openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/drivers/mmc/core/queue.c

514 lines
12 KiB
C
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Copyright (C) 2003 Russell King, All Rights Reserved.
mmc: Move queue functions to mmc_block The mmc block queue functions are tailored for the mmc_block driver, so move those functions into that module. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2006-12-24 03:03:02 +08:00
* Copyright 2006-2007 Pierre Ossman
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
*/
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <linux/module.h>
#include <linux/blkdev.h>
Freezer: make kernel threads nonfreezable by default Currently, the freezer treats all tasks as freezable, except for the kernel threads that explicitly set the PF_NOFREEZE flag for themselves. This approach is problematic, since it requires every kernel thread to either set PF_NOFREEZE explicitly, or call try_to_freeze(), even if it doesn't care for the freezing of tasks at all. It seems better to only require the kernel threads that want to or need to be frozen to use some freezer-related code and to remove any freezer-related code from the other (nonfreezable) kernel threads, which is done in this patch. The patch causes all kernel threads to be nonfreezable by default (ie. to have PF_NOFREEZE set by default) and introduces the set_freezable() function that should be called by the freezable kernel threads in order to unset PF_NOFREEZE. It also makes all of the currently freezable kernel threads call set_freezable(), so it shouldn't cause any (intentional) change of behaviour to appear. Additionally, it updates documentation to describe the freezing of tasks more accurately. [akpm@linux-foundation.org: build fixes] Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Nigel Cunningham <nigel@nigel.suspend2.net> Cc: Pavel Machek <pavel@ucw.cz> Cc: Oleg Nesterov <oleg@tv-sign.ru> Cc: Gautham R Shenoy <ego@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-17 19:03:35 +08:00
#include <linux/freezer.h>
mmc: remove kernel_thread() Replace kernel_thread() with kthread_run()/kthread_stop(). Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2006-11-14 03:23:52 +08:00
#include <linux/kthread.h>
[SG] Update drivers to use sg helpers Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2007-10-23 03:19:53 +08:00
#include <linux/scatterlist.h>
ARM: 7797/1: mmc: Use dma_max_pfn(dev) helper for bounce_limit calculations DMA bounce limit is the maximum direct DMA'able memory beyond which bounce buffers has to be used to perform dma operations. MMC queue layr relies on dma_mask but its calculation is based on max_*pfn which don't have uniform meaning across architectures. So make use of dma_max_pfn() which is expected to return the DMAable maximum pfn value across architectures. Cc: Chris Ball <cjb@laptop.org> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2013-07-29 21:20:15 +08:00
#include <linux/dma-mapping.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <linux/mmc/card.h>
#include <linux/mmc/host.h>
mmc: card: do away with indirection pointer We have enough vtables in the kernel as it is, we don't need this one to create even more artificial separation of concerns. As is proved by the Makefile: obj-$(CONFIG_MMC_BLOCK) += mmc_block.o mmc_block-objs := block.o queue.o block.c and queue.c are baked into the same mmc_block.o object. So why would one of these objects access a function in the other object by dereferencing a pointer? Create a new block.h header file for the single shared function from block to queue and remove the function pointer and just call the queue request function. Apart from making the code more readable, this also makes link optimizations possible and probably speeds up the call as well. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-09-20 17:34:38 +08:00
mmc: Move queue functions to mmc_block The mmc block queue functions are tailored for the mmc_block driver, so move those functions into that module. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2006-12-24 03:03:02 +08:00
#include "queue.h"
mmc: card: do away with indirection pointer We have enough vtables in the kernel as it is, we don't need this one to create even more artificial separation of concerns. As is proved by the Makefile: obj-$(CONFIG_MMC_BLOCK) += mmc_block.o mmc_block-objs := block.o queue.o block.c and queue.c are baked into the same mmc_block.o object. So why would one of these objects access a function in the other object by dereferencing a pointer? Create a new block.h header file for the single shared function from block to queue and remove the function pointer and just call the queue request function. Apart from making the code more readable, this also makes link optimizations possible and probably speeds up the call as well. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-09-20 17:34:38 +08:00
#include "block.h"
mmc: core: Move public functions from core.h to private headers A significant amount of functions are available through the public mmc core.h header file. Let's slim down this public mmc interface, as to prevent users from abusing it, by moving some of the functions to private mmc header files. This change concentrates on moving the functions into private mmc headers, following changes may continue with additional clean-ups, as an example some functions can be turned into static. Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
2017-01-13 21:14:08 +08:00
#include "core.h"
mmc: core: Move public functions from card.h to private headers A significant amount of functions and other definitions are available through the public mmc card.h header file. Let's slim down this public mmc interface, as to prevent users from abusing it, by moving some of the functions/definitions to private mmc header files. This change concentrates on moving the functions into private mmc headers, following changes may continue with additional clean-ups. Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
2017-01-13 21:14:14 +08:00
#include "card.h"
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
#include "host.h"
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
static inline bool mmc_cqe_dcmd_busy(struct mmc_queue *mq)
{
/* Allow only 1 DCMD at a time */
return mq->in_flight[MMC_ISSUE_DCMD];
}
void mmc_cqe_check_busy(struct mmc_queue *mq)
{
if ((mq->cqe_busy & MMC_CQE_DCMD_BUSY) && !mmc_cqe_dcmd_busy(mq))
mq->cqe_busy &= ~MMC_CQE_DCMD_BUSY;
mq->cqe_busy &= ~MMC_CQE_QUEUE_FULL;
}
static inline bool mmc_cqe_can_dcmd(struct mmc_host *host)
{
return host->caps2 & MMC_CAP2_CQE_DCMD;
}
mmc: block: make function mmc_cqe_issue_type static The function mmc_cqe_issue_type is local to the source and does not need to be in global scope, so make it static. Cleans up sparse warning: drivers/mmc/core/queue.c:62:21: warning: symbol 'mmc_cqe_issue_type' was not declared. Should it be static? Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-11-30 19:37:38 +08:00
static enum mmc_issue_type mmc_cqe_issue_type(struct mmc_host *host,
struct request *req)
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
{
switch (req_op(req)) {
case REQ_OP_DRV_IN:
case REQ_OP_DRV_OUT:
case REQ_OP_DISCARD:
case REQ_OP_SECURE_ERASE:
return MMC_ISSUE_SYNC;
case REQ_OP_FLUSH:
return mmc_cqe_can_dcmd(host) ? MMC_ISSUE_DCMD : MMC_ISSUE_SYNC;
default:
return MMC_ISSUE_ASYNC;
}
}
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
enum mmc_issue_type mmc_issue_type(struct mmc_queue *mq, struct request *req)
{
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
struct mmc_host *host = mq->card->host;
if (mq->use_cqe)
return mmc_cqe_issue_type(host, req);
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
if (req_op(req) == REQ_OP_READ || req_op(req) == REQ_OP_WRITE)
return MMC_ISSUE_ASYNC;
return MMC_ISSUE_SYNC;
}
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
static void __mmc_cqe_recovery_notifier(struct mmc_queue *mq)
{
if (!mq->recovery_needed) {
mq->recovery_needed = true;
schedule_work(&mq->recovery_work);
}
}
void mmc_cqe_recovery_notifier(struct mmc_request *mrq)
{
struct mmc_queue_req *mqrq = container_of(mrq, struct mmc_queue_req,
brq.mrq);
struct request *req = mmc_queue_req_to_req(mqrq);
struct request_queue *q = req->q;
struct mmc_queue *mq = q->queuedata;
unsigned long flags;
spin_lock_irqsave(q->queue_lock, flags);
__mmc_cqe_recovery_notifier(mq);
spin_unlock_irqrestore(q->queue_lock, flags);
}
static enum blk_eh_timer_return mmc_cqe_timed_out(struct request *req)
{
struct mmc_queue_req *mqrq = req_to_mmc_queue_req(req);
struct mmc_request *mrq = &mqrq->brq.mrq;
struct mmc_queue *mq = req->q->queuedata;
struct mmc_host *host = mq->card->host;
enum mmc_issue_type issue_type = mmc_issue_type(mq, req);
bool recovery_needed = false;
switch (issue_type) {
case MMC_ISSUE_ASYNC:
case MMC_ISSUE_DCMD:
if (host->cqe_ops->cqe_timeout(host, mrq, &recovery_needed)) {
if (recovery_needed)
__mmc_cqe_recovery_notifier(mq);
return BLK_EH_RESET_TIMER;
}
/* No timeout */
return BLK_EH_HANDLED;
default:
/* Timeout is handled by mmc core */
return BLK_EH_RESET_TIMER;
}
}
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
static enum blk_eh_timer_return mmc_mq_timed_out(struct request *req,
bool reserved)
{
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
struct request_queue *q = req->q;
struct mmc_queue *mq = q->queuedata;
unsigned long flags;
int ret;
spin_lock_irqsave(q->queue_lock, flags);
if (mq->recovery_needed || !mq->use_cqe)
ret = BLK_EH_RESET_TIMER;
else
ret = mmc_cqe_timed_out(req);
spin_unlock_irqrestore(q->queue_lock, flags);
return ret;
}
static void mmc_mq_recovery_handler(struct work_struct *work)
{
struct mmc_queue *mq = container_of(work, struct mmc_queue,
recovery_work);
struct request_queue *q = mq->queue;
mmc_get_card(mq->card, &mq->ctx);
mq->in_recovery = true;
mmc: block: blk-mq: Add support for direct completion For blk-mq, add support for completing requests directly in the ->done callback. That means that error handling and urgent background operations must be handled by recovery_work in that case. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:07 +08:00
if (mq->use_cqe)
mmc_blk_cqe_recovery(mq);
else
mmc_blk_mq_recovery(mq);
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
mq->in_recovery = false;
spin_lock_irq(q->queue_lock);
mq->recovery_needed = false;
spin_unlock_irq(q->queue_lock);
mmc_put_card(mq->card, &mq->ctx);
blk_mq_run_hw_queues(q, true);
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
}
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
static struct scatterlist *mmc_alloc_sg(int sg_len, gfp_t gfp)
mmc: block: add member in mmc queue struct to hold request data The way the request data is organized in the mmc queue struct, it only allows processing of one request at a time. This patch adds a new struct to hold mmc queue request data such as sg list, request, blk request and bounce buffers, and updates any functions depending on the mmc queue struct. This prepares for using multiple active requests in one mmc queue. Signed-off-by: Per Forlin <per.forlin@linaro.org> Acked-by: Kyungmin Park <kyungmin.park@samsung.com> Acked-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Venkatraman S <svenkatr@ti.com> Tested-by: Sourav Poddar <sourav.poddar@ti.com> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-07-10 05:12:36 +08:00
{
struct scatterlist *sg;
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
sg = kmalloc_array(sg_len, sizeof(*sg), gfp);
mmc: queue: Share mmc request array between partitions eMMC can have multiple internal partitions that are represented as separate disks / queues. However switching between partitions is only done when the queue is empty. Consequently the array of mmc requests that are queued can be shared between partitions saving memory. Keep a pointer to the mmc request queue on the card, and use that instead of allocating a new one for each partition. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-03-13 20:36:36 +08:00
if (sg)
mmc: block: add member in mmc queue struct to hold request data The way the request data is organized in the mmc queue struct, it only allows processing of one request at a time. This patch adds a new struct to hold mmc queue request data such as sg list, request, blk request and bounce buffers, and updates any functions depending on the mmc queue struct. This prepares for using multiple active requests in one mmc queue. Signed-off-by: Per Forlin <per.forlin@linaro.org> Acked-by: Kyungmin Park <kyungmin.park@samsung.com> Acked-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Venkatraman S <svenkatr@ti.com> Tested-by: Sourav Poddar <sourav.poddar@ti.com> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-07-10 05:12:36 +08:00
sg_init_table(sg, sg_len);
return sg;
}
mmc: queue: let host controllers specify maximum discard timeout Some host controllers will not operate without a hardware timeout that is limited in value. However large discards require large timeouts, so there needs to be a way to specify the maximum discard size. A host controller driver may now specify the maximum discard timeout possible so that max_discard_sectors can be calculated. However, for eMMC when the High Capacity Erase Group Size is not in use, the timeout calculation depends on clock rate which may change. For that case Preferred Erase Size is used instead. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-06-28 22:16:02 +08:00
static void mmc_queue_setup_discard(struct request_queue *q,
struct mmc_card *card)
{
unsigned max_discard;
max_discard = mmc_calc_max_discard(card);
if (!max_discard)
return;
queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
block: have drivers use blk_queue_max_discard_sectors() Some drivers use it now, others just set the limits field manually. But in preparation for splitting this into a hard and soft limit, ensure that they all call the proper function for setting the hw limit for discards. Reviewed-by: Jeff Moyer <jmoyer@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-07-14 22:15:12 +08:00
blk_queue_max_discard_sectors(q, max_discard);
mmc: queue: let host controllers specify maximum discard timeout Some host controllers will not operate without a hardware timeout that is limited in value. However large discards require large timeouts, so there needs to be a way to specify the maximum discard size. A host controller driver may now specify the maximum discard timeout possible so that max_discard_sectors can be calculated. However, for eMMC when the High Capacity Erase Group Size is not in use, the timeout calculation depends on clock rate which may change. For that case Preferred Erase Size is used instead. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-06-28 22:16:02 +08:00
q->limits.discard_granularity = card->pref_erase << 9;
/* granularity must not be greater than max. discard */
if (card->pref_erase > max_discard)
q->limits.discard_granularity = 0;
mmc: card: Adding support for sanitize in eMMC 4.5 The sanitize support is added as a user-app ioctl call, and was removed from the block-device request, since its purpose is to be invoked not via File-System but by a user. This feature deletes the unmap memory region of the eMMC card, by writing to a specific register in the EXT_CSD. unmap region is the memory region that was previously deleted (by erase, trim or discard operation). In order to avoid timeout when sanitizing large-scale cards, the timeout for sanitize operation is 240 seconds. Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org> Signed-off-by: Maya Erez <merez@codeaurora.org> Signed-off-by: Chris Ball <cjb@laptop.org>
2013-04-18 20:41:55 +08:00
if (mmc_can_secure_erase_trim(card))
block: add a separate operation type for secure erase Instead of overloading the discard support with the REQ_SECURE flag. Use the opportunity to rename the queue flag as well, and remove the dead checks for this flag in the RAID 1 and RAID 10 drivers that don't claim support for secure erase. Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@fb.com>
2016-06-09 22:00:36 +08:00
queue_flag_set_unlocked(QUEUE_FLAG_SECERASE, q);
mmc: queue: let host controllers specify maximum discard timeout Some host controllers will not operate without a hardware timeout that is limited in value. However large discards require large timeouts, so there needs to be a way to specify the maximum discard size. A host controller driver may now specify the maximum discard timeout possible so that max_discard_sectors can be calculated. However, for eMMC when the High Capacity Erase Group Size is not in use, the timeout calculation depends on clock rate which may change. For that case Preferred Erase Size is used instead. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-06-28 22:16:02 +08:00
}
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
/**
* mmc_init_request() - initialize the MMC-specific per-request data
* @q: the request queue
* @req: the request
* @gfp: memory allocation policy
*/
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
static int __mmc_init_request(struct mmc_queue *mq, struct request *req,
gfp_t gfp)
mmc: queue: Factor out mmc_queue_alloc_sgs() In preparation for supporting a queue of requests, factor out mmc_queue_alloc_sgs(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Harjani Ritesh <riteshh@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:13 +08:00
{
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
struct mmc_queue_req *mq_rq = req_to_mmc_queue_req(req);
struct mmc_card *card = mq->card;
struct mmc_host *host = card->host;
mmc: queue: Factor out mmc_queue_alloc_sgs() In preparation for supporting a queue of requests, factor out mmc_queue_alloc_sgs(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Harjani Ritesh <riteshh@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:13 +08:00
mmc: Delete bounce buffer handling In may, Steven sent a patch deleting the bounce buffer handling and the CONFIG_MMC_BLOCK_BOUNCE option. I chose the less invasive path of making it a runtime config option, and we merged that successfully for kernel v4.12. The code is however just standing in the way and taking up space for seemingly no gain on any systems in wide use today. Pierre says the code was there to improve speed on TI SDHCI controllers on certain HP laptops and possibly some Ricoh controllers as well. Early SDHCI controllers lacked the scatter-gather feature, which made software bounce buffers a significant speed boost. We are clearly talking about the list of SDHCI PCI-based MMC/SD card readers found in the pci_ids[] list in drivers/mmc/host/sdhci-pci-core.c. The TI SDHCI derivative is not supported by the upstream kernel. This leaves the Ricoh. What we can however notice is that the x86 defconfigs in the kernel did not enable CONFIG_MMC_BLOCK_BOUNCE option, which means that any such laptop would have to have a custom configured kernel to actually take advantage of this bounce buffer speed-up. It simply seems like there was a speed optimization for the Ricoh controllers that noone was using. (I have not checked the distro defconfigs but I am pretty sure the situation is the same there.) Bounce buffers increased performance on the OMAP HSMMC at one point, and was part of the original submission in commit a45c6cb81647 ("[ARM] 5369/1: omap mmc: Add new omap hsmmc controller for 2430 and 34xx, v3") This optimization was removed in commit 0ccd76d4c236 ("omap_hsmmc: Implement scatter-gather emulation") which found that scatter-gather emulation provided even better performance. The same was introduced for SDHCI in commit 2134a922c6e7 ("sdhci: scatter-gather (ADMA) support") I am pretty positively convinced that software scatter-gather emulation will do for any host controller what the bounce buffers were doing. Essentially, the bounce buffer was a reimplementation of software scatter-gather-emulation in the MMC subsystem, and it should be done away with. Cc: Pierre Ossman <pierre@ossman.eu> Cc: Juha Yrjola <juha.yrjola@solidboot.com> Cc: Steven J. Hill <Steven.Hill@cavium.com> Cc: Shawn Lin <shawn.lin@rock-chips.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Suggested-by: Steven J. Hill <Steven.Hill@cavium.com> Suggested-by: Shawn Lin <shawn.lin@rock-chips.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-09-20 16:56:14 +08:00
mq_rq->sg = mmc_alloc_sg(host->max_segs, gfp);
if (!mq_rq->sg)
return -ENOMEM;
mmc: queue: Factor out mmc_queue_alloc_sgs() In preparation for supporting a queue of requests, factor out mmc_queue_alloc_sgs(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Harjani Ritesh <riteshh@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:13 +08:00
mmc: queue: Introduce queue depth and use it to allocate and free Add a mmc_queue member to record the size of the queue, which currently supports 2 requests on-the-go at a time. Instead of allocating resources for 2 slots in the queue, allow for an arbitrary number. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:15 +08:00
return 0;
}
mmc: queue: Factor out mmc_queue_alloc_sgs() In preparation for supporting a queue of requests, factor out mmc_queue_alloc_sgs(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Harjani Ritesh <riteshh@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:13 +08:00
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
static void mmc_exit_request(struct request_queue *q, struct request *req)
mmc: queue: Introduce queue depth and use it to allocate and free Add a mmc_queue member to record the size of the queue, which currently supports 2 requests on-the-go at a time. Instead of allocating resources for 2 slots in the queue, allow for an arbitrary number. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:15 +08:00
{
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
struct mmc_queue_req *mq_rq = req_to_mmc_queue_req(req);
mmc: queue: Share mmc request array between partitions eMMC can have multiple internal partitions that are represented as separate disks / queues. However switching between partitions is only done when the queue is empty. Consequently the array of mmc requests that are queued can be shared between partitions saving memory. Keep a pointer to the mmc request queue on the card, and use that instead of allocating a new one for each partition. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-03-13 20:36:36 +08:00
mmc: core: Allocate per-request data using the block layer core The mmc_queue_req is a per-request state container the MMC core uses to carry bounce buffers, pointers to asynchronous requests and so on. Currently allocated as a static array of objects, then as a request comes in, a mmc_queue_req is assigned to it, and used during the lifetime of the request. This is backwards compared to how other block layer drivers work: they usally let the block core provide a per-request struct that get allocated right beind the struct request, and which can be obtained using the blk_mq_rq_to_pdu() helper. (The _mq_ infix in this function name is misleading: it is used by both the old and the MQ block layer.) The per-request struct gets allocated to the size stored in the queue variable .cmd_size initialized using the .init_rq_fn() and cleaned up using .exit_rq_fn(). The block layer code makes the MMC core rely on this mechanism to allocate the per-request mmc_queue_req state container. Doing this make a lot of complicated queue handling go away. We only need to keep the .qnct that keeps count of how many request are currently being processed by the MMC layer. The MQ block layer will replace also this once we transition to it. Doing this refactoring is necessary to move the ioctl() operations into custom block layer requests tagged with REQ_OP_DRV_[IN|OUT] instead of the custom code using the BigMMCHostLock that we have today: those require that per-request data be obtainable easily from a request after creating a custom request with e.g.: struct request *rq = blk_get_request(q, REQ_OP_DRV_IN, __GFP_RECLAIM); struct mmc_queue_req *mq_rq = req_to_mq_rq(rq); And this is not possible with the current construction, as the request is not immediately assigned the per-request state container, but instead it gets assigned when the request finally enters the MMC queue, which is way too late for custom requests. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> [Ulf: Folded in the fix to drop a call to blk_cleanup_queue()] Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Heiner Kallweit <hkallweit1@gmail.com>
2017-05-18 17:29:32 +08:00
kfree(mq_rq->sg);
mq_rq->sg = NULL;
mmc: queue: Factor out mmc_queue_reqs_free_bufs() In preparation for supporting a queue of requests, factor out mmc_queue_reqs_free_bufs(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Harjani Ritesh <riteshh@codeaurora.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2016-11-29 18:09:14 +08:00
}
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
static int mmc_mq_init_request(struct blk_mq_tag_set *set, struct request *req,
unsigned int hctx_idx, unsigned int numa_node)
{
return __mmc_init_request(set->driver_data, req, GFP_KERNEL);
}
static void mmc_mq_exit_request(struct blk_mq_tag_set *set, struct request *req,
unsigned int hctx_idx)
{
struct mmc_queue *mq = set->driver_data;
mmc_exit_request(mq->queue, req);
}
/*
* We use BLK_MQ_F_BLOCKING and have only 1 hardware queue, which means requests
* will not be dispatched in parallel.
*/
static blk_status_t mmc_mq_queue_rq(struct blk_mq_hw_ctx *hctx,
const struct blk_mq_queue_data *bd)
{
struct request *req = bd->rq;
struct request_queue *q = req->q;
struct mmc_queue *mq = q->queuedata;
struct mmc_card *card = mq->card;
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
struct mmc_host *host = card->host;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
enum mmc_issue_type issue_type;
enum mmc_issued issued;
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
bool get_card, cqe_retune_ok;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
int ret;
if (mmc_card_removed(mq->card)) {
req->rq_flags |= RQF_QUIET;
return BLK_STS_IOERR;
}
issue_type = mmc_issue_type(mq, req);
spin_lock_irq(q->queue_lock);
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
if (mq->recovery_needed) {
spin_unlock_irq(q->queue_lock);
return BLK_STS_RESOURCE;
}
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
switch (issue_type) {
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
case MMC_ISSUE_DCMD:
if (mmc_cqe_dcmd_busy(mq)) {
mq->cqe_busy |= MMC_CQE_DCMD_BUSY;
spin_unlock_irq(q->queue_lock);
return BLK_STS_RESOURCE;
}
break;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
case MMC_ISSUE_ASYNC:
break;
default:
/*
* Timeouts are handled by mmc core, and we don't have a host
* API to abort requests, so we can't handle the timeout anyway.
* However, when the timeout happens, blk_mq_complete_request()
* no longer works (to stop the request disappearing under us).
* To avoid racing with that, set a large timeout.
*/
req->timeout = 600 * HZ;
break;
}
mq->in_flight[issue_type] += 1;
get_card = (mmc_tot_in_flight(mq) == 1);
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
cqe_retune_ok = (mmc_cqe_qcnt(mq) == 1);
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
spin_unlock_irq(q->queue_lock);
if (!(req->rq_flags & RQF_DONTPREP)) {
req_to_mmc_queue_req(req)->retries = 0;
req->rq_flags |= RQF_DONTPREP;
}
if (get_card)
mmc_get_card(card, &mq->ctx);
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
if (mq->use_cqe) {
host->retune_now = host->need_retune && cqe_retune_ok &&
!host->hold_retune;
}
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
blk_mq_start_request(req);
issued = mmc_blk_mq_issue_rq(mq, req);
switch (issued) {
case MMC_REQ_BUSY:
ret = BLK_STS_RESOURCE;
break;
case MMC_REQ_FAILED_TO_START:
ret = BLK_STS_IOERR;
break;
default:
ret = BLK_STS_OK;
break;
}
if (issued != MMC_REQ_STARTED) {
bool put_card = false;
spin_lock_irq(q->queue_lock);
mq->in_flight[issue_type] -= 1;
if (mmc_tot_in_flight(mq) == 0)
put_card = true;
spin_unlock_irq(q->queue_lock);
if (put_card)
mmc_put_card(card, &mq->ctx);
}
return ret;
}
static const struct blk_mq_ops mmc_mq_ops = {
.queue_rq = mmc_mq_queue_rq,
.init_request = mmc_mq_init_request,
.exit_request = mmc_mq_exit_request,
.complete = mmc_blk_mq_complete,
.timeout = mmc_mq_timed_out,
};
mmc: block: Factor out mmc_setup_queue() Factor out some common code that will also be used with blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-09-22 20:36:57 +08:00
static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
{
struct mmc_host *host = card->host;
u64 limit = BLK_BOUNCE_HIGH;
if (mmc_dev(host)->dma_mask && *mmc_dev(host)->dma_mask)
limit = (u64)dma_max_pfn(mmc_dev(host)) << PAGE_SHIFT;
queue_flag_set_unlocked(QUEUE_FLAG_NONROT, mq->queue);
queue_flag_clear_unlocked(QUEUE_FLAG_ADD_RANDOM, mq->queue);
if (mmc_can_erase(card))
mmc_queue_setup_discard(mq->queue, card);
blk_queue_bounce_limit(mq->queue, limit);
blk_queue_max_hw_sectors(mq->queue,
min(host->max_blk_count, host->max_req_size / 512));
blk_queue_max_segments(mq->queue, host->max_segs);
blk_queue_max_segment_size(mq->queue, host->max_seg_size);
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
INIT_WORK(&mq->recovery_work, mmc_mq_recovery_handler);
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
INIT_WORK(&mq->complete_work, mmc_blk_mq_complete_work);
mutex_init(&mq->complete_lock);
init_waitqueue_head(&mq->wait);
}
static int mmc_mq_init_queue(struct mmc_queue *mq, int q_depth,
const struct blk_mq_ops *mq_ops, spinlock_t *lock)
{
int ret;
memset(&mq->tag_set, 0, sizeof(mq->tag_set));
mq->tag_set.ops = mq_ops;
mq->tag_set.queue_depth = q_depth;
mq->tag_set.numa_node = NUMA_NO_NODE;
mq->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_SG_MERGE |
BLK_MQ_F_BLOCKING;
mq->tag_set.nr_hw_queues = 1;
mq->tag_set.cmd_size = sizeof(struct mmc_queue_req);
mq->tag_set.driver_data = mq;
ret = blk_mq_alloc_tag_set(&mq->tag_set);
if (ret)
return ret;
mq->queue = blk_mq_init_queue(&mq->tag_set);
if (IS_ERR(mq->queue)) {
ret = PTR_ERR(mq->queue);
goto free_tag_set;
}
mq->queue->queue_lock = lock;
mq->queue->queuedata = mq;
return 0;
free_tag_set:
blk_mq_free_tag_set(&mq->tag_set);
return ret;
}
/* Set queue depth to get a reasonable value for q->nr_requests */
#define MMC_QUEUE_DEPTH 64
static int mmc_mq_init(struct mmc_queue *mq, struct mmc_card *card,
spinlock_t *lock)
{
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
struct mmc_host *host = card->host;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
int q_depth;
int ret;
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
/*
* The queue depth for CQE must match the hardware because the request
* tag is used to index the hardware queue.
*/
if (mq->use_cqe)
q_depth = min_t(int, card->ext_csd.cmdq_depth, host->cqe_qdepth);
else
q_depth = MMC_QUEUE_DEPTH;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
ret = mmc_mq_init_queue(mq, q_depth, &mmc_mq_ops, lock);
if (ret)
return ret;
blk_queue_rq_timeout(mq->queue, 60 * HZ);
mmc_setup_queue(mq, card);
return 0;
mmc: block: Factor out mmc_setup_queue() Factor out some common code that will also be used with blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-09-22 20:36:57 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/**
* mmc_init_queue - initialise a queue structure.
* @mq: mmc queue
* @card: mmc card to attach this queue
* @lock: queue lock
mmc: queue: append partition subname to queue thread name For example, an eMMC with 2 boot partitions will have 3 threads. The names change from: 40 ? 00:00:00 mmcqd/0 41 ? 00:00:00 mmcqd/0 42 ? 00:00:00 mmcqd/0 to: 40 ? 00:00:00 mmcqd/0 41 ? 00:00:00 mmcqd/0boot0 42 ? 00:00:00 mmcqd/0boot1 Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Cc: Andrei Warkentin <andreiw@motorola.com> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-06-23 18:40:28 +08:00
* @subname: partition subname
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*
* Initialise a MMC card request queue.
*/
mmc: queue: append partition subname to queue thread name For example, an eMMC with 2 boot partitions will have 3 threads. The names change from: 40 ? 00:00:00 mmcqd/0 41 ? 00:00:00 mmcqd/0 42 ? 00:00:00 mmcqd/0 to: 40 ? 00:00:00 mmcqd/0 41 ? 00:00:00 mmcqd/0boot0 42 ? 00:00:00 mmcqd/0boot1 Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Cc: Andrei Warkentin <andreiw@motorola.com> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-06-23 18:40:28 +08:00
int mmc_init_queue(struct mmc_queue *mq, struct mmc_card *card,
spinlock_t *lock, const char *subname)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct mmc_host *host = card->host;
mq->card = card;
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
mmc: block: Add CQE support Add CQE support to the block driver, including: - optionally using DCMD for flush requests - "manually" issuing discard requests - issuing read / write requests to the CQE - supporting block-layer timeouts - handling recovery - supporting re-tuning CQE offers 25% - 50% better random multi-threaded I/O. There is a slight (e.g. 2%) drop in sequential read speed but no observable change to sequential write. CQE automatically sends the commands to complete requests. However it only supports reads / writes and so-called "direct commands" (DCMD). Furthermore DCMD is limited to one command at a time, but discards require 3 commands. That makes issuing discards through CQE very awkward, but some CQE's don't support DCMD anyway. So for discards, the existing non-CQE approach is taken, where the mmc core code issues the 3 commands one at a time i.e. mmc_erase(). Where DCMD is used, is for issuing flushes. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:04 +08:00
mq->use_cqe = host->cqe_enabled;
mmc: block: Remove code no longer needed after the switch to blk-mq Remove code no longer needed after the switch to blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-11-29 21:41:18 +08:00
return mmc_mq_init(mq, card, lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
mmc: block: Remove code no longer needed after the switch to blk-mq Remove code no longer needed after the switch to blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-11-29 21:41:18 +08:00
void mmc_queue_suspend(struct mmc_queue *mq)
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
{
blk_mq_quiesce_queue(mq->queue);
/*
* The host remains claimed while there are outstanding requests, so
* simply claiming and releasing here ensures there are none.
*/
mmc_claim_host(mq->card->host);
mmc_release_host(mq->card->host);
}
mmc: block: Remove code no longer needed after the switch to blk-mq Remove code no longer needed after the switch to blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-11-29 21:41:18 +08:00
void mmc_queue_resume(struct mmc_queue *mq)
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
{
blk_mq_unquiesce_queue(mq->queue);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
void mmc_cleanup_queue(struct mmc_queue *mq)
{
[BLOCK] Get rid of request_queue_t typedef Some of the code has been gradually transitioned to using the proper struct request_queue, but there's lots left. So do a full sweet of the kernel and get rid of this typedef and replace its uses with the proper type. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2007-07-24 15:28:11 +08:00
struct request_queue *q = mq->queue;
mmc: Flush block queue when removing card After mmc_block's remove function has exited, we must not touch the card structure in any way. This means we not only must remove the gendisk, we must also flush out any remaning requests already queued up. We previously removed the disk, but didn't flush it, causing oops:es when removing a card in the middle of a transfer. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2006-11-15 05:08:16 +08:00
mmc: block: Remove code no longer needed after the switch to blk-mq Remove code no longer needed after the switch to blk-mq. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-11-29 21:41:18 +08:00
/*
* The legacy code handled the possibility of being suspended,
* so do that here too.
*/
if (blk_queue_quiesced(q))
blk_mq_unquiesce_queue(q);
mmc_block: fix queue cleanup The main bug was that 'blk_cleanup_queue()' was called while the block device could still be in use, for example, because the card was removed while files were still open. In addition, to be sure that 'mmc_request()' will get called for all new requests (so it can error them out), the queue is emptied during cleanup. This is done after the worker thread is stopped to avoid racing with it. Finally, it is not a device error for this to be happening, so quiet the (sometimes very many) error messages. Signed-off-by: Adrian Hunter <adrian.hunter@nokia.com> Cc: <linux-mmc@vger.kernel.org> Cc: <stable@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-01-09 06:43:00 +08:00
mmc: block: Simplify cleaning up the queue Use blk_cleanup_queue() to shutdown the queue when the driver is removed, and instead get an extra reference to the queue to prevent the queue being freed before the final mmc_blk_put(). Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:40:59 +08:00
blk_cleanup_queue(q);
mmc: block: Add blk-mq support Define and use a blk-mq queue. Discards and flushes are processed synchronously, but reads and writes asynchronously. In order to support slow DMA unmapping, DMA unmapping is not done until after the next request is started. That means the request is not completed until then. If there is no next request then the completion is done by queued work. Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Linus Walleij <linus.walleij@linaro.org>
2017-11-29 21:41:03 +08:00
/*
* A request can be completed before the next request, potentially
* leaving a complete_work with nothing to do. Such a work item might
* still be queued at this point. Flush it.
*/
flush_work(&mq->complete_work);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
mq->card = NULL;
}
mmc_block: bounce buffer highmem support Support highmem pages in the bounce buffer code by using the sg_copy_from/to_buffer() functions. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-07-22 20:35:42 +08:00
/*
* Prepare the sg list(s) to be handed of to the host driver
*/
mmc: block: add member in mmc queue struct to hold request data The way the request data is organized in the mmc queue struct, it only allows processing of one request at a time. This patch adds a new struct to hold mmc queue request data such as sg list, request, blk request and bounce buffers, and updates any functions depending on the mmc queue struct. This prepares for using multiple active requests in one mmc queue. Signed-off-by: Per Forlin <per.forlin@linaro.org> Acked-by: Kyungmin Park <kyungmin.park@samsung.com> Acked-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Venkatraman S <svenkatr@ti.com> Tested-by: Sourav Poddar <sourav.poddar@ti.com> Tested-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Chris Ball <cjb@laptop.org>
2011-07-10 05:12:36 +08:00
unsigned int mmc_queue_map_sg(struct mmc_queue *mq, struct mmc_queue_req *mqrq)
mmc: bounce requests for simple hosts Some hosts cannot do scatter/gather in hardware. Since not doing sg is such a big performance hit, we (optionally) bounce the requests to a simple linear buffer that we hand over to the driver. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2007-05-12 06:26:16 +08:00
{
mmc: block: remove req back pointer Just as we can use blk_mq_rq_from_pdu() to get the per-request tag we can use blk_mq_rq_to_pdu() to get a request from a tag. Introduce a static inline helper so we are on the clear what is happening. Suggested-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-05-19 21:37:27 +08:00
struct request *req = mmc_queue_req_to_req(mqrq);
mmc: bounce requests for simple hosts Some hosts cannot do scatter/gather in hardware. Since not doing sg is such a big performance hit, we (optionally) bounce the requests to a simple linear buffer that we hand over to the driver. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2007-05-12 06:26:16 +08:00
mmc: Delete bounce buffer handling In may, Steven sent a patch deleting the bounce buffer handling and the CONFIG_MMC_BLOCK_BOUNCE option. I chose the less invasive path of making it a runtime config option, and we merged that successfully for kernel v4.12. The code is however just standing in the way and taking up space for seemingly no gain on any systems in wide use today. Pierre says the code was there to improve speed on TI SDHCI controllers on certain HP laptops and possibly some Ricoh controllers as well. Early SDHCI controllers lacked the scatter-gather feature, which made software bounce buffers a significant speed boost. We are clearly talking about the list of SDHCI PCI-based MMC/SD card readers found in the pci_ids[] list in drivers/mmc/host/sdhci-pci-core.c. The TI SDHCI derivative is not supported by the upstream kernel. This leaves the Ricoh. What we can however notice is that the x86 defconfigs in the kernel did not enable CONFIG_MMC_BLOCK_BOUNCE option, which means that any such laptop would have to have a custom configured kernel to actually take advantage of this bounce buffer speed-up. It simply seems like there was a speed optimization for the Ricoh controllers that noone was using. (I have not checked the distro defconfigs but I am pretty sure the situation is the same there.) Bounce buffers increased performance on the OMAP HSMMC at one point, and was part of the original submission in commit a45c6cb81647 ("[ARM] 5369/1: omap mmc: Add new omap hsmmc controller for 2430 and 34xx, v3") This optimization was removed in commit 0ccd76d4c236 ("omap_hsmmc: Implement scatter-gather emulation") which found that scatter-gather emulation provided even better performance. The same was introduced for SDHCI in commit 2134a922c6e7 ("sdhci: scatter-gather (ADMA) support") I am pretty positively convinced that software scatter-gather emulation will do for any host controller what the bounce buffers were doing. Essentially, the bounce buffer was a reimplementation of software scatter-gather-emulation in the MMC subsystem, and it should be done away with. Cc: Pierre Ossman <pierre@ossman.eu> Cc: Juha Yrjola <juha.yrjola@solidboot.com> Cc: Steven J. Hill <Steven.Hill@cavium.com> Cc: Shawn Lin <shawn.lin@rock-chips.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Suggested-by: Steven J. Hill <Steven.Hill@cavium.com> Suggested-by: Shawn Lin <shawn.lin@rock-chips.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2017-09-20 16:56:14 +08:00
return blk_rq_map_sg(mq->queue, req, mqrq->sg);
mmc: bounce requests for simple hosts Some hosts cannot do scatter/gather in hardware. Since not doing sg is such a big performance hit, we (optionally) bounce the requests to a simple linear buffer that we hand over to the driver. Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2007-05-12 06:26:16 +08:00
}