`CONFIG_SATA_LPM_MOBILE_POLICY` reflects a configuration to apply only to
mobile chipsets.  As some desktop boards may want to use this policy by
default as well, rename the configuration item to `SATA_LPM_POLICY`.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

`AHCI_HFLAG_IS_MOBILE` designates that a chipset should be using the
default link power management policy from a kernel configuration item.

As desktop chipsets may also be interested in this default policy
configuration, rename the flag to `AHCI_HFLAG_USE_LPM_POLICY` to more
accurately reflect that a chipset doesn't have to be mobile to adopt it.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

This board definition was originally created for mobile devices to
designate default link power managmeent policy to influence runtime
power consumption.

As this is interesting for more than just mobile designs, rename the
board to `board_ahci_low_power` to make it clear it is about default
policy.
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

After commit e0afcf140e6e ("ata: pata_hpt37x: disable fast interrupts in
prereset() method") HPT370's and HPT372+'s PIO/DMA mode setting functions
have become identical -- merge them.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

In ata_sff_dev_classify(), replace a string of the *if* statements checking
the device's class with the *switch* statement that fits better here...

While at it, fix the multi-line comment style in the vicinity...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Add the explicit error and status register fields to 'struct ata_taskfile'
using the anonymous *union*s ('struct ide_taskfile' had that for ages!) and
update the libata taskfile code accordingly. There should be no object code
changes resulting from that...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

When compile testing the sata gemini driver, CONFIG_OF is required to
avoid the warning:

drivers/ata/sata_gemini.c:421:34: error: ‘gemini_sata_of_match’ defined
but not used [-Werror=unused-const-variable=]
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Replace strings of the *if* statements checking the SCSI command code
with the *switch* statements that fit better here...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The driver's calls to ata_sff_altstatus() are mostly surrounded by some
clumsy checks. Refactor ata_sff_altstatus() to include the repetitive
checks and return a 'bool' result indicating if the alternate status
register exists or not.

While at it, further update the 'kernel-doc' comment -- the alternate
status register has never been a part of the taskfile, despite what
Jeff and co. think! :-)

In ata_sff_lost_interrupt(), wrap the ata_sff_altstatus() call in a
WARN_ON_ONCE() check to issue a warning if the device control register
does not exist. And while at it, fix the strange argument indentation
in the ata_port_warn() call following the call to ata_sff_altstatus().
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Commit 41dec29b ("libata: introduce sff_set_devctl() method") left some
clumsy checks surrounding calls to ata_sff_set_devctl() which Jeff Garzik
suggested to factor out...  and I never followed up. :-(

At last, refactor ata_sff_set_devctl() to include the repetitive checks and
return a 'bool' result indicating if the device control register exists or
not.

While at it, further update the 'kernel-doc' comment -- the device control
register has never been a part of the taskfile, despite what Jeff and co.
think! :-)
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

ata_resources_present() returns 1 if the primary/secondary channel's PCI
resources are present,  0 if not -- the 'bool' type fits somewhat better
here than 'int'...

Use the *= operator, while at it...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The PIO/DMA mode setting function is hardly a good place for disabling
the fast interrupts on a channel -- let's move that code to the driver's
prereset() method instead.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The PIO/DMA mode setting functions are hardly a good place for disabling
the fast interrupts on a channel -- let's move that code to the driver's
prereset() method instead.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The PIO/DMA mode setting function is hardly a good place for disabling
the fast interrupts on a channel -- let's move that code to the driver's
prereset() method instead.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Platform_driver probe functions aren't called with locks held
and thus don't need GFP_ATOMIC. Use GFP_KERNEL instead.

Problem found with Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

This driver has never used the SH-Navi2G/ATAPI-ATA compatible taskfile
registers (the driver uses the taskfile registers in another location
anyway), so drop their #define's...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

HighPoint HPT36x chips did turn out to have the channel enable bits --
however, badly implemented. Make use of them, despite that is probably
only going to burden the driver's code -- assuming both channels are
always enabled by the HighPoint BIOS anyway...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

sata_rcar_ata_devchk() returns 1 if a device is present, 0 if not --
the 'bool' type clearly fits better here than 'unsigned int'...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

pata_s3c_devchk() returns 1 if a device is present, 0 if not -- the 'bool'
type clearly fits better here than 'unsigned int'...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

ata_devchk() returns 1 if a device is present, 0 if not -- the 'bool' type
clearly fits better here than 'unsigned int'...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The driver has never used 'struct hpt_chip' -- drop its declaration.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The driver has never used HPT_PCI_FAST -- drop it.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

This driver uses a string of the *if* statements in atp8xx_fixup() where
a *switch* statement would fit better...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

This driver uses a string of the *if* statements in artop_init_one()
where the *switch* statement would fit better.  While fixing this,
refactor the 6280 code to e.g. avoid a compound statement inside
the *case* section...
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The driver's bmdma_stop() method writes to the wrong PCI config register
(0x52 intead of 0x54) when trying to clear the state machine on secondary
channel -- "luckily", the write falls on a read-only part of the primary
channel MISC. control 3 register, so no collateral damage is done...

Alan Cox fixed the HPT37x driver in commit 6929da44 ("[PATCH] hpt37x:
Two important bug fixes") but forgot to check the HPT3x2N driver which has
the same bug. :-/
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The driver's prereset() method still doesn't check the channel enable bits.
The bug was there for the entire time the driver has existed. :-/

Alan Cox fixed the HPT37x driver in commit b5bf24b9 ("[PATCH] hpt37x:
Check the enablebits") but forgot to check the HPT3x2N driver which has
the same bug. :-/
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

ata_host_suspend() always returns 0, so the result checks in many drivers
look pointless.  Let's make this function return *void* instead of *int*.

Found by Linux Verification Center (linuxtesting.org) with the SVACE static
analysis tool.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

ata_std_prereset() always returns 0, hence the check in ata_sff_prereset()
is pointless and thus it also can return only 0 (however, we cannot change
the prototypes of ata_{sff|std}_prereset() as they implement the driver's
prereset() method).

Found by Linux Verification Center (linuxtesting.org) with the SVACE static
analysis tool.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The 200 ms delay before debouncing the PHY in `sata_link_resume()` is
not needed for the Marvell 88SE9235.

    $ lspci -nn -s 0021:0e:00.0
    0021:0e:00.0 SATA controller [0106]: Marvell Technology Group Ltd. 88SE9235 PCIe 2.0 x2 4-port SATA 6 Gb/s Controller [1b4b:9235] (rev 11)

So, remove it using the board_ahci_no_debounce_delay board definition.

Tested on IBM S822LC with current Linux 5.17-rc1:

Currently, without this patch (with 200 ms delay), device probe for ata1
takes 485 ms:

    [    3.358158] ata1: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3fe881000100 irq 39
    [    3.358175] ata2: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3fe881000180 irq 39
    [    3.358191] ata3: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3fe881000200 irq 39
    [    3.358207] ata4: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3fe881000280 irq 39
    […]
    [    3.677542] ata3: SATA link down (SStatus 0 SControl 300)
    [    3.677719] ata4: SATA link down (SStatus 0 SControl 300)
    [    3.839242] ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
    [    3.839828] ata2.00: ATA-10: ST1000NX0313         00LY266 00LY265IBM, BE33, max UDMA/133
    [    3.840029] ata2.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
    [    3.841796] ata2.00: configured for UDMA/133
    [    3.843231] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
    [    3.844083] ata1.00: ATA-10: ST1000NX0313         00LY266 00LY265IBM, BE33, max UDMA/133
    [    3.844313] ata1.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
    [    3.846043] ata1.00: configured for UDMA/133

With this patch (no delay) device probe for ata1 takes 273 ms:

    [    3.624259] ata1: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3f e881000100 irq 39
    [    3.624436] ata2: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3f e881000180 irq 39
    [    3.624452] ata3: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3f e881000200 irq 39
    [    3.624468] ata4: SATA max UDMA/133 abar m2048@0x3fe881000000 port 0x3f e881000280 irq 39
    […]
    [    3.731966] ata3: SATA link down (SStatus 0 SControl 300)
    [    3.732069] ata4: SATA link down (SStatus 0 SControl 300)
    [    3.897448] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
    [    3.897678] ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
    [    3.898140] ata1.00: ATA-10: ST1000NX0313         00LY266 00LY265IBM, BE33, max UDMA/133
    [    3.898175] ata2.00: ATA-10: ST1000NX0313         00LY266 00LY265IBM, BE33, max UDMA/133
    [    3.898287] ata1.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
    [    3.898349] ata2.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
    [    3.900070] ata1.00: configured for UDMA/133
    [    3.900166] ata2.00: configured for UDMA/133
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Since the commit c05e6ff0 ("libata-acpi: implement
and use ata_acpi_init_gtm()") ata_acpi_on_suspend() just returns 0, so
its call from ata_eh_handle_port_suspend() doesn't make sense anymore.
Remove the function completely, at last...

Found by Linux Verification Center (linuxtesting.org) with the SVACE
static analysis tool.
Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

In scsi_10_lba_len() and scsi_16_lba_len() functions, use
get_unaligned_bexx() to access a cdb LBA and length fields instead of
hardcoding the byte retrieval. With these simplification, the functions
can also be declared inline.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>

Use get_unaligned_be16() instead of using hardcoded accesses to
16-bits big endian cdb fields.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>

Remove the unnecessary "break" after the return statement in the
MODE_SELECT/MODE_SELECT_10 case.
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>

The static array pio_timing is read-only so it make sense to make
it const.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

The static arrays pio_timings and mwdma_timings are read-only so
it make sense to make them const.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Make use of platform_get_mem_or_io() to simplify the code.

While at it, drop use of unlikely() from pata_platform_probe()
as it isn't a hotpath.
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>

Pull irq fixes from Borislav Petkov:

 - Drop an unused private data field in the AIC driver

 - Various fixes to the realtek-rtl driver

 - Make the GICv3 ITS driver compile again in !SMP configurations

 - Force reset of the GICv3 ITSs at probe time to avoid issues during kexec

 - Yet another kfree/bitmap_free conversion

 - Various DT updates (Renesas, SiFive)

* tag 'irq_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  dt-bindings: interrupt-controller: sifive,plic: Group interrupt tuples
  dt-bindings: interrupt-controller: sifive,plic: Fix number of interrupts
  dt-bindings: irqchip: renesas-irqc: Add R-Car V3U support
  irqchip/gic-v3-its: Reset each ITS's BASERn register before probe
  irqchip/gic-v3-its: Fix build for !SMP
  irqchip/loongson-pch-ms: Use bitmap_free() to free bitmap
  irqchip/realtek-rtl: Service all pending interrupts
  irqchip/realtek-rtl: Fix off-by-one in routing
  irqchip/realtek-rtl: Map control data to virq
  irqchip/apple-aic: Drop unused ipi_hwirq field

Pull perf fixes from Borislav Petkov:

 - Prevent accesses to the per-CPU cgroup context list from another CPU
   except the one it belongs to, to avoid list corruption

 - Make sure parent events are always woken up to avoid indefinite hangs
   in the traced workload

* tag 'perf_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/core: Fix cgroup event list management
  perf: Always wake the parent event

Pull scheduler fix from Borislav Petkov:
 "Make sure the membarrier-rseq fence commands are part of the reported
  set when querying membarrier(2) commands through MEMBARRIER_CMD_QUERY"

* tag 'sched_urgent_for_v5.17_rc2_p2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/membarrier: Fix membarrier-rseq fence command missing from query bitmask