dont advertise non-removable sdmmc devices as removable to the scsi layer.

this is cosmetic more than anything cos it gets rid of "removable"
in dmesg output when sd attaches.

handle some more manufacturer ids.

constify SCSI adapter entry points

ok krw@

Fix SDMMC_DEBUG build

- Replace undefined SDMMCDEVNAME macro with usual DEVNAME from sdmmcvar.h
- typofix struct member name

Replace simple one-line bcopy()/memcpy() of faked data into scsi_xfer with call
to scsi_copy_internal_data(). Thus getting xs->resid properly set and adding the
usual uio/size sanity checks.

Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd'
member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic
cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*'
from cmd. Take the address of cmd as required by the various casts.

No intentional functional change.

luna88k test by aoyama@, sparc64 test by jmatthew@

Identification of 2009's last *cmd use and ok jmatthew@

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

constify SCSI adapter entry points

ok krw@

Fix SDMMC_DEBUG build

- Replace undefined SDMMCDEVNAME macro with usual DEVNAME from sdmmcvar.h
- typofix struct member name

Replace simple one-line bcopy()/memcpy() of faked data into scsi_xfer with call
to scsi_copy_internal_data(). Thus getting xs->resid properly set and adding the
usual uio/size sanity checks.

Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd'
member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic
cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*'
from cmd. Take the address of cmd as required by the various casts.

No intentional functional change.

luna88k test by aoyama@, sparc64 test by jmatthew@

Identification of 2009's last *cmd use and ok jmatthew@

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Fix SDMMC_DEBUG build

- Replace undefined SDMMCDEVNAME macro with usual DEVNAME from sdmmcvar.h
- typofix struct member name

Replace simple one-line bcopy()/memcpy() of faked data into scsi_xfer with call
to scsi_copy_internal_data(). Thus getting xs->resid properly set and adding the
usual uio/size sanity checks.

Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd'
member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic
cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*'
from cmd. Take the address of cmd as required by the various casts.

No intentional functional change.

luna88k test by aoyama@, sparc64 test by jmatthew@

Identification of 2009's last *cmd use and ok jmatthew@

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Replace simple one-line bcopy()/memcpy() of faked data into scsi_xfer with call
to scsi_copy_internal_data(). Thus getting xs->resid properly set and adding the
usual uio/size sanity checks.

Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd'
member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic
cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*'
from cmd. Take the address of cmd as required by the various casts.

No intentional functional change.

luna88k test by aoyama@, sparc64 test by jmatthew@

Identification of 2009's last *cmd use and ok jmatthew@

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Since dlg@'s 2009 scsi midlayer refactoring the 'struct scsi_generic *cmd'
member of 'struct scsi_xfer' has always been pointed at the 'struct scsi_generic
cmdstore' member of the same instance. So nuke 'cmdstore' and remove the '*'
from cmd. Take the address of cmd as required by the various casts.

No intentional functional change.

luna88k test by aoyama@, sparc64 test by jmatthew@

Identification of 2009's last *cmd use and ok jmatthew@

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Replace '32' with SID_SCSI2_ALEN (a.k.a. 31) when building emulated INQUIRY
responses. This is what the SCSI specifications say is the correct value and
already used in several cases.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.

When building emulated INQUIRY results use the SCSI_REV_* #define's to
initialize the 'version' field. Not numbers.

Rename [READ|WRITE]_BIG to [READ|WRITE]_10. Rename struct scsi_rw_big to
struct scsi_rw_10.

ok gnezdo@ jmatthew@ (who also did sparc64 compile test)

No need for a bzero(), just init saa.sf to NULL. All the saa.saa
fields are explicitly initialized.

Nuke unused struct scsi_link members of adapter softc's where the
driver successfully compiles on one or more of amd64, i386, hppa.

Move remaining scsi bus initialization info from "prototype scsi link"
fields to struct scsibus_attach_args. Nuke the struct scsi_link *
(saa_sc_link) in scaibus_attach_args.

Explicitly initialize each field in scsibus_attach_args variables.

Move the adapter related items (luns, adapter, adapter_target,
adapter_buswidth, adapter_softc) from struct scsi_link to struct
scsibus_attach_args.

Additional compile tests by jmatthew@ (sparc64) and aoyam@ (luna88k).

Access adapter softc via link->bus->sb_adapter_softc.

In sparc64 autoconf access 'luns' via sb_luns and 'adapter_buswidth'
via sb_adapter_buswidth.

Removes last post-config uses of the copies of bus related
information in scsi_link.

Remove some pointless casting of void * when assigning one
(scsi_link.adapter_softc) to a pointer of a particular softc type.

Pass the expected 'scsibus_attach_args *' to scsibus_attach() via
config_found(), not a 'scsi_link *'. Part of matthew@'s 2010 work.

Relax the grip the dead hand of QBUS (a.k.a. MAXPHYS) has on the
throat of physio().

Allows individual devices to eventually be modified to support larger
physio() (a.k.a. 'raw') i/o's if they prove capable of them.

No immediate functional change.

Tested for many weeks by and ok robert@.

Bring a few stragglers into line with the idiom used by the other 40+
scsi drivers. i.e. eliminate the struct scsi_adapter member in the
softc and rely on the pointer to a static scsi_adapter in the struct
scsi_link member.

Shuffle some names around to make reading the code less headache
inducing. Rename scsi_adapter member 'scsi_minphys' to 'dev_minphys'
to reflect what it is supposed to do. Use consistent naming convention
(<dev>_minphys) for the actual device functions.

No functional change.

Drivers that implement their own *minphys() don't need to call the
system minphys(). scsi_minphys() will do that and cd/sd/st will call
scsi_minphys().

ok jmatthew@ as part of larger diff

Add some sensible error checking in the hibernate io path, helpfully
suggested by coverity (via daniel@)

Add hibernate IO path for sdmmc(4). This requires some help from the
sdmmc chipset driver, currently only implemented in sdhc(4), but
mostly uses the regular path. sdhc(4) also needed the ability to
perform IO while cold.

ok deraadt@

Sizes for the remaining free() calls, and use mallocarray()
ok visa

add sizes to free() calls
ok dhill

p_comm is the process's command and isn't per thread, so move it from
struct proc to struct process.

ok deraadt@ kettenis@

sdmmc sd(4) devices should be marked SID_REMOVABLE, so that eject(1) can
succeed. No downside in the bottom part of the driver.
ok dlg krw

Populate SCSI inquiry data with information from SD and MMC cards.
Code is slightly more convoluted to avoid using strncpy(9).

ok jsg@, millert@, deraadt@

Remove some includes include-what-you-use claims don't
have any direct symbols used. Tested for indirect use by compiling
amd64/i386/sparc64 kernels.

ok tedu@ deraadt@

More malloc() -> mallocarray() in the kernel.

ok deraadt@ tedu@

fix the build when SDMMC_DEBUG is defined

remove uneeded proc.h includes
ok mpi@ kspillner@

add a size argument to free. will be used soon, but for now default to 0.
after discussions with beck deraadt kettenis.

Backout a bunch of my SCSI commits from c2k11. At least one of these
is causing problems when trying to boot sparc64 from an isp(4).

Verified to fix the sparc64/isp(4) regression by krw@; ok deraadt@

Simplify sdmmc_scsi by using SCSI_NO_ADAPTER_TARGET.

Tested by drahn@

Kludge around sdmmc(4)'s poor design: replace "struct scsi_link *"
with "struct scsibus_attach_args" in sdmmc_attach_args.

ok krw@

dont respond to VPD inquiries with standard inquiry data. add a
check to make sure cmdlen is correct while there.

ok miod@

Iopoolification. Tested by mcbride@, dlg@, and Gabriel Kihlman on tech@.

ok dlg@

lockmgr -> rwlock conversion

and with this, my adventure in sdmmc-land is over
bikeshed is still banana-shaped

ok oga@

- fold several occurences of DEVNAME() into one located in sdmmcvar.h.
making DEVNAME() usage a lot more consistent.

ok krw@

Don't use and maintain a function pointer that is only ever set to one
function. Just call the function where needed.

ok krw dlg

New scsi code seems to be stable. Pluck previously identified
low-hanging splbio/splx pairs that are no longer needed and see if
this reveals any hidden scsi flaws.

ok dlg@

Change the scsi_cmd function member of scsi_adapter from int to
void. Use XS_NO_CCB error in the scsi command (xs) to report the
NO_CCB condition. Eliminates all SUCCESSFULLY_QUEUED and COMPLETE
confusion and untangles the midlayer from the adapter a bit more.

Eyes and some fixes by miod@

There may be some compile issues on little used (i.e. I don't have
any) drivers but the change is mechanical and thus easy to remedy.

ok dlg@

Zap all setting of ITSDONE in drivers that don't look at it. Nobody
else cares so it's just noise. Drivers that actually look at ITSDONE
are unchanged.

ok marco@ (for his files) dlg@ beck@

Add missing <sys/device.h>.

ok deraadt@, miod@

scsi_done before COMPLETE

Remove recursion from the sdmmc locking regime.

In snaps for a while.

Originally hacked on phessler@'s couch.
Testing by many, input from jsg@

"I'm tired of seeing the Ms" deraadt@

Remove redundant casts.

Let the sdmmc adapters tell the sdmmc layer how much sectors they can
transfer with one command. Build on this and the recent minphys() changes
in the sdmmc layer to crank transfers at the maximum possible size instead
of a sad DEV_BSIZE.

Depending on your controller, this can speed up sdmmc I/O up to 2.5 times.

Extend the scsi_adapter minphys() callback to take a struct scsi_link *
as additional argument. This will allow intermediate layers between
scsi devices such as sd and scsi host adapters to take appropriate
action if necessary.

Fix crash on zaurus when sd memory card plugged by providing an attachment
structure that slave devices like sbt expect.

Replace timeout_add(msecs * hz / 1000) with timeout_add_msec(msecs).
No functional changes.

ok krw@ miod@

backout: would have been nice if this had been tested, to see that it
crashes the moment a card is plugged in.

Rework sdmmc locking to remove recursion
Add SDMMC_ASSERT_LOCKED(), replacing many locking calls, and sprinkling
the assertion to make sure that future changes don't catch us out

"please commit" uwe@

KNF

prompted and "much better" by marco@, ok pyr@

more MALLOC/FREE -> malloc/free and M_ZERO changes

ok pyr@

give scsi controllers a real attach args to fill in when attaching scsibus.

ok miod@ marco@ deraadt@

dont pass a copy of sdmmcs scsi_link struct on the stack to the midlayer.
it triggered a panic on detach of the scsibus when it tried to use the
adapters scsi_link struct that was no longer there.

issue found by and this fix tested by pedro@
ok krw@ looks right deraadt@

Protect scsi_done() with splbio(), discussed with and okay dlg@, marco@
and fgsch@, uwe@ time-out :(

First and foremost, avoid an obious race between two or more processes
trying to get MMC commands through to the SD/MMC host controller via the
emulated SCSI layer. This is achieved by ensuring exclusive access to
the host controller for one process during any MMC command and during a
sequence of commands at the sdmmc(4) layer.

While at it, the command processing thread has been moved to sdmmc(4),
so as to simplify the implementation of future host controller drivers.
This should also pave the way for further cleanup of the code and for
new drivers.

Minor cleanups are included in this commit, as well.

SDIO card identification

don't #undef SDMMC_DEBUG

Support for standard SD host controllers like the Ricoh 5C822, a small
generic bus layer, and SCSI emulation for SD/MMC memory cards.

Replace '.response_format = 2' with '.response_format = SID_SCSI2_RESPONSE'.