From 9bc89cd82d6f88fb0ca39b30445c329a430fd66b Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Tue, 2 Jan 2007 11:10:44 -0700 Subject: async_tx: add the async_tx api The async_tx api provides methods for describing a chain of asynchronous bulk memory transfers/transforms with support for inter-transactional dependencies. It is implemented as a dmaengine client that smooths over the details of different hardware offload engine implementations. Code that is written to the api can optimize for asynchronous operation and the api will fit the chain of operations to the available offload resources. I imagine that any piece of ADMA hardware would register with the 'async_*' subsystem, and a call to async_X would be routed as appropriate, or be run in-line. - Neil Brown async_tx exploits the capabilities of struct dma_async_tx_descriptor to provide an api of the following general format: struct dma_async_tx_descriptor * async_(..., struct dma_async_tx_descriptor *depend_tx, dma_async_tx_callback cb_fn, void *cb_param) { struct dma_chan *chan = async_tx_find_channel(depend_tx, ); struct dma_device *device = chan ? chan->device : NULL; int int_en = cb_fn ? 1 : 0; struct dma_async_tx_descriptor *tx = device ? device->device_prep_dma_(chan, len, int_en) : NULL; if (tx) { /* run asynchronously */ ... tx->tx_set_dest(addr, tx, index); ... tx->tx_set_src(addr, tx, index); ... async_tx_submit(chan, tx, flags, depend_tx, cb_fn, cb_param); } else { /* run synchronously */ ... ... async_tx_sync_epilog(flags, depend_tx, cb_fn, cb_param); } return tx; } async_tx_find_channel() returns a capable channel from its pool. The channel pool is organized as a per-cpu array of channel pointers. The async_tx_rebalance() routine is tasked with managing these arrays. In the uniprocessor case async_tx_rebalance() tries to spread responsibility evenly over channels of similar capabilities. For example if there are two copy+xor channels, one will handle copy operations and the other will handle xor. In the SMP case async_tx_rebalance() attempts to spread the operations evenly over the cpus, e.g. cpu0 gets copy channel0 and xor channel0 while cpu1 gets copy channel 1 and xor channel 1. When a dependency is specified async_tx_find_channel defaults to keeping the operation on the same channel. A xor->copy->xor chain will stay on one channel if it supports both operation types, otherwise the transaction will transition between a copy and a xor resource. Currently the raid5 implementation in the MD raid456 driver has been converted to the async_tx api. A driver for the offload engines on the Intel Xscale series of I/O processors, iop-adma, is provided in a later commit. With the iop-adma driver and async_tx, raid456 is able to offload copy, xor, and xor-zero-sum operations to hardware engines. On iop342 tiobench showed higher throughput for sequential writes (20 - 30% improvement) and sequential reads to a degraded array (40 - 55% improvement). For the other cases performance was roughly equal, +/- a few percentage points. On a x86-smp platform the performance of the async_tx implementation (in synchronous mode) was also +/- a few percentage points of the original implementation. According to 'top' on iop342 CPU utilization drops from ~50% to ~15% during a 'resync' while the speed according to /proc/mdstat doubles from ~25 MB/s to ~50 MB/s. The tiobench command line used for testing was: tiobench --size 2048 --block 4096 --block 131072 --dir /mnt/raid --numruns 5 * iop342 had 1GB of memory available Details: * if CONFIG_DMA_ENGINE=n the asynchronous path is compiled away by making async_tx_find_channel a static inline routine that always returns NULL * when a callback is specified for a given transaction an interrupt will fire at operation completion time and the callback will occur in a tasklet. if the the channel does not support interrupts then a live polling wait will be performed * the api is written as a dmaengine client that requests all available channels * In support of dependencies the api implicitly schedules channel-switch interrupts. The interrupt triggers the cleanup tasklet which causes pending operations to be scheduled on the next channel * Xor engines treat an xor destination address differently than a software xor routine. To the software routine the destination address is an implied source, whereas engines treat it as a write-only destination. This patch modifies the xor_blocks routine to take a an explicit destination address to mirror the hardware. Changelog: * fixed a leftover debug print * don't allow callbacks in async_interrupt_cond * fixed xor_block changes * fixed usage of ASYNC_TX_XOR_DROP_DEST * drop dma mapping methods, suggested by Chris Leech * printk warning fixups from Andrew Morton * don't use inline in C files, Adrian Bunk * select the API when MD is enabled * BUG_ON xor source counts <= 1 * implicitly handle hardware concerns like channel switching and interrupts, Neil Brown * remove the per operation type list, and distribute operation capabilities evenly amongst the available channels * simplify async_tx_find_channel to optimize the fast path * introduce the channel_table_initialized flag to prevent early calls to the api * reorganize the code to mimic crypto * include mm.h as not all archs include it in dma-mapping.h * make the Kconfig options non-user visible, Adrian Bunk * move async_tx under crypto since it is meant as 'core' functionality, and the two may share algorithms in the future * move large inline functions into c files * checkpatch.pl fixes * gpl v2 only correction Cc: Herbert Xu Signed-off-by: Dan Williams Acked-By: NeilBrown --- drivers/dma/Kconfig | 5 ++--- drivers/md/Kconfig | 3 ++- drivers/md/raid5.c | 54 ++++++++++++++++++++++++++--------------------------- 3 files changed, 31 insertions(+), 31 deletions(-) (limited to 'drivers') diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig index 72be6c63edfc..492aa080562d 100644 --- a/drivers/dma/Kconfig +++ b/drivers/dma/Kconfig @@ -8,8 +8,8 @@ menu "DMA Engine support" config DMA_ENGINE bool "Support for DMA engines" ---help--- - DMA engines offload copy operations from the CPU to dedicated - hardware, allowing the copies to happen asynchronously. + DMA engines offload bulk memory operations from the CPU to dedicated + hardware, allowing the operations to happen asynchronously. comment "DMA Clients" @@ -31,5 +31,4 @@ config INTEL_IOATDMA default m ---help--- Enable support for the Intel(R) I/OAT DMA engine. - endmenu diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig index 24d93d02a1f3..bfd9b9c6252c 100644 --- a/drivers/md/Kconfig +++ b/drivers/md/Kconfig @@ -109,7 +109,8 @@ config MD_RAID10 config MD_RAID456 tristate "RAID-4/RAID-5/RAID-6 mode" depends on BLK_DEV_MD - select XOR_BLOCKS + select ASYNC_MEMCPY + select ASYNC_XOR ---help--- A RAID-5 set of N drives with a capacity of C MB per drive provides the capacity of C * (N - 1) MB, and protects against a failure diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index 5adbe0b22684..4f51dfa8e487 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -916,25 +916,25 @@ static void copy_data(int frombio, struct bio *bio, } } -#define check_xor() do { \ - if (count == MAX_XOR_BLOCKS) { \ - xor_blocks(count, STRIPE_SIZE, ptr); \ - count = 1; \ - } \ +#define check_xor() do { \ + if (count == MAX_XOR_BLOCKS) { \ + xor_blocks(count, STRIPE_SIZE, dest, ptr);\ + count = 0; \ + } \ } while(0) static void compute_block(struct stripe_head *sh, int dd_idx) { int i, count, disks = sh->disks; - void *ptr[MAX_XOR_BLOCKS], *p; + void *ptr[MAX_XOR_BLOCKS], *dest, *p; PRINTK("compute_block, stripe %llu, idx %d\n", (unsigned long long)sh->sector, dd_idx); - ptr[0] = page_address(sh->dev[dd_idx].page); - memset(ptr[0], 0, STRIPE_SIZE); - count = 1; + dest = page_address(sh->dev[dd_idx].page); + memset(dest, 0, STRIPE_SIZE); + count = 0; for (i = disks ; i--; ) { if (i == dd_idx) continue; @@ -948,8 +948,8 @@ static void compute_block(struct stripe_head *sh, int dd_idx) check_xor(); } - if (count != 1) - xor_blocks(count, STRIPE_SIZE, ptr); + if (count) + xor_blocks(count, STRIPE_SIZE, dest, ptr); set_bit(R5_UPTODATE, &sh->dev[dd_idx].flags); } @@ -957,14 +957,14 @@ static void compute_parity5(struct stripe_head *sh, int method) { raid5_conf_t *conf = sh->raid_conf; int i, pd_idx = sh->pd_idx, disks = sh->disks, count; - void *ptr[MAX_XOR_BLOCKS]; + void *ptr[MAX_XOR_BLOCKS], *dest; struct bio *chosen; PRINTK("compute_parity5, stripe %llu, method %d\n", (unsigned long long)sh->sector, method); - count = 1; - ptr[0] = page_address(sh->dev[pd_idx].page); + count = 0; + dest = page_address(sh->dev[pd_idx].page); switch(method) { case READ_MODIFY_WRITE: BUG_ON(!test_bit(R5_UPTODATE, &sh->dev[pd_idx].flags)); @@ -987,7 +987,7 @@ static void compute_parity5(struct stripe_head *sh, int method) } break; case RECONSTRUCT_WRITE: - memset(ptr[0], 0, STRIPE_SIZE); + memset(dest, 0, STRIPE_SIZE); for (i= disks; i-- ;) if (i!=pd_idx && sh->dev[i].towrite) { chosen = sh->dev[i].towrite; @@ -1003,9 +1003,9 @@ static void compute_parity5(struct stripe_head *sh, int method) case CHECK_PARITY: break; } - if (count>1) { - xor_blocks(count, STRIPE_SIZE, ptr); - count = 1; + if (count) { + xor_blocks(count, STRIPE_SIZE, dest, ptr); + count = 0; } for (i = disks; i--;) @@ -1037,9 +1037,9 @@ static void compute_parity5(struct stripe_head *sh, int method) check_xor(); } } - if (count != 1) - xor_blocks(count, STRIPE_SIZE, ptr); - + if (count) + xor_blocks(count, STRIPE_SIZE, dest, ptr); + if (method != CHECK_PARITY) { set_bit(R5_UPTODATE, &sh->dev[pd_idx].flags); set_bit(R5_LOCKED, &sh->dev[pd_idx].flags); @@ -1132,7 +1132,7 @@ static void compute_parity6(struct stripe_head *sh, int method) static void compute_block_1(struct stripe_head *sh, int dd_idx, int nozero) { int i, count, disks = sh->disks; - void *ptr[MAX_XOR_BLOCKS], *p; + void *ptr[MAX_XOR_BLOCKS], *dest, *p; int pd_idx = sh->pd_idx; int qd_idx = raid6_next_disk(pd_idx, disks); @@ -1143,9 +1143,9 @@ static void compute_block_1(struct stripe_head *sh, int dd_idx, int nozero) /* We're actually computing the Q drive */ compute_parity6(sh, UPDATE_PARITY); } else { - ptr[0] = page_address(sh->dev[dd_idx].page); - if (!nozero) memset(ptr[0], 0, STRIPE_SIZE); - count = 1; + dest = page_address(sh->dev[dd_idx].page); + if (!nozero) memset(dest, 0, STRIPE_SIZE); + count = 0; for (i = disks ; i--; ) { if (i == dd_idx || i == qd_idx) continue; @@ -1159,8 +1159,8 @@ static void compute_block_1(struct stripe_head *sh, int dd_idx, int nozero) check_xor(); } - if (count != 1) - xor_blocks(count, STRIPE_SIZE, ptr); + if (count) + xor_blocks(count, STRIPE_SIZE, dest, ptr); if (!nozero) set_bit(R5_UPTODATE, &sh->dev[dd_idx].flags); else clear_bit(R5_UPTODATE, &sh->dev[dd_idx].flags); } -- cgit v1.2.3