue, 27 Sep 2005) | 16 lines Changed paths: M /trunk/video1394.c trivial __user cleanup (video1394) tree 91c4a76517f5d166d31c47a920ec6f24cdb07fa6 parent 18fb9442576c79fc2f4bca1fd90c3aa4d1115e97 author viro@ZenIV.linux.org.uk Fri, 09 Sep 2005 20:26:01 +0100 committer Linus Torvalds Sat, 10 Sep 2005 04:05:54 -0700 [PATCH] trivial __user cleanup (video1394) Signed-off-by: Al Viro Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre drivers/ieee1394/video1394.c | 2 +- 1 files changed, 1 insertion(+), 1 deletion(-) ------------------------------------------------------------------------ r1324 | scjody | 2005-09-27 16:39:34 +0200 (Tue, 27 Sep 2005) | 15 lines Changed paths: M /trunk/sbp2.c sbp2: default to serialize_io=1 sbp2: Set serialize_io=1 by default. This is safer and required by seemingly more and more hardware. It causes little or no performance loss for S400 devices. Performance of S800 1394b devices may drop by 25...30%. Therefore make the parameter's description and dmesg message clearer about performance impact. Update description of the max_speed parameter too. IEEE1394_SPEED_MAX is currently S800. Signed-off-by: Stefan Richter Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1323 | scjody | 2005-09-27 16:38:37 +0200 (Tue, 27 Sep 2005) | 7 lines Changed paths: M /trunk/sbp2.c sbp2: make CONFIG_IEEE_SBP2_DEBUG #define match Linus's tree An 'undefined symbol' problem related to CONFIG_IEEE_SBP2_DEBUG was fixed differently in Linus's tree. Make the fix here match that fix. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1322 | bcollins | 2005-08-15 22:13:58 +0200 (Mon, 15 Aug 2005) | 5 lines Changed paths: M /trunk/amdtp.c Use schedule_timeout_interruptible() instead of set_current_state()/schedule_timeout() to reduce kernel size. Signed-off-by: Nishanth Aravamudan ------------------------------------------------------------------------ r1321 | bcollins | 2005-08-15 22:11:50 +0200 (Mon, 15 Aug 2005) | 6 lines Changed paths: M /trunk/amdtp.c M /trunk/dv1394.c M /trunk/raw1394.c M /trunk/video1394.c amdtp, dv1394, raw1394, video1394: Delete legacy module aliases. The macros did not work and the aliases are not needed nowadays. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1320 | bcollins | 2005-08-15 22:09:42 +0200 (Mon, 15 Aug 2005) | 15 lines Changed paths: M /trunk/sbp2.c Add some more paranoia in sbp2scsi_queuecommand. When a FireWire CD-ROM or RBC harddisk was detached, the ieee1394 subsystem and probably the scsi subsystem was locked up because scsi_remove_device() is prone to hang after scsi_block_requests(). sbp2 now no longer blocks requests after a node was detached but rather completes them with DID_NO_CONNECT. The same is done if the FireWire host is shutting down. This fixes "modprobe -r ohci1394" hanging if sbp2 was still connected to CD-ROMs or RBC disks. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1319 | bcollins | 2005-08-15 21:27:18 +0200 (Mon, 15 Aug 2005) | 5 lines Changed paths: M /trunk/hosts.h Trivial correction of a few comments Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1318 | bcollins | 2005-08-15 21:19:31 +0200 (Mon, 15 Aug 2005) | 6 lines Changed paths: M /trunk/sbp2.c Fix some warnings after -Wno-def was added to global CFLAGS original patch by Olaf Hering, 27 July 2005, slightly touched up Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1317 | scjody | 2005-08-09 21:03:44 +0200 (Tue, 09 Aug 2005) | 5 lines Changed paths: M /trunk/csr1212.h Remove an include that's no longer needed. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1316 | bcollins | 2005-07-23 18:09:58 +0200 (Sat, 23 Jul 2005) | 7 lines Changed paths: M /trunk/sbp2.c Reactivates the currently unused sdev member in struct scsi_id_instance_data, points it to the scsi device obtained by scsi_add_device(), and calls scsi_remove_device() before logout. Signed-off-by: Stefan Richter Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1315 | bcollins | 2005-07-15 00:45:47 +0200 (Fri, 15 Jul 2005) | 7 lines Changed paths: M /trunk/hosts.c Use of time_before() macro, defined at linux/jiffies.h, which deal with wrapping correctly and are nicer to read. Signed-off-by: Marcelo Feitoza Parisi Signed-off-by: Domen Puncer Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1314 | bcollins | 2005-07-14 23:40:40 +0200 (Thu, 14 Jul 2005) | 6 lines Changed paths: M /trunk/ieee1394_core.c Fix debug code so it prints the correct speed (was defaulting to 100, so anything > 400 showed only 100). Signed-off-by: Stefan Richter Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1313 | bcollins | 2005-07-14 18:22:50 +0200 (Thu, 14 Jul 2005) | 6 lines Changed paths: M /trunk/ohci1394.c printk_ratelimit() for IR legacy context enabling Signed-off-by: Olaf Hering Signed-off-by: Andrew Morton Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1312 | bcollins | 2005-07-14 05:32:37 +0200 (Thu, 14 Jul 2005) | 3 lines Changed paths: M /trunk/eth1394.c Better workaround for limiting iso recv size to 2048. Makes sure we use the lesser of what the host supports and 2048. ------------------------------------------------------------------------ r1311 | bcollins | 2005-07-13 18:19:49 +0200 (Wed, 13 Jul 2005) | 5 lines Changed paths: M /trunk/sbp2.c M /trunk/sbp2.h Remove our tracking of device type, since we no longer need to worry about it. Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1310 | bcollins | 2005-07-13 18:02:31 +0200 (Wed, 13 Jul 2005) | 5 lines Changed paths: M /trunk/sbp2.c M /trunk/sbp2.h Merge TYPE_RBC and 10byte removal patch from scsi maintainers. Added more cleanups, to remove unused code. Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1309 | bcollins | 2005-07-11 20:15:51 +0200 (Mon, 11 Jul 2005) | 7 lines Changed paths: M /trunk/eth1394.c Work around in eth1394 for limitations in buffer size for rawiso. Occurs on platforms with PAGE_SIZE >= 4096, and using S800 packet sizes. Original patch by Steve Kinneberg. Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1308 | bcollins | 2005-07-10 19:36:32 +0200 (Sun, 10 Jul 2005) | 5 lines Changed paths: M /trunk/sbp2.c Call scsi_device_put() after scsi_add_device() as suggested by James Bottomley. Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1307 | bcollins | 2005-07-10 01:47:36 +0200 (Sun, 10 Jul 2005) | 4 lines Changed paths: M /trunk/dma.c Get rid of compiler warning Signed-off-by: Ben Collins ------------------------------------------------------------------------ r1306 | bcollins | 2005-07-10 01:26:53 +0200 (Sun, 10 Jul 2005) | 2 lines Changed paths: M /trunk/sbp2.c Oops, RBC, not RDC. ------------------------------------------------------------------------ r1305 | bcollins | 2005-07-10 01:25:18 +0200 (Sun, 10 Jul 2005) | 2 lines Changed paths: M /trunk/sbp2.c M /trunk/sbp2.h s/TYPE_SDAD/TYPE_RDC/ ------------------------------------------------------------------------ r1304 | bcollins | 2005-07-10 00:42:58 +0200 (Sun, 10 Jul 2005) | 6 lines Changed paths: M /trunk/sbp2.c Taken from a patch by Jeff Garzik, committed to Linus' tree on June 17: allow sleeping in ->eh_device_reset_handler() Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1303 | bcollins | 2005-07-09 23:18:27 +0200 (Sat, 09 Jul 2005) | 7 lines Changed paths: M /trunk/sbp2.c Taken from a patch by James Bottomley from May, submitted to Linus' tree by Jeff Garzik on June 17: "The SCSI ->done() hook should not be called from inside a spinlock." Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1302 | bcollins | 2005-07-09 23:10:36 +0200 (Sat, 09 Jul 2005) | 8 lines Changed paths: M /trunk/Kconfig M /trunk/ieee1394_core.c IEEE 1394 subsystem: Wrap symbols that are neither used by the 1394 drivers in the mainline kernel (a.k.a. Linux1394) nor by dfg1394 (a.k.a the video-2-1394 project) into a new configuration option. Switch this option on by default. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1301 | bcollins | 2005-07-09 20:18:28 +0200 (Sat, 09 Jul 2005) | 11 lines Changed paths: M /trunk/ieee1394_core.c M /trunk/nodemgr.c Update to 2.6.13's refrigerator calls. Required to insert linux1394's trunk into the mainline kernel pre-release. Patch is derived from commit 3e1d1d28d99dabe63c64f7f40f1ca1d646de1f73, author Christoph Lameter , 25 Jun 2005. Warning: This breaks linux1394's compatibility with Linux 2.6.11 and 2.6.12 at source level. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1300 | bcollins | 2005-07-09 20:15:59 +0200 (Sat, 09 Jul 2005) | 36 lines Changed paths: M /trunk/dv1394.c M /trunk/ieee1394_core.c M /trunk/ieee1394_core.h M /trunk/nodemgr.c M /trunk/raw1394.c M /trunk/sbp2.c M /trunk/video1394.c Syncing patch from mainline. author gregkh@suse.de Wed, 23 Mar 2005 17:53:36 +0000 (09:53 -0800) committer Greg Kroah-Hartman Mon, 20 Jun 2005 22:15:08 +0000 (15:15 -0700) commit 7e25ab9155aef04e83da69545742cf65c9b801ce [PATCH] class: convert drivers/ieee1394/* to use the new class api instead of class_simple ----- author gregkh@suse.de Fri, 25 Mar 2005 19:45:31 +0000 (11:45 -0800) committer Greg Kroah-Hartman Mon, 20 Jun 2005 22:15:26 +0000 (15:15 -0700) commit 64360322ab3330d4881166380ad43a1eec2f123d [PATCH] Use device_for_each_child() to unregister devices in nodemgr_remove_host_dev() ----- author Yani Ioannou Tue, 17 May 2005 10:42:58 +0000 (06:42 -0400) committer Greg Kroah-Hartman Mon, 20 Jun 2005 22:15:34 +0000 (15:15 -0700) commit e404e274f62665f3333d6a539d0d3701f678a598 [PATCH] Driver Core: drivers/i2c/chips/w83781d.c - drivers/s390/block/dcssblk.c: update device attribute callbacks ----- combined patch: Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1299 | bcollins | 2005-07-09 19:48:15 +0200 (Sat, 09 Jul 2005) | 7 lines Changed paths: M /trunk/ohci1394.c Replace CONFIG_PMAC_PBOOK by CONFIG_PPC_PMAC/ in ohci1394's suspend and resume callbacks. Patch taken from linux-2.6.13-rc2. Original author of the patch is not known to me. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1298 | bcollins | 2005-07-09 19:31:32 +0200 (Sat, 09 Jul 2005) | 6 lines Changed paths: M /trunk/ohci1394.h M /trunk/video1394.c Patch removes a few trailing spaces. Brings linux1394.org's trunk on par with linux-2.6.13-rc2, whitespace-wise. Signed-off-by: Ben Collins Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1297 | ddennedy | 2005-07-08 18:51:48 +0200 (Fri, 08 Jul 2005) | 8 lines Changed paths: M /trunk/raw1394.c Fix the sparse warning "implicit cast to nocast type" File/Subsystem: drivers/ieee1394/raw1394.c Signed-off-by: Victor Fusco Signed-off-by: Domen Puncer Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1296 | ddennedy | 2005-07-08 01:07:42 +0200 (Fri, 08 Jul 2005) | 21 lines Changed paths: M /trunk/csr1212.c Add index and pointer checks to various csr1212 library functions. When parameters are out of bound... csr1212_init_local_csr() - max_rom: ...call BUG_ON when in kernel, insert an impossible value which safely fails later when in userspace csr1212_new_icon_descriptor_leaf() - palette_depth, color_space: ...return error value NULL csr1212_append_new_cache() - csr, csr->ops..., csr->max_rom: ...return EINVAL instead of ENOMEM csr1212_generate_csr_image() - csr->max_rom, _csr1212_read_keyval() - csr->max_rom, csr1212_parse_csr() - csr->ops, csr->ops->get_max_rom(): ...return EINVAL Signed-off-by: Stefan Richter Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1295 | ddennedy | 2005-07-08 00:48:04 +0200 (Fri, 08 Jul 2005) | 18 lines Changed paths: M /trunk/ieee1394_core.h Platforms differ WRT hpsb_packet.ack_code and hpsb_packet.tcode being signed char or unsigned char. This leads to different behaviour when ack_code is cast to int in hpsb_packet_success(). All ACKX_* codes may be mistaken for invalid ack, although this only adds a spurious warning. There does not seem to be a problem with tcode, but better be safe and use the same datatype on all platforms. Spotted by Paul Mackerras (?) and Benjamin Herrenschmidt, http://marc.theaimsgroup.com/?l=linux1394-devel&m=111465364509329 ACKX_* mismatch seen in a report by Sjoerd Simons, http://marc.theaimsgroup.com/?l=linux1394-user&m=111692728832395 Signed-off-by: Stefan Richter Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1294 | ddennedy | 2005-07-02 02:04:25 +0200 (Sat, 02 Jul 2005) | 5 lines Changed paths: M /trunk/raw1394.c Fix previous erroneous checkin r1291, validation changes incorrectly applied to receive function, should have been on transmit. Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1293 | ddennedy | 2005-07-01 21:44:01 +0200 (Fri, 01 Jul 2005) | 11 lines Changed paths: M /trunk/csr.c M /trunk/ohci1394.c ieee1394's disable_irm parameter missed to switch off every last reminder of IRM capability. csr.c, add_host(): Clear IRMC in the initial ROM image. ohci1394.c, ohci_initialize(): Explicitly clear IRMC in the bus options and selfID. We should not rely on a power reset default of 0. Signed-off-by: Stefan Richter Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1292 | ddennedy | 2005-07-01 20:54:25 +0200 (Fri, 01 Jul 2005) | 5 lines Changed paths: M /trunk/pcilynx.c Bugfix inverted binary instead of logical when masking the contender bit out. Signed-off-by: Stefan Richter Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1291 | ddennedy | 2005-07-01 20:51:31 +0200 (Fri, 01 Jul 2005) | 7 lines Changed paths: M /trunk/iso.c M /trunk/raw1394.c Specify different errnos for different validations on raw1394_iso_send_packets(). Prevent rawiso xmit from stalling on start by prebuffering buf_packets or by using -1 (default) parameter value. Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1290 | bcollins | 2005-07-01 18:41:19 +0200 (Fri, 01 Jul 2005) | 7 lines Changed paths: M /trunk/ieee1394_core.c M /trunk/nodemgr.c M /trunk/pcilynx.c M /trunk/sbp2.c Remove some hard 0 initialized static variables and let gcc put things where they are supposed to be (.bss) and allow the kernel to zero them. Signed-off-by: Ben Collins Signed-off-by: Carlo Perassi Signed-off-by: Domen Puncer ------------------------------------------------------------------------ r1289 | ddennedy | 2005-06-25 20:13:21 +0200 (Sat, 25 Jun 2005) | 5 lines Changed paths: M /trunk/raw1394.c Do not let user submit max packets for rawiso transmission, which would prevent DMA from resuming operation. Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1288 | bcollins | 2005-06-24 22:09:41 +0200 (Fri, 24 Jun 2005) | 7 lines Changed paths: M /trunk/sbp2.c Fix allocation of scsi host so we get enough space to use for hostdata[0]. Fixes crashes related to sbp2 (slab corruption). Signed-off-by: Ben Collins Signed-off-by: Alexandre Oliva ------------------------------------------------------------------------ r1287 | ddennedy | 2005-06-22 22:31:01 +0200 (Wed, 22 Jun 2005) | 4 lines Changed paths: M /trunk/iso.c bugfix to prevent divide by zero when iso buffer size is less than 4 packets. Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1286 | ddennedy | 2005-05-19 14:51:44 +0200 (Thu, 19 May 2005) | 9 lines Changed paths: M /branches/linux-2.4/iso.c M /branches/linux-2.4/iso.h M /branches/linux-2.4/ohci1394.c This patch fixes a memory leak with the circular buffers used to hold incoming Isochronous packets from an OHCI interface. The symptom is that an application using libraw1394 to receive Iso packets simply stops receiving them after a while, and the application has to be restarted to continue reception. Signed-off-by: Keith Bengston Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1285 | ddennedy | 2005-05-19 05:04:43 +0200 (Thu, 19 May 2005) | 9 lines Changed paths: M /trunk/iso.c M /trunk/iso.h M /trunk/ohci1394.c This patch fixes a memory leak with the circular buffers used to hold incoming Isochronous packets from an OHCI interface. The symptom is that an application using libraw1394 to receive Iso packets simply stops receiving them after a while, and the application has to be restarted to continue reception. Signed-off-by: Keith Bengston Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1284 | scjody | 2005-05-06 19:32:48 +0200 (Fri, 06 May 2005) | 7 lines Changed paths: M /trunk/ieee1394_core.c Set the initial sendtime to be 10 seconds in the future, to avoid the packet timing out while it's still queued to be sent. This fixes furthur "no tlabel match" problems caused by premature expiry. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1283 | scjody | 2005-05-06 19:17:57 +0200 (Fri, 06 May 2005) | 7 lines Changed paths: M /trunk/ohci1394.h Fix cross_bound to not return 1 for zero-length regions. Fixes regression when sending null ISO packets. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1282 | scjody | 2005-04-22 16:03:21 +0200 (Fri, 22 Apr 2005) | 5 lines Changed paths: M /trunk/video1394.c Removed useless + 0 * whatever; terms to avoid confusion. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1281 | scjody | 2005-04-21 22:49:42 +0200 (Thu, 21 Apr 2005) | 10 lines Changed paths: M /trunk/pcilynx.c Use the DMA_32BIT_MASK constant from dma-mapping.h when calling pci_set_dma_mask() or pci_set_consistent_dma_mask() These patches include dma-mapping.h explicitly because it caused errors on some architectures otherwise. See http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for details Signed-off-by: Tobias Klauser Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1280 | scjody | 2005-04-21 22:48:11 +0200 (Thu, 21 Apr 2005) | 6 lines Changed paths: M /trunk/nodemgr.c M /trunk/ohci1394.c M /trunk/video1394.c This patch removes redundant NULL pointer checks before kfree() in all of drivers/ieee1394/ Signed-off-by: Jesper Juhl Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1279 | scjody | 2005-04-21 22:44:48 +0200 (Thu, 21 Apr 2005) | 10 lines Changed paths: M /trunk/ieee1394_core.c M /trunk/ieee1394_transactions.c M /trunk/ieee1394_transactions.h This patch contains the following possible cleanups: - #if 0 the following unused global functions: - hpsb_lock - hpsb_send_gasp - ieee1394_transactions.h: remove the stale hpsb_lock64 prototype Signed-off-by: Adrian Bunk Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1278 | scjody | 2005-04-21 22:42:44 +0200 (Thu, 21 Apr 2005) | 7 lines Changed paths: M /trunk/Kconfig M /trunk/pcilynx.c M /trunk/pcilynx.h The options CONFIG_IEEE1394_PCILYNX_LOCALRAM and CONFIG_IEEE1394_PCILYNX_PORTS are not available for some time. Removed. Signed-off-by: Adrian Bunk Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1277 | scjody | 2005-04-21 22:40:24 +0200 (Thu, 21 Apr 2005) | 9 lines Changed paths: M /trunk/ohci1394.c Fix non-legacy multichannel ISO receive, broken by Parag Wardukar's allocation fix. Multichannel ISO receive still sucks; it should be possible to use both legacy and non-legacy modes at the same time, but with this patch, things are no worse than they were in 2.6.11 and allocation is still done at the correct time. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1276 | scjody | 2005-04-21 22:40:02 +0200 (Thu, 21 Apr 2005) | 11 lines Changed paths: M /trunk/ohci1394.c Subject: [PATCH] ohci1394: tlabels misprinted in DBGMSG From: Stefan Richter - Print the correct value in the DBGMSG in dma_rcv_tasklet(). See OHCI 1.1 section 8.7, page 103 ff. - Print tlabels as %d everywhere. Signed-off-by: Stefan Richter Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1275 | ddennedy | 2005-04-20 03:45:25 +0200 (Wed, 20 Apr 2005) | 6 lines Changed paths: M /trunk/video1394.c Apply and fixup patch from Markus Tavenrath for video1394 to allow only a single buffer on receive and two buffers on transmit. Tested with libdc1394 and dvconnect (libdv). Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1274 | scjody | 2005-04-19 19:54:37 +0200 (Tue, 19 Apr 2005) | 12 lines Changed paths: M /trunk/raw1394.c ChangeSet 1.2071, 2005/03/13 16:28:06-08:00, juhl-lkml@dif.dk [PATCH] verify_area cleanup : drivers part 1 This patch converts the first half of drivers from verify_area to access_ok. Signed-off-by: Jesper Juhl Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds ------------------------------------------------------------------------ r1273 | scjody | 2005-03-07 21:26:44 +0100 (Mon, 07 Mar 2005) | 7 lines Changed paths: M /trunk/ohci1394.c M /trunk/pcilynx.c convert from pci_module_init to pci_register_driver Signed-off-by: Christophe Lucas Signed-off-by: Domen Puncer ------------------------------------------------------------------------ r1272 | scjody | 2005-03-07 20:52:06 +0100 (Mon, 07 Mar 2005) | 9 lines Changed paths: M /trunk/video1394.c Description: Use wait_event_interruptible() instead of the deprecated interruptible_sleep_on(). The first change is simply to clean up the code a little to make it clearer. The second actually does a replacement, mimicking exactly the first. I removed the #if 1/#else/endif logic, as it duplicated the same code. Patch is compile-tested. Signed-off-by: Nishanth Aravamudan ------------------------------------------------------------------------ r1271 | scjody | 2005-03-07 20:26:36 +0100 (Mon, 07 Mar 2005) | 20 lines Changed paths: M /trunk/ohci1394.c ohci1394.c allocates the legacy IR DMA Context on demand. This happens in IRQ path resulting in call to dma_pool_create from within interrupt. Same is true for de-allocation of the IR DMA Context - it happens again in IRQ path resulting in call to dma_pool_destroy. After attempting to fix the above by using schedule_work for delayed allocation and de-allocation, it was determined that it should be fixed simply by moving the allocation to the drivers probe() function and de-allocation to remove() function. This way it reduces the complexity and does not result into any significant wastage of resources. This patch also fixes duplicate pci pool names and removes spaces from the names to make it easier for parsing by scripts. Thanks go to Dave Brownell for greatly simplifying this. Signed-off-by: Parag Warudkar Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1270 | scjody | 2005-03-07 20:17:26 +0100 (Mon, 07 Mar 2005) | 7 lines Changed paths: M /trunk/video1394.c Description: Use wait_event_interruptible() instead of the deprecated interruptible_sleep_on(). Add a helper function to make the condition for wait_event_interruptible() sane and lock-safe. Patch is compile-tested. Signed-off-by: Nishanth Aravamudan ------------------------------------------------------------------------ r1269 | ddennedy | 2005-02-27 04:32:29 +0100 (Sun, 27 Feb 2005) | 13 lines Changed paths: M /trunk/dv1394.c Fix a partial conversion to unlocked_ioctl(). drivers/ieee1394/dv1394.c: In function `handle_dv1394_init': drivers/ieee1394/dv1394.c:2529: warning: comparison of distinct pointer types lacks a cast drivers/ieee1394/dv1394.c: In function `handle_dv1394_get_status': drivers/ieee1394/dv1394.c:2559: warning: comparison of distinct pointer types lacks a cast drivers/ieee1394/dv1394.c: In function `dv1394_compat_ioctl': drivers/ieee1394/dv1394.c:2593: warning: unused variable `err' Signed-off-by: Andrew Morton Signed-off-by: Dan Dennedy ------------------------------------------------------------------------ r1268 | scjody | 2005-02-23 23:28:30 +0100 (Wed, 23 Feb 2005) | 10 lines Changed paths: M /trunk/nodemgr.c M /trunk/ohci1394.c Hi! This should fix u32 vs. pm_message_t confusion in firewire. No code changes. Please apply, Pavel Signed-off-by: Pavel Machek Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1267 | scjody | 2005-02-23 23:26:53 +0100 (Wed, 23 Feb 2005) | 6 lines Changed paths: M /trunk/raw1394.c Move hpsb_unregister_protocol, which fixes a hang on rmmod experienced by Parag Warudkar Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1266 | scjody | 2005-02-13 00:40:29 +0100 (Sun, 13 Feb 2005) | 2 lines Changed paths: M /trunk/pcilynx.c Fixed comment; this is also a svn test. ------------------------------------------------------------------------ r1265 | ddennedy | 2005-02-05 20:38:35 +0100 (Sat, 05 Feb 2005) | 5 lines Changed paths: M /trunk/sbp2.c sbp2: add precautionary log notice to new exit branch from last patch Signed-off-by: Stefan Richter ------------------------------------------------------------------------ r1264 | ddennedy | 2005-02-05 20:32:47 +0100 (Sat, 05 Feb 2005) | 5 lines Changed paths: M /trunk/eth1394.c Change the initialization message for eth1394 to KERN_INFO, requested by Steffen Zieger Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1263 | ddennedy | 2005-02-03 20:28:46 +0100 (Thu, 03 Feb 2005) | 4 lines Changed paths: M /trunk/sbp2.c apply patch from Nishanth Aravamudan to use sleep_interruptible for clarity and prevent early return on wait_queue events. ------------------------------------------------------------------------ r1262 | ddennedy | 2005-02-03 19:46:32 +0100 (Thu, 03 Feb 2005) | 6 lines Changed paths: M /trunk/pcilynx.c pcilynx: reduce stack usage in add_card(), from 800 to 308 bytes (on i386), by dynamically allocating struct i2c_adapter i2c_adapter; /* 492 bytes */ Signed-off-by: Randy Dunlap ------------------------------------------------------------------------ r1261 | ddennedy | 2005-02-02 04:32:06 +0100 (Wed, 02 Feb 2005) | 4 lines Changed paths: M /trunk/sbp2.c M /trunk/sbp2.h Fix indentiation and comments to match Linux 2.6 BK. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1260 | ddennedy | 2005-02-02 04:27:22 +0100 (Wed, 02 Feb 2005) | 10 lines Changed paths: M /trunk/nodemgr.c This patch is from Nigel's swsusp2, it kills ugly #include from all over the tree, and makes code slightly nicer. I only left those parts that do not change any code. From: Nigel Cunningham Signed-off-by: Pavel Machek Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1259 | ddennedy | 2005-02-02 04:25:15 +0100 (Wed, 02 Feb 2005) | 4 lines Changed paths: M /trunk/eth1394.c Fix incorrect merge from Linux 2.6 BK (r1235). Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1258 | ddennedy | 2005-02-02 04:22:59 +0100 (Wed, 02 Feb 2005) | 4 lines Changed paths: M /trunk/eth1394.c M /trunk/ohci1394.c Fix whitespace to match Linux 2.6 BK. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1257 | ddennedy | 2005-02-02 04:19:56 +0100 (Wed, 02 Feb 2005) | 7 lines Changed paths: M /trunk/eth1394.c This was missed during the merge to Linux 2.6 BK (r1235). # 2004/08/29 17:50:16-04:00 viro@www.linux.org.uk # [netdrvr eth1394] use netdev_priv Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1256 | ddennedy | 2005-02-02 04:17:18 +0100 (Wed, 02 Feb 2005) | 7 lines Changed paths: M /trunk/dv1394.c Convert dv1394 driver to compat_ioctl Signed-off-by: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1255 | ddennedy | 2005-02-02 04:15:17 +0100 (Wed, 02 Feb 2005) | 7 lines Changed paths: M /trunk/dv1394.c M /trunk/highlevel.c M /trunk/raw1394.c Use the new lock initializers DEFINE_SPIN_LOCK and DEFINE_RW_LOCK Signed-off-by: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1254 | ddennedy | 2005-02-02 04:12:14 +0100 (Wed, 02 Feb 2005) | 5 lines Changed paths: M /trunk/video1394.c Fix bug introduced during merge with Linux 2.6 BK (in r1235). Also fix whitespace to match BK. Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1253 | ddennedy | 2005-02-02 04:10:30 +0100 (Wed, 02 Feb 2005) | 7 lines Changed paths: M /trunk/video1394.c Convert video1394 driver to compat_ioctl Signed-off-by: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1252 | ddennedy | 2005-02-02 04:07:00 +0100 (Wed, 02 Feb 2005) | 5 lines Changed paths: M /trunk/amdtp.c Convert amdtp driver to compat_ioctl Signed-off-by: Andi Kleen Signed-off-by: Jody McIntyre ------------------------------------------------------------------------ r1251 | ddennedy | 2005-01-17 00:06:51 +0100 (Mon, 17 Jan 2005) | 1 line Changed paths: M /trunk/nodemgr.c reorganise LUN handling to resolve oops ------------------------------------------------------------------------