- 24 Jun, 2005 40 commits
-
-
NeilBrown authored
We're running the laundromat work on the default kevent worker thread. But the laundromat takes the nfsv4 state semaphore, which is used for way too much stuff, and the potential for deadlocks is high. Better to have this on a separate workqueue. Signed-off-by:
J. Bruce Fields <bfields@citi.umich.edu> Signed-off-by:
Neil Brown <neilb@cse.unsw.edu.au> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
NeilBrown authored
Minor cleanup. Signed-off-by:
-
NeilBrown authored
rpc_create_client was modified recently to do its own (synchronous) NULL ping of the server. We'd rather do that on our own, asynchronously, so that we don't have to block the nfsd thread doing the probe, and so that setclientid handling (hence, client mounts) can proceed normally whether the callback is succesful or not. (We can still function fine without the callback channel--we just won't be able to give out delegations till it's verified to work.) Signed-off-by:
-
NeilBrown authored
We're trying to read and write from a struct file that we may not hold a reference to any more (since a close could be processed as soon as we drop the state lock). Signed-off-by:
-
NeilBrown authored
Silence another annoying "failed to contact portmap (errno -512)" on shutdown. Signed-off-by:
-
NeilBrown authored
Add a struct kref to each nfs4_file and take a reference to it from each stateid and delegation that refers to it. The atomicity guarantees are overkill given that all this stuff is done under the single nfsd4 state lock, but a) we'd like finer-grained locking some day, and b) this simplifies the cleanup of the structures a bit, something that has previously been a bit complicated and bug-prone. Signed-off-by:
-
NeilBrown authored
Trivial renaming patch: I can never remember, while looking at various lists relating the nfsd4 state structures, which are the "heads" and which are items on other lists, or which structures are actually on the various lists. The following convention helps me: given structures foo and bar, with foo containing the head of a list of bars, use "bars" for the name of the head of the list contained in the struct foo, and use "per_foo" for the entries in the struct bars. Go ahead and do this for struct nfs4_file. Signed-off-by:
-
NeilBrown authored
These remaining debugging counters haven't proved that useful. Signed-off-by:
-
NeilBrown authored
Allocate delegations from a slab cache. Signed-off-by:
-
NeilBrown authored
Allocate stateid's from a slab cache. Signed-off-by:
-
NeilBrown authored
The structures the server uses to keep track of various pieces of nfsv4 state (open files, outstanding delegations, etc.) are likely to be allocated and deallocated frequently and seem reasonable candidates for slab caches. While we're at it, the slab code keeps statistics that help catch leaks and such, so we may as well take this chance to eliminate some debugging counters that we've been keeping ourselves. Start with the struct nfs4_file. Signed-off-by:
-
NeilBrown authored
We currently return err_grace if a user attempts a non-reclaim open during the grace period. But we also need to prevent renames and removes, at least, to ensure clients have the chance to recover state on files before they are moved or deleted. Of course, local users could also do renames and removes during the lease period, and there's not much we can do about that. This at least will help with remote users. Signed-off-by:
-
NeilBrown authored
We're returning NFS4_FH_NOEXPIRE_WITH_OPEN | NFS4_FH_VOL_RENAME for the fh_expire_type attribute. This is incorrect: 1. The spec actually only allows NOEXPIRE_WITH_OPEN when VOLATILE_ANY is also set. 2. Filehandles for open files can expire, if the file is removed and there is a reboot. 3. Filehandles are only volatile on rename in the nosubtree check case. Unfortunately, there's no way to indicate that we only expire on remove. So our only choice is FH4_VOLATILE_ANY. Although it's redundant, we also set FH4_VOL_RENAME in the subtree check case, since subtreecheck does actually cause problems in practice and it seems possibly useful to give clients some way to distinguish that case. Fix a mispelled #define while we're at it. Signed-off-by:
-
NeilBrown authored
Add OPEN claim type NFS4_OPEN_CLAIM_DELEGATE_CUR to nfsd4_open(). A delegation stateid and a name are provided. OPEN with O_CREAT is not legal with this claim type; otherwise, use the NFS4_OPEN_CLAIM_NULL code path to lookup the filename to be opened. Signed-off-by:
Andy Adamson <andros@citi.umich.edu> Signed-off-by:
-
NeilBrown authored
State logic for OPEN with claim type CLAIM_DELEGATE_CUR, which the NFSv4 client uses to report local OPENs on a delegated file back to the NFSv4 server. nfs4_check_deleg() performs input delegation stateid lookup and sanity check. Signed-off-by:
-
NeilBrown authored
We don't really need to be doing a separate open for every stateid. And in the case of an open from a client that already has a delegation on a file, it unnecessarily results in a delegation recall. Signed-off-by:
-
NeilBrown authored
Additional minor code reshuffling to prepare for claim_deleg_cur support. Signed-off-by:
-
NeilBrown authored
Factor out a bit of common code that will be useful elsewhere. Signed-off-by:
-
John Levon authored
The below patch passes samples from anonymous regions to userspace instead of just dropping them. This provides the support needed for reporting anonymous-region code samples (today: basic accumulated results; later: Java and other dynamically compiled code). As this changes the format, an upgrade to the just-released 0.9 release of the userspace tools is required. This patch is based upon an earlier one by Will Cohen <wcohen@redhat.com> Signed-off-by:
John Levon <levon@movementarian.org> Signed-off-by:
-
Mauro Carvalho Chehab authored
tuner-core.c, tuner.h: - tuner-core changed to support multiple I2C devices used on some adapters; - Kconfig now has an option (CONFIG_TUNER_MULTI_I2C) to enable this new behavor; - By default, even enabling CONFIG_TUNER_MULTI_I2C, tuner-core emulates the old behavor, using first I2C device for both FM and TV; - There is a new i2c command (TUNER_SET_ADDR) to allow tuner clients to select I2C address for FM or TV tuner; - Tuner I2C dettach now generates a warning on syslog if failed. tuner-simple.c: - TVision TVF-8531MF and TVF-5533 MF tuner included. It uses, by default, I2C on 0xC2 address for TV and on 0xC0 for Radio. Both TV and FM Radio mode are working. Signed-off-by:
Mauro Carvalho Chehab <mchehab@brturbo.com.br> Signed-off-by:
-
Johannes Stezenbach authored
Added support for the Broadcom BCM3510 ATSC (8VSB/16VSB & ITU J83 AnnexB FEC QAM64/256) demodulator used in the first generation of Air2PC ATSC PCI-cards/USB-boxes made by B2C2. Signed-off-by:
Patrick Boettcher <pb@linuxtv.org> Signed-off-by:
Johannes Stezenbach <js@linuxtv.org> Signed-off-by:
-
Johannes Stezenbach authored
Use time_after() macro. Signed-off-by:
Marcelo Feitoza Parisi <marcelo@feitoza.com.br> Signed-off-by:
Domen Puncer <domen@coderock.org> Signed-off-by:
-
Johannes Stezenbach authored
Fix error checking during initialization. Thanks to Gerolf Wendland for discovering. Signed-off-by:
-
Johannes Stezenbach authored
Add generalized dvb-usb driver which supports a wide variety of devices. Signed-off-by:
Randy Dunlap <rdunlap@xenotime.net> Signed-off-by:
-
Johannes Stezenbach authored
Remove the dibusb driver which has been obsoleted by the generalized dvb-usb driver. Signed-off-by:
-
Yani Ioannou authored
This patch updates all the device attribute callbacks that weren't updated with the new parameter, I guess because they weren't in Greg's tree (including drivers/pcmcia/ds.c). Without the patch these callbacks are probably broken (and generate a warning along the lines of "assignment from incompatible pointer type"). Please see http://lkml.org/lkml/2005/5/19/40 for the scripts I used to update the attributes automatically. Signed-off-by:
Yani Ioannou <yani.ioannou@gmail.com> Signed-off-by:
-
Andrew Morton authored
LD .tmp_vmlinux1 drivers/built-in.o: In function `i2o_cfg_parms': config-osm.c:(.text+0x12764a): undefined reference to `i2o_parm_issue' Cc: Markus Lidel <Markus.Lidel@shadowconnect.com> Signed-off-by:
-
Markus Lidel authored
Set max sectors to 256 for Promise controllers. Signed-off-by:
Markus Lidel <Markus.Lidel@shadowconnect.com> Signed-off-by:
-
Markus Lidel authored
Lindent run and replaced printk() through the corresponding osm_*() function Signed-off-by:
-
Markus Lidel authored
Changes: - Added header "core.h" for i2o_core.ko internal definitions - More sparse fixes - Changed display of TID's in sysfs attributes from XXX to 0xXXX - Use the right functions for accessing I/O and normal memory - Removed error handling of SCSI device errors and let the SCSI layer take care of it - Added new device / removed device handling to SCSI-OSM - Make status access volatile - Cleaned up activation of I2O controller - Removed unnecessary wmb() and rmb() calls - Use own struct i2o_io for I/O memory instead of struct i2o_dma Signed-off-by:
-
Markus Lidel authored
Changes: - Provide SG_IO access to BLOCK and EXECUTIVE class on Adaptec controllers - Use PRIVATE messages in SCSI-OSM because on some controllers normal SCSI class commands like READ or READ CAPACITY cause errors - Use new DMA and SG list creation function - Added workaround to limit sectors per request for Adaptec 2400A controllers Signed-off-by:
-
Markus Lidel authored
Changes: - Added Bus-OSM which could be used by user space programs to reset a channel on the controller - Make ioctl's in Config-OSM obsolete in prefer for sysfs attributes and move those to its own file - Added sysfs attribute for firmware read and write access for I2O controllers - Added special handling of firmware read and write access for Adaptec controllers - Added vendor id and product id as sysfs-attribute to Executive classes - Added automatic notification of LCT change handling to Exec-OSM - Added flushing function to Block-OSM for later barrier implementation - Use PRIVATE messages for Block access on Adaptec controllers, which are faster then BLOCK class access - Cleaned up support for Promise controller - New messages are now detected using the IRQ status register as suggested by the I2O spec - Added i2o_dma_high() and i2o_dma_low() functions - Added facility for SG tablesize calculation when using 32-bit and 64-bit DMA addresses - Added i2o_dma_map_single() and i2o_dma_map_sg() which could build the SG list for 32-bit as well as 64-bit DMA addresses Signed-off-by:
-
Markus Lidel authored
Changes: - Removed unnecessary checking of NULL before calling kfree() - Make some functions static - Changed pr_debug() into osm_debug() - Use i2o_msg_in_to_virt() for getting a pointer to the message frame - Cleaned up some comments - Changed some le32_to_cpu() into readl() where necessary - Make error messages of OSM's look the same - Cleaned up error handling in i2o_block_end_request() - Removed unused error handling of failed messages in Block-OSM, which are not allowed by the I2O spec - Corrected the blocksize detection in i2o_block - Added hrt and lct sysfs-attribute to controller - Call done() function in SCSI-OSM after freeing DMA buffers - Removed unneeded variable for message size calculation in i2o_scsi_queuecommand() - Make some changes to remove sparse warnings - Reordered some functions - Cleaned up controller initialization - Replaced some magic numbers by defines - Removed unnecessary dma_sync_single_for_cpu() call on coherent DMA - Removed some unused fields in i2o_controller and removed some unused functions Signed-off-by:
-
Markus Lidel authored
Changes: - Fixed sysfs bug where user and parent links where added to the I2O device itself - Fixed bug when calculating TID for the event handler and cleaned up the workflow of i2o_driver_dispatch() - Fixed oops when no I2O device could be found for an event delivered to Exec-OSM - Fixed initialization of spinlock in Exec-OSM - Fixed memory leak in i2o_cfg_passthru() and i2o_cfg_passthru() - Removed MTRR support - Added PCI ID of Promise SX6000 with firmware >= 1.20.x.x - Turn of caching for ioremapped memory of in_queue - Added initialization sequence for Promise controllers - Moved definition of u8 / u16 / u32 for raidutils before first use Signed-off-by:
-
Kylene Jo Hall authored
Since after reconsideration this is more debug output than an error (the TPM is operating correctly given the current state) I have changed the statements to dbg rather than err. Also this patch corrects a memory leak if the error path is taken in the tpm_show_pubek function. Signed-off-by:
Kylene Hall <kjhall@us.ibm.com> Signed-off-by:
-
Kylene Hall authored
Signed-off-by:
-
Kylene Hall authored
Realized the tpm_lpc_init function isn't really necessary. Replaced it with vendor specific logic to find out the address the BIOS mapped the TPM to. This patch removes the tpm_lpc_init function, enums associated with it and calls to it. The patch also implements the replacement functionality. Signed-off-by:
-
Kylene Hall authored
Add support for TPMs on additional LPC buses. Signed-off-by:
-
Kylene Hall authored
Add a missing lock in the register hardware and fix a misplaced lock release release. Signed-off-by:
-
Kylene Hall authored
This patch provides the logic to check if an operation has been canceled while waiting for the response to arrive. Signed-off-by:
-
