- 05 Oct, 2012 40 commits
-
-
Linus Torvalds authored
Merge misc patches from Andrew Morton: "The MM tree is rather stuck while I wait to find out what the heck is happening with sched/numa. Probably I'll need to route around all the code which was added to -next, sigh. So this is "everything else", or at least most of it - other small bits are still awaiting resolutions of various kinds." * emailed patches from Andrew Morton <akpm@linux-foundation.org>: (180 commits) lib/decompress.c add __init to decompress_method and data kernel/resource.c: fix stack overflow in __reserve_region_with_split() omfs: convert to use beXX_add_cpu() taskstats: cgroupstats_user_cmd() may leak on error aoe: update aoe-internal version number to 50 aoe: update documentation to better reflect aoe-plus-udev usage aoe: remove unused code aoe: make dynamic block minor numbers the default aoe: update and specify AoE address guards and error messages aoe: retain static block device numbers for backwards compatibility aoe: support more AoE addresses with dynamic block device minor numbers aoe: update documentation with new URL and VM settings reference aoe: update copyright year in touched files aoe: update internal version number to 49 aoe: remove unused code and add cosmetic improvements aoe: increase net_device reference count while using it aoe: associate frames with the AoE storage target aoe: disallow unsupported AoE minor addresses aoe: do revalidation steps in order aoe: failover remote interface based on aoe_deadsecs parameter ...
-
Hein Tibosch authored
Fix the warning: WARNING: vmlinux.o(.text+0x14cfd8): Section mismatch in reference from the variable compressed_formats to the function .init.text:gunzip() The function compressed_formats() references the function __init gunzip(). etc.. Within decompress.c, compressed_formats[] needs 'a __initdata annotation', because some of it's data members refer to functions which will be unloaded after init. Consequently, its user decompress_method() will get the __init prefix. Signed-off-by:
Hein Tibosch <hein_tibosch@yahoo.es> Cc: Albin Tonnerre <albin.tonnerre@free-electrons.com> Cc: Phillip Lougher <phillip@lougher.demon.co.uk> Cc: "H. Peter Anvin" <hpa@zytor.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
T Makphaibulchoke authored
Using a recursive call add a non-conflicting region in __reserve_region_with_split() could result in a stack overflow in the case that the recursive calls are too deep. Convert the recursive calls to an iterative loop to avoid the problem. Tested on a machine containing 135 regions. The kernel no longer panicked with stack overflow. Also tested with code arbitrarily adding regions with no conflict, embedding two consecutive conflicts and embedding two non-consecutive conflicts. Signed-off-by:
T Makphaibulchoke <tmac@hp.com> Reviewed-by:
Ram Pai <linuxram@us.ibm.com> Cc: Paul Gortmaker <paul.gortmaker@gmail.com> Cc: Wei Yang <weiyang@linux.vnet.ibm.com> Signed-off-by:
-
Wei Yongjun authored
Convert cpu_to_beXX(beXX_to_cpu(E1) + E2) to use beXX_add_cpu(). dpatch engine is used to auto generate this patch. (https://github.com/weiyj/dpatch) Signed-off-by:
Wei Yongjun <yongjun_wei@trendmicro.com.cn> Acked-by:
Bob Copeland <me@bobcopeland.com> Signed-off-by:
-
Jesper Juhl authored
If prepare_reply() succeeds we have allocated memory for 'rep_skb'. If nla_reserve() then subsequently fails and returns NULL we fail to release the memory we allocated, thus causing a leak. Signed-off-by:
Jesper Juhl <jj@chaosbits.net> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by:
-
Ed Cashin authored
Signed-off-by:
Ed Cashin <ecashin@coraid.com> Signed-off-by:
-
Ed Cashin authored
Signed-off-by:
-
Ed Cashin authored
Signed-off-by:
-
Ed Cashin authored
Because udev use is so widespread, making the old static mapping the default is too conservative, given the severe limitations it places on usable AoE addresses. Storage virtualization and larger shelves have made the old limitations too confining. These changes make the dynamic block device minor numbers the default, removing the limitations on usable AoE addresses. The static arrangement is still available with aoe_dyndevs=0, and the aoe-stat tool from the userland aoetools package, the user space counterpart to the aoe driver, recognizes the case where there is a mismatch between the minor number in sysfs and the minor number in a special device file. Signed-off-by:
-
Ed Cashin authored
In general, specific is better when it comes to messages about AoE usage problems. Also, explicit checks for the AoE broadcast addresses are added. Signed-off-by:
-
Ed Cashin authored
The old mapping between AoE target shelf and slot addresses and the block device minor number is retained as a backwards-compatible feature, with a new "aoe_dyndevs" module parameter available for enabling dynamic block device minor numbers. Signed-off-by:
-
Ed Cashin authored
The ATA over Ethernet protocol uses a major (shelf) and minor (slot) address to identify a particular storage target. These changes remove an artificial limitation the aoe driver imposes on the use of AoE addresses. For example, without these changes, the slot address has a maximum of 15, but users commonly use slot numbers much greater than that. The AoE shelf and slot address space is often used sparsely. Instead of using a static mapping between AoE addresses and the block device minor number, the block device minor numbers are now allocated on demand. Signed-off-by:
-
Ed Cashin authored
The old area has a new URL. Also, now that the driver can perform better, it is worth mentioning the VM settings that help aoe to sink dirty pages out early, avoiding unecessary memory pressure when much I/O is going on. Signed-off-by:
-
Ed Cashin authored
Signed-off-by:
-
Ed Cashin authored
The internal version number of the aoe driver appears in a console message when the driver loads and is usually obtained by the user with the userland aoe-version tool, part of the aoetools.[1] Although this patchset includes bugfixes backported from higher-numbered versions published on the coraid.com website, it is a form of version 49. 1. http://aoetools.sourceforge.net/Signed-off-by:
-
Ed Cashin authored
This change removes some unused code and attempts to increase code consistency. Signed-off-by:
-
Ed Cashin authored
This change eliminates the danger that the user could rmmod the driver for a network interface that is being used for AoE by the aoe driver. Signed-off-by:
-
Ed Cashin authored
In the driver code, "target" and aoetgt refer to a particular remote interface on the AoE storage target. The latter is identified by its AoE major and minor addresses. Commands that are being sent to an AoE storage target {major, minor} can be sent or retransmitted to any of the remote MAC addresses associated with the AoE storage target. That is, frames are naturally associated with not an aoetgt (AoE major, AoE minor, remote MAC address) but an aoedev (AoE major, AoE minor). Making the code reflect that reality simplifies the driver, especially when the path to a remote MAC address becomes unusable. Signed-off-by: -
Ed Cashin authored
A guard is inserted to prevent AoE minor addresses (slot addresses) higher than 15 to be used, as they are not yet supported by the driver. There is a change coming that will allow the aoe driver to overcome this limit by using system device minor numbers dynamically, but until then, this guard prevents unexpected targets from being used by the driver when AoE targets with high minor numbers are on the AoE network. Signed-off-by:
-
Ed Cashin authored
The discovery process begins with an optional AoE config query command and an AoE config query response. Normally when an aoe device is already open, the config query response does not trigger an ATA identify device command to be sent out, since the response contains storage capacity information that, if changed, could surprise the user of the device. The userland "aoe-revalidate" tool uses a character device to trigger an AoE config query for a particular AoE storage target and an ATA device identify command, even when the device is open. This change causes the config query to go out first, reflecting the normal discovery sequence. The responses could come back in any order, so this change is fairly cosmetic. Signed-off-by:
-
Ed Cashin authored
The aoe_deadsecs module parameter allows the user to specify a hard limit on the number of seconds an AoE command can be retransmitted before the AoE block device is considered to have failed. Using aoe_deadsecs to determine the time we try using a different remote interface helps to ensure that the hard limit is not reached before we've tried to recover by sending to a different remote port. As a data storage target, the AoE target is unambiguously identified by its {major, minor} AoE address tuple, and an AoE target can have multiple MAC addresses. However, note that "target" in the driver code and comments means a {major, minor, MAC address} tuple, as in "somewhere to send packets". Signed-off-by: -
Ed Cashin authored
Users with several network interfaces dedicated to AoE generally do not configure them to support different-sized AoE data payloads on purpose. For a given AoE target, there will be a set of local network interfaces that can reach it. Using only the payload that will fit in the smallest-sized MTU of all those local interfaces greatly simplifies the driver, especially in failure scenarios. Signed-off-by:
-
Ed Cashin authored
The dev_queue_xmit function needs to have interrupts enabled, so the most simple way to get the locking right but still fulfill that requirement is to use a process that can call dev_queue_xmit serially over queued transmissions. Signed-off-by:
-
Ed Cashin authored
To allow users to choose an elevator algorithm for their particular workloads, change from a make_request-style driver to an I/O-request-queue-handler-style driver. We have to do a couple of things that might be surprising. We manipulate the page _count directly on the assumption that we still have no guarantee that users of the block layer are prohibited from submitting bios containing pages with zero reference counts.[1] If such a prohibition now exists, I can get rid of the _count manipulation. Just as before this patch, we still keep track of the sk_buffs that the network layer still hasn't finished yet and cap the resources we use with a "pool" of skbs.[2] Now that the block layer maintains the disk stats, the aoe driver's diskstats function can go away. 1. https://lkml.org/lkml/2007/3/1/374 2. https://lkml.org/lkml/2007/7/6/241Signed-off-by:
-
Ed Cashin authored
Make the frames the aoe driver uses to track the relationship between bios and packets more flexible and detached, so that they can be passed to an "aoe_ktio" thread for completion of I/O. The frames are handled much like skbs, with a capped amount of preallocation so that real-world use cases are likely to run smoothly and degenerate gracefully even under memory pressure. Decoupling I/O completion from the receive path and serializing it in a process makes it easier to think about the correctness of the locking in the driver, especially in the case of a remote MAC address becoming unusable. [dan.carpenter@oracle.com: cleanup an allocation a bit] Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Ed Cashin authored
tAdd adds the ability to work with large packets composed of a number of segments, using the scatter gather feature of the block layer (biovecs) and the network layer (skb frag array). The motivation is the performance gained by using a packet data payload greater than a page size and by using the network card's scatter gather feature. Users of the out-of-tree aoe driver already had these changes, but since early 2011, they have complained of increased memory utilization and higher CPU utilization during heavy writes.[1] The commit below appears related, as it disables scatter gather on non-IP protocols inside the harmonize_features function, even when the NIC supports sg. commit f01a5236 Author: Jesse Gross <jesse@nicira.com> Date: Sun Jan 9 06:23:31 2011 +0000 net offloading: Generalize netif_get_vlan_features(). With that regression in place, transmits always linearize sg AoE packets, but in-kernel users did not have this patch. Before 2.6.38, though, these changes were working to allow sg to increase performance. 1. http://www.spinics.net/lists/linux-mm/msg15184.htmlSigned-off-by:
-
Paul Clements authored
Add discard support to nbd. If the nbd-server supports discard, it will send NBD_FLAG_SEND_TRIM to the client. The client will then set the flag in the kernel via NBD_SET_FLAGS, which tells the kernel to enable discards for the device (QUEUE_FLAG_DISCARD). If discard support is enabled, then when the nbd client system receives a discard request, this will be passed along to the nbd-server. When the discard request is received by the nbd-server, it will perform: fallocate(.. FALLOC_FL_PUNCH_HOLE ..) To punch a hole in the backend storage, which is no longer needed. Signed-off-by:
Paul Clements <paul.clements@steeleye.com> Signed-off-by:
-
Paul Clements authored
Add a set-flags ioctl, allowing various option flags to be set on an nbd device. This allows the nbd-client to set the device flags (to enable read-only mode, or enable discard support, etc.). Flags are typically specified by the nbd-server. During the negotiation phase of the nbd connection, the server sends its flags to the client. The client then uses NBD_SET_FLAGS to inform the kernel of the options. Also included is a one-line fix to debug output for the set-timeout ioctl. Signed-off-by:
-
Alexandre Bounine authored
Replace the single global destination ID counter with per-net allocation mechanism to allow independent destID management for each available RapidIO network. Using bitmap based mechanism instead of counters allows destination ID release and reuse in systems that support hot-swap. Signed-off-by:
Alexandre Bounine <alexandre.bounine@idt.com> Cc: Matt Porter <mporter@kernel.crashing.org> Cc: Li Yang <leoli@freescale.com> Signed-off-by:
-
Alexandre Bounine authored
Make RIONET driver multi-net safe/capable by introducing per-net lists of RapidIO network peers. Rework registration of network adapters to support all available RIO master port devices. Signed-off-by:
-
Alexandre Bounine authored
Modify mport initialization routine to run the RapidIO discovery process asynchronously. This allows to have an arbitrary order of enumerating and discovering ports in systems with multiple RapidIO controllers without creating a deadlock situation if enumerator port is registered after a discovering one. Making netID matching to mportID ensures consistent net ID assignment in multiport RapidIO systems with asynchronous discovery process (global counter implementation is affected by race between threads). [akpm@linux-foundation.org: tweak code layput] Signed-off-by:
-
Alexandre Bounine authored
Modify handling of device lists to resolve issues caused by using single global list of RIO devices during enumeration/discovery. The most common sign of existing issue is incorrect contents of switch routing tables in systems with multiple mport controllers while single-port configuration performs as expected. Signed-off-by:
-
Alexandre Bounine authored
The following set of patches provides modifications targeting support of multiple RapidIO master port (mport) devices on a CPU-side of RapidIO-capable board. While the RapidIO subsystem code has definitions suitable for multi-controller/multi-net support, the existing implementation cannot be considered ready for multiple mport configurations. =========== NOTES: ============= a) The patches below do not address RapidIO side view of multiport processing elements defined in Part 6 of RapidIO spec Rev.2.1 (section 6.4.1). These devices have Base Device ID CSR (0x60) and Component Tag CSR (0x6C) shared by all SRIO ports. For example, Freescale's P4080, P3041 and P5020 have a dual-port SRIO controller implemented according the specification. Enumeration/discovery of such devices from RapidIO side may require device-specific fixups. b) Devices referenced above may also require implementation specific code to setup a host device ID for mport device. These operations are not addressed by patches in this package. ================================= Details about provided patches: 1. Fix blocking wait for discovery ready While it does not happen on PowerPC based platforms, there is possibility of stalled CPU warning dump on x86 based platforms that run RapidIO discovery process if they wait too long for being enumerated. Currently users can avoid it by disabling the soft-lockup detector using "nosoftlockup" kernel parameter OR by ensuring that enumeration is completed before soft-lockup is detected. This patch eliminates blocking wait and keeps a scheduler running. It also is required for patch 3 below which introduces asynchronous discovery process. 2. Use device lists handling on per-net basis This patch allows to correctly support multiple RapidIO nets and resolves possible issues caused by using single global list of devices during RapidIO system enumeration/discovery. The most common sign of existing issue is incorrect contents of switch routing tables in systems with multiple mport controllers while single-port configuration performs as expected. The patch does not eliminate the global RapidIO device list but changes some routines in enumeration/discovery to use per-net device lists instead. This way compatibility with upper layer RIO routines is preserved. 3. Run discovery as an asynchronous process This patch modifies RapidIO initialization routine to asynchronously run the discovery process for each corresponding mport. This allows having an arbitrary order of enumerating and discovering mports without creating a deadlock situation if an enumerator port was registered after a discovering one. On boards with multiple discovering mports it also eliminates order dependency between mports and may reduce total time of RapidIO subsystem initialization. Making netID matching to mportID ensures consistent netID assignment in multiport RapidIO systems with asynchronous discovery process (global counter implementation is affected by race between threads). 4. Rework RIONET to support multiple RIO master ports In the current version of the driver rionet_probe() has comment "XXX Make multi-net safe". Now it is a good time to address this comment. This patch makes RIONET driver multi-net safe/capable by introducing per-net lists of RapidIO network peers. It also enables to register network adapters for all available mport devices. 5. Add destination ID allocation mechanism The patch replaces a single global destination ID counter with per-net allocation mechanism to allow independent destID management for each available RapidIO network. Using bitmap based mechanism instead of counters allows destination ID release and reuse in systems that support hot-swap. This patch: Fix blocking wait loop in the RapidIO discovery routine to avoid warning dumps about stalled CPU on x86 platforms. Signed-off-by:
-
Alexandre Bounine authored
Apply port RX/TX enable operations only to active switch ports. RapidIO specification (Part 6: LP-Serial Physical Layer) recommends to keep Output Port Enable (TX) and Input Port Enable (RX) control bits in disabled state (0b0) after device reset. It also allows to have implementation specific reset state for these bits. This patch ensures that TX/RX enable action is applied only to active switch's ports while preserving an initial state of inactive ones. This patch is intended to keep inactive switch ports with inbound and outbound packet transfers disabled to block unexpected packets during hot insertion event. While it does not fix any visible malfunction it is intended to prevent such events in future. Signed-off-by:
-
Alexandre Bounine authored
Add Tsi721 routines to support RapidIO subsystem's inbound memory mapping interface (RapidIO to system's local memory). Signed-off-by:
-
Alexandre Bounine authored
Add common inbound memory mapping/unmapping interface. This allows to make local memory space accessible from the RapidIO side using hardware mapping capabilities of RapidIO bridging devices. The new interface is intended to enable data transfers between RapidIO devices in combination with DMA engine support. This patch is based on patch submitted by Li Yang <leoli@freescale.com> (https://lists.ozlabs.org/pipermail/linuxppc-dev/2009-April/071210.html) Signed-off-by:
-
Peter Senna Tschudin authored
Convert a nonnegative error return code to a negative one, as returned elsewhere in the function. A simplified version of the semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> ( if@p1 (\(ret < 0\|ret != 0\)) { ... return ret; } | ret@p1 = 0 ) ... when != ret = e1 when != &ret *if(...) { ... when != ret = e2 when forall return ret; } // </smpl> Signed-off-by:
Peter Senna Tschudin <peter.senna@gmail.com> Acked-by:
-
Alexandre Bounine authored
Signed-off-by:
Robert P. J. Day <rpjday@crashcourse.ca> Cc: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by:
-
Alexandre Bounine authored
Modify RapidIO mport device name assignment to include device name of PCIe side of Tsi721 bridge. The new name format is intended to provide definitive reference between RapidIO and PCIe sides of the bridge in systems with multiple Tsi721 bridges. Signed-off-by:
-
Alexandre Bounine authored
Fix multicast packet transmit logic to account for repetitive transmission of single skb: - correct check for available buffers (this bug may produce NULL pointer crash dump in case of heavy traffic); - update skb user count (incorrect user counter causes a warning dump from net_tx_action routine during multicast transfers in systems with three or more rionet participants). Signed-off-by:
-
