Commit Graph

98 Commits

Author SHA1 Message Date
Blue Swirl bf4b9889ab ESP: convert to trace framework
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2011-10-01 09:28:40 +00:00
Blue Swirl 9a975d6356 esp: avoid structure holes spotted by pahole
Report from pahole on amd64 host:
struct ESPState {
	SysBusDevice               busdev;               /*     0  5648 */
	/* --- cacheline 88 boundary (5632 bytes) was 16 bytes ago --- */
	uint32_t                   it_shift;             /*  5648     4 */

	/* XXX 4 bytes hole, try to pack */

	qemu_irq                   irq;                  /*  5656     8 */
	uint8_t                    rregs[16];            /*  5664    16 */
	uint8_t                    wregs[16];            /*  5680    16 */
	/* --- cacheline 89 boundary (5696 bytes) --- */
	int32_t                    ti_size;              /*  5696     4 */
	uint32_t                   ti_rptr;              /*  5700     4 */
	uint32_t                   ti_wptr;              /*  5704     4 */
	uint8_t                    ti_buf[16];           /*  5708    16 */
	uint32_t                   status;               /*  5724     4 */
	uint32_t                   dma;                  /*  5728     4 */

	/* XXX 4 bytes hole, try to pack */

	SCSIBus                    bus;                  /*  5736  2120 */
	/* --- cacheline 122 boundary (7808 bytes) was 48 bytes ago --- */
	SCSIDevice *               current_dev;          /*  7856     8 */
	SCSIRequest *              current_req;          /*  7864     8 */
	/* --- cacheline 123 boundary (7872 bytes) --- */
	uint8_t                    cmdbuf[16];           /*  7872    16 */
	uint32_t                   cmdlen;               /*  7888     4 */
	uint32_t                   do_cmd;               /*  7892     4 */
	uint32_t                   dma_left;             /*  7896     4 */
	uint32_t                   dma_counter;          /*  7900     4 */
	uint8_t *                  async_buf;            /*  7904     8 */
	uint32_t                   async_len;            /*  7912     4 */

	/* XXX 4 bytes hole, try to pack */

	ESPDMAMemoryReadWriteFunc  dma_memory_read;      /*  7920     8 */
	ESPDMAMemoryReadWriteFunc  dma_memory_write;     /*  7928     8 */
	/* --- cacheline 124 boundary (7936 bytes) --- */
	void *                     dma_opaque;           /*  7936     8 */
	int                        dma_enabled;          /*  7944     4 */

	/* XXX 4 bytes hole, try to pack */

	void                       (*dma_cb)(ESPState *); /*  7952     8 */

	/* size: 7960, cachelines: 125 */
	/* sum members: 7944, holes: 4, sum holes: 16 */
	/* last cacheline: 24 bytes */
};	/* definitions: 1 */

Fix by rearranging the structure to avoid padding.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2011-08-21 19:52:36 +00:00
Paolo Bonzini c39ce112b6 scsi: pass cdb already to scsi_req_new
Right now the CDB is not passed to the SCSIBus until scsi_req_enqueue.
Passing it to scsi_req_new will let scsi_req_new dispatch common requests
through different reqops.

Moving the memcpy to scsi_req_new is a hack that will go away as
soon as scsi_req_new will also take care of the parsing.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2011-08-12 08:27:37 -05:00
Hannes Reinecke c5bf71a9a3 scsi: Add 'hba_private' to SCSIRequest
'tag' is just an abstraction to identify the command
from the driver. So we should make that explicit by
replacing 'tag' with a driver-defined pointer 'hba_private'.
This saves the lookup for driver handling several commands
in parallel.
'tag' is still being kept for tracing purposes.

Signed-off-by: Hannes Reinecke <hare@suse.de>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2011-07-19 15:38:58 +02:00
Hervé Poussineau 429bef6912 esp: cancel current request only if some request is in flight
This bug was introduced in 94d3f98a3f3caddd7875f9a11776daeb84962a7b:
scsi_cancel_io was checking if some request was pending before trying
to cancel it, while scsi_req_cancel always cancels the request.

This may lead to a crash of Qemu due to dereferencing a NULL pointer,
as exhibited by NetBSD 5.1 installer on MIPS Magnum emulation.

Signed-off-by: Hervé Poussineau <hpoussin@reactos.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2011-07-12 21:29:39 +00:00
Hervé Poussineau 75ef849696 esp: correctly fill bus id with requested lun
This bug showed up after 1455084ea2, and
may be seen only on operating systems *not* using DMA to give commands
to SCSI adapter.

Signed-off-by: Hervé Poussineau <hpoussin@reactos.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2011-07-02 18:50:19 +00:00
Paolo Bonzini aba1f02363 scsi: rename arguments to the new callbacks
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:17 +02:00
Paolo Bonzini c6df7102f5 scsi: split command_complete callback in two
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:17 +02:00
Paolo Bonzini 3944966d95 esp: rename sense to status
This mirrors the LSI patch that was recently committed.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:17 +02:00
Paolo Bonzini 0c34459b6a scsi: introduce scsi_req_get_buf
... and remove some SCSIDevice variables or fields that now become unused.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:16 +02:00
Paolo Bonzini ad3376cc55 scsi: introduce scsi_req_continue
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:16 +02:00
Paolo Bonzini 43a2b33957 scsi: introduce scsi_req_new
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:16 +02:00
Paolo Bonzini fc4f0754c7 scsi: do not call send_command directly
Move the common part of scsi-disk.c and scsi-generic.c to the SCSI layer.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:16 +02:00
Paolo Bonzini 94d3f98a3f scsi: introduce scsi_req_cancel
This is for when the request must be dropped in the void,
but still memory should be freed.  To this end, the devices
register a second callback in SCSIBusOps.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:15 +02:00
Hannes Reinecke 5c6c0e5136 scsi: Use 'SCSIRequest' directly
Currently the SCSIRequest structure is abstracted away and cannot accessed
directly from the driver. This requires the handler to do a lookup on
an abstract 'tag' which identifies the SCSIRequest structure.

With this patch the SCSIRequest structure is exposed to the driver. This
allows use to use it directly as an argument to the SCSIDeviceInfo
callback functions and remove the lookup.

A new callback function 'alloc_req' is introduced matching 'free
req'; unref'ing to free up resources after use is moved into the
scsi_command_complete callbacks.

This temporarily introduces a leak of requests that are cancelled,
when they are removed from the queue and not from the driver.  This
is fixed later by introducing scsi_req_cancel.  That patch in turn
depends on this one, because the argument to scsi_req_cancel is a
SCSIRequest.

Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:15 +02:00
Paolo Bonzini cfdc1bb06e scsi: introduce SCSIBusOps
There are more operations than a SCSI bus can handle, besides completing
commands.  One example, which this series will introduce, is cleaning up
after a request is cancelled.

More long term, a "SCSI bus" can represent the LUNs attached to a
target; in this case, while all commands will ultimately reach a logical
unit, it is the target who is in charge of answering REPORT LUNs.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
2011-05-26 12:14:14 +02:00
Alexander Graf 2507c12ab0 Add endianness as io mem parameter
As stated before, devices can be little, big or native endian. The
target endianness is not of their concern, so we need to push things
down a level.

This patch adds a parameter to cpu_register_io_memory that allows a
device to choose its endianness. For now, all devices simply choose
native endian, because that's the same behavior as before.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-12-11 15:24:25 +00:00
Blue Swirl 73d7434279 ESP: fix ESP DMA access when DMA is not enabled
Sending ESP a command caused it to trigger DMA immediately
even if DMA was not enabled at the DMA controller.

Add a signal from DMA controller to ESP to tell ESP about changes in
DMA enable bit. Also use the correct function for setting up GPIO outputs.

This fixes NetBSD 1.6.1 through 3.0 boot.

Thanks to Artyom Tarasenko for extensive debugging of the problem.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-09-11 16:38:33 +00:00
Markus Armbruster fa66b909f3 scsi: scsi_bus_legacy_handle_cmdline() can fail, fix callers
None of its callers checks for failure.  scsi_hot_add() can crash
because of that:

(qemu) drive_add 4 if=scsi,format=host_device,file=/dev/sg1
scsi-generic: scsi generic interface too old
Segmentation fault (core dumped)

Fix all callers, not just scsi_hot_add().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2010-07-02 13:18:01 +02:00
Blue Swirl 85948643b8 esp: lower IRQ on soft reset
42f1ced228 removed irq lowering
during reset. However, for chip reset command and DMA reset signal,
its actually the correct thing to do.

Lower IRQ on soft reset only.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-06-10 17:57:39 +00:00
Blue Swirl ff9868ecaa esp: use CamelCaseFunc for function types
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-02-07 09:17:35 +00:00
Blue Swirl dca47eddba ESP: improve IRQ debugging
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-01-16 09:06:34 +00:00
Gerd Hoffmann 43b443b668 scsi: move scsi-disk.h -> scsi.h
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-11-09 08:43:03 -06:00
Gerd Hoffmann 1cd3af5480 scsi: move scsi.h -> esp.h
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-11-09 08:43:03 -06:00
Blue Swirl a01d6ef446 sparc32 (mostly): remove unneeded calls to device reset
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-11-07 08:55:24 +00:00
Blue Swirl 63235df8a1 esp: convert to reset + vmsd
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-10-24 16:34:21 +00:00
Markus Armbruster e23a1b33b5 New qdev_init_nofail()
Like qdev_init(), but terminate program via hw_error() instead of
returning an error value.

Use it instead of qdev_init() where terminating the program on failure
is okay, either because it's during machine construction, or because
we know that failure can't happen.

Because relying in the latter is somewhat unclean, and the former is
not always obvious, it would be nice to go back to qdev_init() in the
not-so-obvious cases, only with proper error handling.  I'm leaving
that for another day, because it involves making sure that error
values are properly checked by all callers.

Patchworks-ID: 35168
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-07 08:54:54 -05:00
Gerd Hoffmann ca9c39faed switch scsi bus to inplace allocation.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-05 09:32:46 -05:00
Anthony Liguori c227f0995e Revert "Get rid of _t suffix"
In the very least, a change like this requires discussion on the list.

The naming convention is goofy and it causes a massive merge problem.  Something
like this _must_ be presented on the list first so people can provide input
and cope with it.

This reverts commit 99a0949b72.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-01 16:12:16 -05:00
malc 99a0949b72 Get rid of _t suffix
Some not so obvious bits, slirp and Xen were left alone for the time
being.

Signed-off-by: malc <av1474@comtv.ru>
2009-10-01 22:45:02 +04:00
Blue Swirl cc9952f3ac ESP: convert to VMState
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-09-19 15:44:50 +00:00
Gerd Hoffmann d52affa7f6 qdev/scsi: add scsi bus support to qdev, convert drivers.
* Add SCSIBus.
 * Add SCSIDeviceInfo, move device callbacks here.
 * add qdev/scsi helper functions.
 * convert drivers.

Adding scsi disks via -device works now, i.e. you can do:

 -drive id=sda,if=none,...
 -device lsi
 -device scsi-disk,drive=sda

legacy command lines (-drive if=scsi,...) continue to work.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-09-09 14:57:19 -05:00
Artyom Tarasenko f2818f2201 esp: handle "select without attention"
Up to now "select without attention" was handled the same way as
"select with attention". According to

http://www.ibiblio.org/pub/historic-linux/early-ports/Sparc/NCR/NCR53C9X.txt

select without ATN sends the CDB (Command Descriptor Block) directly,
whereas select with ATN sends one message phase byte
followed by 6, 10, or 12 command phase bytes.

The attached patch implements the behaviour described above.

Signed-off-by: Artyom Tarasenko <atar4qemu@gmail.com>
[blauwirbel@gmail.com: cleaned up formatting]
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-09-05 06:24:47 +00:00
Artyom Tarasenko 4e2a68c1e1 esp (sparc32) Extra scsi data.
> ESP: Message Accepted (12)
> ESP: Transfer status (sense=0)
> ESP: read reg[5]: 0x20
> ESP: read reg[4]: 0x07
> ESP: read reg[7]: 0x02
> Extra scsi data. Fatal error.

It looks like "Message Accepted" shouldn't write a response. At least
ESP_RFLAGS must definetely be 0. With the following fix OBP goes one
step further:

Signed-off-by: Artyom Tarasenko <atar4qemu@gmail.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-08-31 17:12:11 +00:00
Gerd Hoffmann 81a322d4a1 qdev: add return value to init() callbacks.
Sorry folks, but it has to be.  One more of these invasive qdev patches.

We have a serious design bug in the qdev interface:  device init
callbacks can't signal failure because the init() callback has no
return value.  This patch fixes it.

We have already one case in-tree where this is needed:
Try -device virtio-blk-pci (without drive= specified) and watch qemu
segfault.  This patch fixes it.

With usb+scsi being converted to qdev we'll get more devices where the
init callback can fail for various reasons.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-08-27 20:43:28 -05:00
Blue Swirl d60efc6b0d Make CPURead/WriteFunc structure 'const'
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-08-25 18:29:31 +00:00
Blue Swirl 0fd0eb2161 ESP: implement Transfer Pad
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-08-22 13:55:05 +00:00
Blue Swirl 5e1e0a3bb9 ESP: Implement select without ATN, fix comments
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-08-22 13:54:31 +00:00
Blue Swirl 74ff8d90a1 Sparc32: move sparc32_dma init to sun4m.c
Also connect ESP and Lance reset signals to DMA.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-08-08 21:43:12 +00:00
Blue Swirl 2814df28ef esp: fix interrupt register read
Read of interrupt register should clear it and also sequence step and status.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-07-31 07:26:44 +00:00
Gerd Hoffmann ee6847d19b qdev: rework device properties.
This patch is a major overhaul of the device properties.  The properties
are saved directly in the device state struct now, the linked list of
property values is gone.

Advantages:
  * We don't have to maintain the list with the property values.
  * The value in the property list and the value actually used by
    the device can't go out of sync any more (used to happen for
    the pci.devfn == -1 case) because there is only one place where
    the value is stored.
  * A record describing the property is required now, you can't set
    random properties any more.

There are bus-specific and device-specific properties.  The former
should be used for properties common to all bus drivers.  Typical
use case is bus addressing, i.e. pci.devfn and i2c.address.

Properties have a PropertyInfo struct attached with name, size and
function pointers to parse and print properties.  A few common property
types have PropertyInfos defined in qdev-properties.c.  Drivers are free
to implement their own very special property parsers if needed.

Properties can have default values.  If unset they are zero-filled.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-07-16 17:28:51 -05:00
Jan Kiszka a08d43677f Revert "Introduce reset notifier order"
This reverts commit 8217606e6e (and
updates later added users of qemu_register_reset), we solved the
problem it originally addressed less invasively.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-06-29 14:18:08 -05:00
Blue Swirl 42f1ced228 Don't set IRQs on device reset and loadvm/savevm
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-06-17 19:23:44 +00:00
Avi Kivity 1eed09cb4a Remove io_index argument from cpu_register_io_memory()
The parameter is always zero except when registering the three internal
io regions (ROM, unassigned, notdirty).  Remove the parameter to reduce
the API's power, thus facilitating future change.

Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-06-16 15:18:37 -05:00
Paul Brook 067a3ddc88 Remove qdev irq sink handling
We have both IRQ sinks and GPIO inputs.  These are in principle exactly
the same thing, so remove the former.

Signed-off-by: Paul Brook <paul@codesourcery.com>
2009-05-26 14:56:11 +01:00
Jan Kiszka 8217606e6e Introduce reset notifier order
Add the parameter 'order' to qemu_register_reset and sort callbacks on
registration. On system reset, callbacks with lower order will be
invoked before those with higher order. Update all existing users to the
standard order 0.

Note: At least for x86, the existing users seem to assume that handlers
are called in their registration order. Therefore, the patch preserves
this property. If someone feels bored, (s)he could try to identify this
dependency and express it properly on callback registration.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-05-22 10:50:34 -05:00
Paul Brook cfb9de9ce4 ESP SCSI qdev conversion
Signed-off-by: Paul Brook <paul@codesourcery.com>
2009-05-14 22:35:07 +01:00
Blue Swirl 001faf3269 Replace gcc variadic macro extension with C99 version
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-05-13 17:53:17 +00:00
aliguori 487414f1cb hw: remove error handling from qemu_malloc() callers (Avi Kivity)
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>


git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@6529 c046a42c-6fe2-441c-8c8c-71466251a162
2009-02-05 22:06:05 +00:00
blueswir1 e64d7d595f Remove address masking
git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@5853 c046a42c-6fe2-441c-8c8c-71466251a162
2008-12-02 17:47:02 +00:00