summaryrefslogtreecommitdiffstats
path: root/drivers/ata/libata-pmp.c
AgeCommit message (Collapse)AuthorFilesLines
2010-10-21libata: implement LPM support for port multipliersTejun Heo1-4/+44
Port multipliers can do DIPM on fan-out links fine. Implement support for it. Tested w/ SIMG 57xx and marvell PMPs. Both the host and fan-out links enter power save modes nicely. SIMG 37xx and 47xx report link offline on SStatus causing EH to detach the devices. Blacklisted. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2010-10-21[libata] Add ATA transport classGwendal Grignou1-2/+16
This is a scheleton for libata transport class. All information is read only, exporting information from libata: - ata_port class: one per ATA port - ata_link class: one per ATA port or 15 for SATA Port Multiplier - ata_device class: up to 2 for PATA link, usually one for SATA. Signed-off-by: Gwendal Grignou <gwendal@google.com> Reviewed-by: Grant Grundler <grundler@google.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2010-05-14[libata] Disable R_OK (Early ACK) on SII 3726 PMPGrant Grundler1-3/+29
In 2009, While running "cache read" performance test of drives behind SII PMP we encountered a "all 5 drives" timeout on more than 30% of the machines under test. This patch reduces the rate by a factor of about 70. Low enough that we didn't care to further investigate the issue. Performance impact with any sort of "normal" use was ~2%+ CPU and less than 1% throughput degradation. Worst case impact (cached read) was 6% IOPS reduction. This is with NCQ off (q=1) but I believe FIS based switching enabled in the SATA driver. The patch disables "Early ACK" in the 3726 port multiplier. "Early ACK" is issued when device sends a FIS to the host (via PMP) and the PMP sends an ACK immediately back to the device - well before the host gets the response. Under worst case IOPs load (cached read test) and more than 2 PMPs connected to a 4-port SATA controller, I suspect the time to service all of the PMPs is exceeding the PMPs ability to keep track of outstanding FIS it owes the Host. Reducing the number of PMPs to 2 (or 1) reduces the frequency by several orders of magnitude. Kudos to Gwendal for initial debugging of this issue. [Any errors in the description are mine, not his.] Patch is currently in production on Google servers. Signed-off-by: Grant Grundler <grundler@google.com> Signed-off-by: Gwendal Grignou <gwendal@google.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2010-03-30include cleanup: Update gfp.h and slab.h includes to prepare for breaking ↵Tejun Heo1-0/+1
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>
2009-09-08libata: add SATA PMP revision information for spec 1.2Shane Huang1-0/+2
This small patch is just adding the information for PMP spec 1.2 Signed-off-by: Shane Huang <shane.huang@amd.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2009-02-02libata: add @spd_limit to sata_down_spd_limit()Tejun Heo1-1/+1
Add @spd_limit to sata_down_spd_limit() so that the caller can specify the SPD limit it wants. This parameter doesn't get in the way even when it's too low. The closest possible limit is applied. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-12-28libata: beef up iteratorsTejun Heo1-11/+11
There currently are the following looping constructs. * __ata_port_for_each_link() for all available links * ata_port_for_each_link() for edge links * ata_link_for_each_dev() for all devices * ata_link_for_each_dev_reverse() for all devices in reverse order Now there's a need for looping construct which is similar to __ata_port_for_each_link() but iterates over PMP links before the host link. Instead of adding another one with long name, do the following cleanup. * Implement and export ata_link_next() and ata_dev_next() which take @mode parameter and can be used to build custom loop. * Implement ata_for_each_link() and ata_for_each_dev() which take looping mode explicitly. The following iteration modes are implemented. * ATA_LITER_EDGE : loop over edge links * ATA_LITER_HOST_FIRST : loop over all links, host link first * ATA_LITER_PMP_FIRST : loop over all links, PMP links first * ATA_DITER_ENABLED : loop over enabled devices * ATA_DITER_ENABLED_REVERSE : loop over enabled devices in reverse order * ATA_DITER_ALL : loop over all devices * ATA_DITER_ALL_REVERSE : loop over all devices in reverse order This change removes exlicit device enabledness checks from many loops and makes it clear which ones are iterated over in which direction. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-07-14libata: improve EH retry delay handlingTejun Heo1-10/+0
EH retries were delayed by 5 seconds to ensure that resets don't occur back-to-back. However, this 5 second delay is superflous or excessive in many cases. For example, after IDENTIFY times out, there's no reason to wait five more seconds before retrying. This patch adds ehc->last_reset timestamp and record the timestamp for the last reset trial or success and uses it to space resets by ATA_EH_RESET_COOL_DOWN which is 5 secs and removes unconditional 5 sec sleeps. As this change makes inter-try waits often shorter and they're redundant in nature, this patch also removes the "retrying..." messages. While at it, convert explicit rounding up division to DIV_ROUND_UP(). This change speeds up EH in many cases w/o sacrificing robustness. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-07-14libata: consistently use msecs for time durationsTejun Heo1-1/+2
libata has been using mix of jiffies and msecs for time druations. This is getting confusing. As writing sub HZ values in jiffies is PITA and msecs_to_jiffies() can't be used as initializer, unify unit for all time durations to msecs. So, durations are in msecs and deadlines are in jiffies. ata_deadline() is added to compute deadline from a start time and duration in msecs. While at it, drop now superflous _msec suffix from arguments and rename @timeout to @deadline if it represents a fixed point in time rather than duration. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-05-30libata: SRST can't be trusted on PMP sil3726Tejun Heo1-2/+5
As in sil4726, SRST can't be trusted on sil3726 causing detection problems under certain configuraitons. I thought it was from the Config Disk device but apparently not. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-05-19libata: make sure PMP notification is turned off during recoveryTejun Heo1-16/+20
PMP notification during reset can make some controllers fail reset processing and needs to be turned off during resets. PMP attach and full-revalidation path did this via sata_pmp_configure() but the quick revalidation wasn't. Move the notification disable code right above fan-out port recovery so that it's always turned off. This fixes obscure reset failures. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-05-19libata: increase PMP register access timeout to 3sTejun Heo1-2/+2
This timeout was set low because previously PMP register access was done via polling and register access timeouts could stack up. This is no longer the case. One timeout will make all following accesses fail immediately. In rare cases both marvell and SIMG PMPs need almost a second. Bump it to 3s. While at it, rename it to SATA_PMP_RW_TIMEOUT. It's not specific to SCR access. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-05-19libata: move reset freeze/thaw handling into ata_eh_reset()Tejun Heo1-4/+0
Previously reset freeze/thaw handling lived outside of ata_eh_reset() mainly because the original PMP reset code needed the port frozen while resetting all the fan-out ports, which is no longer the case. This patch moves freeze/thaw handling into ata_eh_reset(). @prereset() and @postreset() are now called w/o freezing the port although @prereset() an be called frozen if the port is frozen prior to entering ata_eh_reset(). This makes code simpler and will help removing hotplug event related races. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
2008-04-17libata: implement PMP helpersTejun Heo1-3/+3
Implement helpers to test whether PMP is supported, attached and determine pmp number to use when issuing SRST to a link. While at it, move ata_is_host_link() so that it's together with the two new PMP helpers. This change simplifies LLDs and helps making PMP support optional. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: separate PMP support code from core codeTejun Heo1-0/+12
Most of PMP support code is already in libata-pmp.c. All that are in libata-core.c are sata_pmp_port_ops and EXPORTs. Move them to libata-pmp.c. Also, collect PMP related prototypes and declarations in header files and move them right above of SFF stuff. This change is to make PMP support optional. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: move PMP SCR access failure during reset to ata_eh_reset()Tejun Heo1-35/+0
If PMP fan-out reset fails and SCR isn't accessible, PMP should be reset. This used to be tested by sata_pmp_std_hardreset() and communicated to EH by -ERESTART. However, this logic is generic and doesn't really have much to do with specific hardreset implementation. This patch moves SCR access failure detection logic to ata_eh_reset() where it belongs. As this makes sata_pmp_std_hardreset() identical to sata_std_hardreset(), the function is killed and replaced with the standard method. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: clear SError after link resumeTejun Heo1-92/+1
SError used to be cleared in ->postreset. This has small hotplug race condition. If a device is plugged in after reset is complete but postreset hasn't run yet, its hotplug event gets lost when SError is cleared. This patch makes sata_link_resume() clear SError. This kills the race condition and makes a lot of sense as some PMP and host PHYs don't work properly without SError cleared. This change makes sata_pmp_std_{pre|post}_reset()'s unnecessary as they become identical to ata_std counterparts. It also simplifies sata_pmp_hardreset() and ahci_vt8251_hardreset(). Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: move generic hardreset code from sata_sff_hardreset() to ↵Tejun Heo1-3/+4
sata_link_hardreset() sata_sff_hardreset() contains link readiness wait logic which isn't SFF specific. Move that part into sata_link_hardreset(), which now takes two more parameters - @online and @check_ready. Both are optional. The former is out parameter for link onlineness after reset. The latter is used to wait for link readiness after hardreset. Users of sata_link_hardreset() is updated to use new funtionality and ahci_hardreset() is updated to use sata_link_hardreset() instead of sata_sff_hardreset(). This doesn't really cause any behavior change. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: make reset related methods proper port operationsTejun Heo1-36/+11
Currently reset methods are not specified directly in the ata_port_operations table. If a LLD wants to use custom reset methods, it should construct and use a error_handler which uses those reset methods. It's done this way for two reasons. First, the ops table already contained too many methods and adding four more of them would noticeably increase the amount of necessary boilerplate code all over low level drivers. Second, as ->error_handler uses those reset methods, it can get confusing. ie. By overriding ->error_handler, those reset ops can be made useless making layering a bit hazy. Now that ops table uses inheritance, the first problem doesn't exist anymore. The second isn't completely solved but is relieved by providing default values - most drivers can just override what it has implemented and don't have to concern itself about higher level callbacks. In fact, there currently is no driver which actually modifies error handling behavior. Drivers which override ->error_handler just wraps the standard error handler only to prepare the controller for EH. I don't think making ops layering strict has any noticeable benefit. This patch makes ->prereset, ->softreset, ->hardreset, ->postreset and their PMP counterparts propoer ops. Default ops are provided in the base ops tables and drivers are converted to override individual reset methods instead of creating custom error_handler. * ata_std_error_handler() doesn't use sata_std_hardreset() if SCRs aren't accessible. sata_promise doesn't need to use separate error_handlers for PATA and SATA anymore. * softreset is broken for sata_inic162x and sata_sx4. As libata now always prefers hardreset, this doesn't really matter but the ops are forced to NULL using ATA_OP_NULL for documentation purpose. * pata_hpt374 needs to use different prereset for the first and second PCI functions. This used to be done by branching from hpt374_error_handler(). The proper way to do this is to use separate ops and port_info tables for each function. Converted. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: kill ata_ehi_schedule_probe()Tejun Heo1-4/+8
ata_ehi_schedule_probe() was created to hide details of link-resuming reset magic. Now that all the softreset workarounds are gone, scheduling probe is very simple - set probe_mask and request RESET. Kill ata_ehi_schedule_probe() and open code it. This also increases consistency as ata_ehi_schedule_probe() couldn't cover individual device probings so they were open-coded even when the helper existed. While at it, define ATA_ALL_DEVICES as mask of all possible devices on a link and always use it when requesting probe on link level for simplicity and consistency. Setting extra bits in the probe_mask doesn't hurt anybody. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: kill ATA_EHI_RESUME_LINKTejun Heo1-1/+0
ATA_EHI_RESUME_LINK has two functions - promote reset to hardreset if ATA_LFLAG_HRST_TO_RESUME is set and preventing EH from shortcutting reset action when probing is requested. The former is gone now and the latter can easily be achieved by making EH to perform at least one reset if reset is requested, which also makes more sense than depending on RESUME_LINK flag. As ATA_EHI_RESUME_LINK was the only EHI reset modifier, this also kills reset modifier handling. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: kill ATA_LFLAG_HRST_TO_RESUMETejun Heo1-16/+0
Now that hardreset is the preferred method of resetting, there's no need for ATA_LFLAG_HRST_TO_RESUME flag. Kill it. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-04-17libata: prefer hardresetTejun Heo1-20/+8
When both soft and hard resets are available, libata preferred softreset till now. The logic behind it was to be softer to devices; however, this doesn't really help much. Rationales for the change: * BIOS may freeze lock certain things during boot and softreset can't unlock those. This by itself is okay but during operation PHY event or other error conditions can trigger hardreset and the device may end up with different configuration. For example, after a hardreset, previously unlockable HPA can be unlocked resulting in different device size and thus revalidation failure. Similar condition can occur during or after resume. * Certain ATAPI devices require hardreset to recover after certain error conditions. On PATA, this is done by issuing the DEVICE RESET command. On SATA, COMRESET has equivalent effect. The problem is that DEVICE RESET needs its own execution protocol. For SFF controllers with bare TF access, it can be easily implemented but more advanced controllers (e.g. ahci and sata_sil24) require specialized implementations. Simply using hardreset solves the problem nicely. * COMRESET initialization sequence is the norm in SATA land and many SATA devices don't work properly if only SRST is used. For example, some PMPs behave this way and libata works around by always issuing hardreset if the host supports PMP. Like the above example, libata has developed a number of mechanisms aiming to promote softreset to hardreset if softreset is not going to work. This approach is time consuming and error prone. Also, note that, dependingon how you read the specs, it could be argued that PMP fan-out ports require COMRESET to start operation. In fact, all the PMPs on the market except one don't work properly if COMRESET is not issued to fan-out ports after PMP reset. * COMRESET is an integral part of SATA connection and any working device should be able to handle COMRESET properly. After all, it's the way to signal hardreset during reboot. This is the most used and recommended (at least by the ahci spec) method of resetting devices. So, this patch makes libata prefer hardreset over softreset by making the following changes. * Rename ATA_EH_RESET_MASK to ATA_EH_RESET and use it whereever ATA_EH_{SOFT|HARD}RESET used to be used. ATA_EH_{SOFT|HARD}RESET is now only used to tell prereset whether soft or hard reset will be issued. * Strip out now unneeded promote-to-hardreset logics from ata_eh_reset(), ata_std_prereset(), sata_pmp_std_prereset() and other places. Signed-off-by: Tejun Heo <htejun@gmail.com>
2008-02-24libata-pmp: clear hob for pmp register accessesMark Lord1-2/+2
>> Mark Lord wrote: >>> Tejun, I've added PMP to sata_mv, and am now trying to get it >>> to work with a Marvell PM attached. >>> >>> And the behaviour I see is very bizarre. >>> >>> After hard+soft resets, the PM signature is found, >>> and libata interrogates the PM registers. >>> >>> It successfully reads register 0, and then register 1. >>> But all subsequent registers read out (incorrectly) as zeros. ... This behavior has been confirmed by Marvell with a SATA analyzer. The Marvell port-multiplier apparently likes to see clean HOB information when accessing PMP registers. Since sata_mv uses PIO shadow register access, this doesn't happen automatically, as it might in a more purely FIS-based driver (eg. ahci). One way to fix this is to flag these commands with ATA_TFLAG_LBA48, forcing libata to write out the HOB fields with known (zero) values. Signed-off-by: Saeed Bishara <saeed@marvell.com> Acked-by: Mark Lord <mlord@pobox.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2008-01-10libata-pmp: 4726 hates SRSTTejun Heo1-7/+5
4726 hates SRST even on non-config ports. Don't use it. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-10-12libata: use ata_exec_internal() for PMP register accessTejun Heo1-117/+92
PMP registers used to be accessed with dedicated accessors ->pmp_read and ->pmp_write. During reset, those callbacks are called with the port frozen so they should be able to run without depending on interrupt delivery. To achieve this, they were implemented polling. However, as resetting the host port makes the PMP to isolate fan-out ports until SError.X is cleared, resetting fan-out ports while port is frozen doesn't buy much additional safety. This patch updates libata PMP support such that PMP registers are accessed using regular ata_exec_internal() mechanism and kills ->pmp_read/write() callbacks. The following changes are made. * PMP access helpers - sata_pmp_read_init_tf(), sata_pmp_read_val(), sata_pmp_write_init_tf() are folded into sata_pmp_read/write() which are now standalone PMP register access functions. * sata_pmp_read/write() returns err_mask instead of rc. This is consistent with other functions which issue internal commands and allows more detailed error reporting. * ahci interrupt handler is modified to ignore BAD_PMP and spurious/illegal completion IRQs while reset is in progress. These conditions are expected during reset. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-10-12libata-pmp: implement qc_defer for command switching PMP supportTejun Heo1-0/+30
Implement sata_pmp_qc_defer_cmd_switch() - standard qc_defer for command switching PMP support. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-10-12libata-pmp: extend ACPI support to cover PMPTejun Heo1-0/+4
Extend ata_acpi_associate_sata_port() such that it can handle PMP and call it when PMP is attached and detached. Build breakage when !CONFIG_ATA_ACPI was spotted and fixed by Petr Vandrovec. Signed-off-by: Tejun Heo <htejun@gmail.com> Cc: Petr Vandrovec <petr@vandrovec.name> Signed-off-by: Jeff Garzik <jeff@garzik.org>
2007-10-12libata-pmp: implement Port Multiplier supportTejun Heo1-0/+1182
Implement Port Multiplier support. To support PMP, a LLDD has to supply ops->pmp_read() and pmp_write(). If non-null, ->pmp_attach and ->pmp_detach are called on PMP attach and detach, respectively. ->pmp_read/write() can be called while the port is frozen, so they must be implemented by polling. This patch supplies several helpers to ease ->pmp_read/write() implementation. Also, irq_handler and error_handler must be PMP aware. Most of PMP aware EH can be done by calling ata_pmp_do_eh() with appropriate methods. PMP EH uses separate set of reset methods and this patch implements standard prereset, hardreset and postreset methods. This patch only implements PMP support. The next patch will integrate PMP into the reset of libata and thus enable PMP support. Signed-off-by: Tejun Heo <htejun@gmail.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>