aboutsummaryrefslogtreecommitdiff
path: root/hw/scsi
AgeCommit message (Collapse)AuthorFilesLines
2022-01-18hw/dma: Use dma_addr_t type definition when relevantPhilippe Mathieu-Daudé1-5/+5
Update the obvious places where dma_addr_t should be used (instead of uint64_t, hwaddr, size_t, int32_t types). This allows to have &dma_addr_t type portable on 32/64-bit hosts. Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-Id: <20220111184309.28637-11-f4bug@amsat.org>
2022-01-18hw/scsi: Rename SCSIRequest::resid as 'residual'Philippe Mathieu-Daudé3-24/+32
The 'resid' field is slightly confusing and could be interpreted as some ID. Rename it as 'residual' which is clearer to review. No logical change. Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: David Hildenbrand <david@redhat.com> Message-Id: <20220111184309.28637-8-f4bug@amsat.org> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
2022-01-12virtio: unify dataplane and non-dataplane ->handle_output()Stefan Hajnoczi1-48/+6
Now that virtio-blk and virtio-scsi are ready, get rid of the handle_aio_output() callback. It's no longer needed. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-7-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2022-01-12virtio-scsi: prepare virtio_scsi_handle_cmd for dataplaneStefan Hajnoczi1-1/+1
Prepare virtio_scsi_handle_cmd() to be used by both dataplane and non-dataplane by making the condition for starting ioeventfd more specific. This way it won't trigger when dataplane has already been started. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-5-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2022-01-12virtio: get rid of VirtIOHandleAIOOutputStefan Hajnoczi1-12/+6
The virtqueue host notifier API virtio_queue_aio_set_host_notifier_handler() polls the virtqueue for new buffers. AioContext previously required a bool progress return value indicating whether an event was handled or not. This is no longer necessary because the AioContext polling API has been split into a poll check function and an event handler function. The event handler is only run when we know there is work to do, so it doesn't return bool. The VirtIOHandleAIOOutput function signature is now the same as VirtIOHandleOutput. Get rid of the bool return value. Further simplifications will be made for virtio-blk and virtio-scsi in the next patch. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-3-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2022-01-07hw/scsi/vhost-scsi: don't double close vhostfd on errorDaniil Tatianin1-1/+8
vhost_dev_init calls vhost_dev_cleanup on error, which closes vhostfd, don't double close it. Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru> Message-Id: <20211129132358.1110372-2-d-tatianin@yandex-team.ru> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-01-07hw/scsi/vhost-scsi: don't leak vqs on errorDaniil Tatianin1-2/+4
vhost_dev_init calls vhost_dev_cleanup in case of an error during initialization, which zeroes out the entire vsc->dev as well as the vsc->dev.vqs pointer. This prevents us from properly freeing it in free_vqs. Keep a local copy of the pointer so we can free it later. Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru> Message-Id: <20211129132358.1110372-1-d-tatianin@yandex-team.ru> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2021-12-31pci: Let ld*_pci_dma() propagate MemTxResultPhilippe Mathieu-Daudé3-20/+33
ld*_dma() returns a MemTxResult type. Do not discard it, return it to the caller. Update the few callers. Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-24-philmd@redhat.com>
2021-12-31pci: Let ld*_pci_dma() take MemTxAttrs argumentPhilippe Mathieu-Daudé3-11/+18
Let devices specify transaction attributes when calling ld*_pci_dma(). Keep the default MEMTXATTRS_UNSPECIFIED in the few callers. Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-22-philmd@redhat.com>
2021-12-31pci: Let st*_pci_dma() take MemTxAttrs argumentPhilippe Mathieu-Daudé2-6/+12
Let devices specify transaction attributes when calling st*_pci_dma(). Keep the default MEMTXATTRS_UNSPECIFIED in the few callers. Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-21-philmd@redhat.com>
2021-12-31dma: Let dma_buf_read() take MemTxAttrs argumentPhilippe Mathieu-Daudé2-13/+13
Let devices specify transaction attributes when calling dma_buf_read(). Keep the default MEMTXATTRS_UNSPECIFIED in the few callers. Reviewed-by: Klaus Jensen <k.jensen@samsung.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-13-philmd@redhat.com>
2021-12-31dma: Let dma_buf_write() take MemTxAttrs argumentPhilippe Mathieu-Daudé2-2/+2
Let devices specify transaction attributes when calling dma_buf_write(). Keep the default MEMTXATTRS_UNSPECIFIED in the few callers. Reviewed-by: Klaus Jensen <k.jensen@samsung.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-12-philmd@redhat.com>
2021-12-31pci: Let pci_dma_rw() take MemTxAttrs argumentPhilippe Mathieu-Daudé1-1/+1
Let devices specify transaction attributes when calling pci_dma_rw(). Keep the default MEMTXATTRS_UNSPECIFIED in the few callers. Reviewed-by: Klaus Jensen <k.jensen@samsung.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-10-philmd@redhat.com>
2021-12-30dma: Have dma_buf_read() / dma_buf_write() take a void pointerPhilippe Mathieu-Daudé1-11/+11
DMA operations are run on any kind of buffer, not arrays of uint8_t. Convert dma_buf_read/dma_buf_write functions to take a void pointer argument and save us pointless casts to uint8_t *. Remove this pointless casts in the megasas device model. Reviewed-by: Klaus Jensen <k.jensen@samsung.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-9-philmd@redhat.com>
2021-12-30hw/scsi/megasas: Use uint32_t for reply queue head/tail valuesPhilippe Mathieu-Daudé2-6/+6
While the reply queue values fit in 16-bit, they are accessed as 32-bit: 661: s->reply_queue_head = ldl_le_pci_dma(pcid, s->producer_pa); 662: s->reply_queue_head %= MEGASAS_MAX_FRAMES; 663: s->reply_queue_tail = ldl_le_pci_dma(pcid, s->consumer_pa); 664: s->reply_queue_tail %= MEGASAS_MAX_FRAMES; Having: 41:#define MEGASAS_MAX_FRAMES 2048 /* Firmware limit at 65535 */ In order to update the ld/st*_pci_dma() API to pass the address of the value to access, it is simpler to have the head/tail declared as 32-bit values. Replace the uint16_t by uint32_t, wasting 4 bytes in the MegasasState structure. Acked-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211223115554.3155328-20-philmd@redhat.com>
2021-12-18hw/scsi/megasas: Fails command if SGL buffer overflowsPhilippe Mathieu-Daudé1-0/+1
If we detect an overflow on the SGL buffer, do not keep processing the command: discard it. TARGET_FAILURE sense code will be returned (MFI_STAT_SCSI_DONE_WITH_ERROR). Reported-by: Alexander Bulekov <alxndr@bu.edu> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/521 Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Darren Kenny <darren.kenny@oracle.com> Message-Id: <20211119201141.532377-2-philmd@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-11-30hw/scsi/lsi53c895a: Do not abort when DMA requested and no data queuedPhilippe Mathieu-Daudé1-2/+1
If asked for DMA request and no data is available, simply wait for data to be queued, do not abort. This fixes: $ cat << EOF | \ qemu-system-i386 -nographic -M q35,accel=qtest -serial none \ -monitor none -qtest stdio -trace lsi* \ -drive if=none,id=drive0,file=null-co://,file.read-zeroes=on,format=raw \ -device lsi53c895a,id=scsi0 -device scsi-hd,drive=drive0,bus=scsi0.0,channel=0,scsi-id=0,lun=0 lsi_reset Reset lsi_reg_write Write reg DSP2 0x2e = 0xff lsi_reg_write Write reg DSP3 0x2f = 0xff lsi_execute_script SCRIPTS dsp=0xffff0000 opcode 0x184a3900 arg 0x4a8b2d75 qemu-system-i386: hw/scsi/lsi53c895a.c:624: lsi_do_dma: Assertion `s->current' failed. (gdb) bt #5 0x00007ffff4e8a3a6 in __GI___assert_fail (assertion=0x5555560accbc "s->current", file=0x5555560acc28 "hw/scsi/lsi53c895a.c", line=624, function=0x5555560adb18 "lsi_do_dma") at assert.c:101 #6 0x0000555555aa33b9 in lsi_do_dma (s=0x555557805ac0, out=1) at hw/scsi/lsi53c895a.c:624 #7 0x0000555555aa5042 in lsi_execute_script (s=0x555557805ac0) at hw/scsi/lsi53c895a.c:1250 #8 0x0000555555aa757a in lsi_reg_writeb (s=0x555557805ac0, offset=47, val=255 '\377') at hw/scsi/lsi53c895a.c:1984 #9 0x0000555555aa875b in lsi_mmio_write (opaque=0x555557805ac0, addr=47, val=255, size=1) at hw/scsi/lsi53c895a.c:2095 Cc: qemu-stable@nongnu.org Cc: Gerd Hoffmann <kraxel@redhat.com> Cc: Vadim Rozenfeld <vrozenfe@redhat.com> Cc: Stefan Hajnoczi <stefanha@redhat.com> Reported-by: Jérôme Poulin <jeromepoulin@gmail.com> Reported-by: Ruhr-University <bugs-syssec@rub.de> Reported-by: Gaoning Pan <pgn@zju.edu.cn> Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr> Fixes: b96a0da06bd ("lsi: move dma_len+dma_buf into lsi_request") BugLink: https://bugs.launchpad.net/qemu/+bug/697510 BugLink: https://bugs.launchpad.net/qemu/+bug/1905521 BugLink: https://bugs.launchpad.net/qemu/+bug/1908515 Resolves: https://gitlab.com/qemu-project/qemu/-/issues/84 Resolves: https://gitlab.com/qemu-project/qemu/-/issues/305 Resolves: https://gitlab.com/qemu-project/qemu/-/issues/552 Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Laurent Vivier <lvivier@redhat.com> Message-Id: <20211123111732.83137-2-philmd@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-11-19esp: ensure that async_len is reset to 0 during esp_hard_reset()Mark Cave-Ayland1-0/+1
If a reset command is sent after data has been transferred into the SCSI buffer ensure that async_len is reset to 0. Otherwise a subsequent TI command assumes the SCSI buffer contains data to be transferred to the device causing it to dereference the stale async_buf pointer. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Fixes: https://gitlab.com/qemu-project/qemu/-/issues/724 Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20211118100327.29061-2-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-11-08hw/scsi/scsi-disk: MODE_PAGE_ALLS not allowed in MODE SELECT commandsMauro Matteo Cascella1-0/+6
This avoids an off-by-one read of 'mode_sense_valid' buffer in hw/scsi/scsi-disk.c:mode_sense_page(). Fixes: CVE-2021-3930 Cc: qemu-stable@nongnu.org Reported-by: Alexander Bulekov <alxndr@bu.edu> Fixes: a8f4bbe2900 ("scsi-disk: store valid mode pages in a table") Fixes: #546 Reported-by: Qiuhao Li <Qiuhao.Li@outlook.com> Signed-off-by: Mauro Matteo Cascella <mcascell@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-11-02esp: ensure in-flight SCSI requests are always cancelledMark Cave-Ayland1-5/+5
There is currently a check in esp_select() to cancel any in-flight SCSI requests to ensure that issuing multiple select commands without continuing through the rest of the ESP state machine ignores all but the last SCSI request. This is also enforced through the addition of assert()s in esp_transfer_data() and scsi_read_data(). The get_cmd() function does not call esp_select() when TC == 0 which means it is possible for a fuzzer to trigger these assert()s by sending a select command when TC == 0 immediately after a valid SCSI CDB has been submitted. Since esp_select() is only called from get_cmd(), hoist the check to cancel in-flight SCSI requests from esp_select() into get_cmd() to ensure it is always called when executing a select command to initiate a new SCSI request. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Closes: https://gitlab.com/qemu-project/qemu/-/issues/662 Closes: https://gitlab.com/qemu-project/qemu/-/issues/663 Message-Id: <20211101183516.8455-2-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-10-06block: introduce max_hw_iov for use in scsi-genericPaolo Bonzini1-1/+1
Linux limits the size of iovecs to 1024 (UIO_MAXIOV in the kernel sources, IOV_MAX in POSIX). Because of this, on some host adapters requests with many iovecs are rejected with -EINVAL by the io_submit() or readv()/writev() system calls. In fact, the same limit applies to SG_IO as well. To fix both the EINVAL and the possible performance issues from using fewer iovecs than allowed by Linux (some HBAs have max_segments as low as 128), introduce a separate entry in BlockLimits to hold the max_segments value from sysfs. This new limit is used only for SG_IO and clamped to bs->bl.max_iov anyway, just like max_hw_transfer is clamped to bs->bl.max_transfer. Reported-by: Halil Pasic <pasic@linux.ibm.com> Cc: Hanna Reitz <hreitz@redhat.com> Cc: Kevin Wolf <kwolf@redhat.com> Cc: qemu-block@nongnu.org Cc: qemu-stable@nongnu.org Fixes: 18473467d5 ("file-posix: try BLKSECTGET on block devices too, do not round to power of 2", 2021-06-25) Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20210923130436.1187591-1-pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2021-09-30qbus: Rename qbus_create_inplace() to qbus_init()Peter Maydell1-1/+1
Rename qbus_create_inplace() to qbus_init(); this is more in line with our usual naming convention for functions that in-place initialize objects. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Message-id: 20210923121153.23754-5-peter.maydell@linaro.org
2021-09-30scsi: Replace scsi_bus_new() with scsi_bus_init(), scsi_bus_init_named()Peter Maydell9-14/+11
The function scsi_bus_new() creates a new SCSI bus; callers can either pass in a name argument to specify the name of the new bus, or they can pass in NULL to allow the bus to be given an automatically generated unique name. Almost all callers want to use the autogenerated name; the only exception is the virtio-scsi device. Taking a name argument that should almost always be NULL is an easy-to-misuse API design -- it encourages callers to think perhaps they should pass in some standard name like "scsi" or "scsi-bus". We don't do this anywhere for SCSI, but we do (incorrectly) do it for other bus types such as i2c. The function name also implies that it will return a newly allocated object, when it in fact does in-place allocation. We more commonly name such functions foo_init(), with foo_new() being the allocate-and-return variant. Replace all the scsi_bus_new() callsites with either: * scsi_bus_init() for the usual case where the caller wants an autogenerated bus name * scsi_bus_init_named() for the rare case where the caller needs to specify the bus name and document that for the _named() version it's then the caller's responsibility to think about uniqueness of bus names. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Acked-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20210923121153.23754-2-peter.maydell@linaro.org
2021-08-26vhost-scsi: Plug memory leak on migrate_add_blocker() failureMarkus Armbruster1-2/+2
Cc: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com> Message-Id: <20210720125408.387910-6-armbru@redhat.com> Acked-by: Michael S. Tsirkin <mst@redhat.com>
2021-07-02virtio: Clarify MR transaction optimizationGreg Kurz1-0/+16
The device model batching its ioeventfds in a single MR transaction is an optimization. Clarify this in virtio-scsi, virtio-blk and generic virtio code. Also clarify that the transaction must commit before closing ioeventfds so that no one is tempted to merge the loops in the start functions error path and in the stop functions. Signed-off-by: Greg Kurz <groug@kaod.org> Message-Id: <162125799728.1394228.339855768563326832.stgit@bahia.lan> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2021-06-30vhost: Add Error parameter to vhost_dev_init()Kevin Wolf2-6/+2
This allows callers to return better error messages instead of making one up while the real error ends up on stderr. Most callers can immediately make use of this because they already have an Error parameter themselves. The others just keep printing the error with error_report_err(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20210609154658.350308-2-kwolf@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2021-06-25block: add max_hw_transfer to BlockLimitsPaolo Bonzini1-1/+1
For block host devices, I/O can happen through either the kernel file descriptor I/O system calls (preadv/pwritev, io_submit, io_uring) or the SCSI passthrough ioctl SG_IO. In the latter case, the size of each transfer can be limited by the HBA, while for file descriptor I/O the kernel is able to split and merge I/O in smaller pieces as needed. Applying the HBA limits to file descriptor I/O results in more system calls and suboptimal performance, so this patch splits the max_transfer limit in two: max_transfer remains valid and is used in general, while max_hw_transfer is limited to the maximum hardware size. max_hw_transfer can then be included by the scsi-generic driver in the block limits page, to ensure that the stricter hardware limit is used. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-25scsi-generic: pass max_segments via max_iov field in BlockLimitsPaolo Bonzini1-2/+4
I/O to a disk via read/write is not limited by the number of segments allowed by the host adapter; the kernel can split requests if needed, and the limit imposed by the host adapter can be very low (256k or so) to avoid that SG_IO returns EINVAL if memory is heavily fragmented. Since this value is only interesting for SG_IO-based I/O, do not include it in the max_transfer and only take it into account when patching the block limits VPD page in the scsi-generic device. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
2021-06-15esp: store lun coming from the MESSAGE OUT phasePaolo Bonzini2-17/+33
The LUN is selected with an IDENTIFY message, and persists until the next message out phase. Instead of passing it to do_busid_cmd, store it in ESPState. Because do_cmd can simply skip the message out phase if cmdfifo_cdb_offset is zero, it can now be used for the S without ATN cases as well. Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Tested-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: fix migration version check in esp_is_version_5()Mark Cave-Ayland1-1/+1
Commit 4e78f3bf35 "esp: defer command completion interrupt on incoming data transfers" added a version check for use with VMSTATE_*_TEST macros to allow migration from older QEMU versions. Unfortunately the version check fails to work in its current form since if the VMStateDescription version_id is incremented, the test returns false and so the fields are not included in the outgoing migration stream. Change the version check to use >= rather == to ensure that migration works correctly when the ESPState VMStateDescription has version_id > 5. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Fixes: 4e78f3bf35 ("esp: defer command completion interrupt on incoming data transfers") Message-Id: <20210613102614.5438-1-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: correctly accumulate extended messages for PDMAMark Cave-Ayland1-3/+20
Commit 799d90d818 "esp: transition to message out phase after SATN and stop command" added logic to correctly handle extended messages for DMA requests but not for PDMA requests. Apply the same logic in esp_do_dma() to do_dma_pdma_cb() so that extended messages terminated with a PDMA request are accumulated correctly. This allows the ESP device to respond correctly to the SDTR negotiation initiated by the NetBSD ESP driver without causing errors and timeouts on boot. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Message-Id: <20210519100803.10293-6-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: revert 75ef849696 "esp: correctly fill bus id with requested lun"Mark Cave-Ayland1-3/+0
This commit from nearly 10 years ago is now broken due to the improvements in esp emulation (or perhaps was never correct). It shows up as a bug in detecting the CDROM drive under MacOS. The error is caused by the MacOS CDROM driver sending this CDB with an "S without ATN" command and without DMA: 0x12 0x00 0x00 0x00 0x05 0x00 (INQUIRY) This is a valid INQUIRY command, however with this logic present the 3rd byte (0x0) is copied over the 1st byte (0x12) which silently converts the INQUIRY command to a TEST UNIT READY command before passing it to the QEMU SCSI layer. Since the TEST UNIT READY command has a zero length response the MacOS CDROM driver never receives a response and assumes the CDROM is not present. The logic was to ignore the IDENTIFY byte and copy the LUN over from the CDB, which did store the LUN in bits 5-7 of the second byte in olden times. This however is all obsolete, so just drop the code. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Message-Id: <20210519100803.10293-5-mark.cave-ayland@ilande.co.uk> [Tweaked commit message. - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: ensure PDMA write transfers are flushed from the FIFO to the target ↵Mark Cave-Ayland1-14/+11
immediately After each PDMA write transfer the MacOS CDROM driver waits until the FIFO is empty (i.e. its contents have been written out to the SCSI bus) by polling the FIFO count register until it reads 0. This doesn't work with the current PDMA write implementation which waits until either the FIFO is full or the transfer is complete before invoking the PDMA callback to process the FIFO contents. Change the PDMA write transfer logic so that the PDMA callback is invoked after each PDMA write to transfer the FIFO contents to the target buffer immediately, and hence avoid getting stuck in the FIFO count register polling loop. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Message-Id: <20210519100803.10293-4-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: handle non-DMA transfers from the target one byte at a timeMark Cave-Ayland1-22/+50
The initial implementation of non-DMA transfers was based upon analysis of traces from the MacOS toolbox ROM for handling unaligned reads but missed one key aspect - during a non-DMA transfer from the target, the bus service interrupt should be raised for every single byte received from the bus and not just at either the end of the transfer or when the FIFO is full. Adjust the non-DMA code accordingly so that esp_do_nodma() is called for every byte received from the target. This also includes special handling for managing the change from DATA IN to STATUS phase as this needs to occur when the final byte is read out from the FIFO, and not at the end of the transfer of the last byte into the FIFO. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Message-Id: <20210519100803.10293-3-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: allow non-DMA callback in esp_transfer_data() initial transferMark Cave-Ayland1-12/+2
The current implementation only resumes DMA transfers when incoming data is received from the target device, but this is also required for non-DMA transfers with the next set of non-DMA changes. Rather than duplicate the DMA/non-DMA dispatch logic in the initial transfer section, update the code so that the initial transfer section can just fallthrough to the main DMA/non-DMA dispatch logic. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Message-Id: <20210519100803.10293-2-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: only set ESP_RSEQ at the start of the select sequenceMark Cave-Ayland1-2/+9
When processing a command to select a target and send a CDB, the ESP device maintains a sequence step register so that if an error occurs the host can determine which part of the selection/CDB submission sequence failed. The old Linux 2.6 driver is really pedantic here: it checks the sequence step register even if a command succeeds and complains loudly on the console if the sequence step register doesn't match the expected bus phase and interrupt flags. This reason this mismatch occurs is because the ESP emulation currently doesn't update the bus phase until the next TI (Transfer Information) command and so the cleared sequence step register is considered invalid for the stale bus phase. Normally this isn't an issue as the host only checks the sequence step register if an error occurs but the old Linux 2.6 driver does this in several places causing a large stream of "esp0: STEP_ASEL for tgt 0" messages to appear on the console during the boot process. Fix this by not clearing the sequence step register when reading the interrupt register and clearing the DMA status, so the guest sees a valid sequence step and bus phase combination at the end of the command phase. No other change is required since the sequence step register is correctly updated throughout the selection/CDB submission sequence once one of the select commands is issued. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Fixes: 1b9e48a5bd ("esp: implement non-DMA transfers in PDMA mode") Message-Id: <20210518212511.21688-3-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15esp: only assert INTR_DC interrupt flag if selection failsMark Cave-Ayland1-1/+1
The datasheet sequence tables confirm that when a target selection fails, only the INTR_DC interrupt flag should be asserted. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Fixes: cf47a41e05 ("esp: latch individual bits in ESP_RINTR register") Message-Id: <20210518212511.21688-2-mark.cave-ayland@ilande.co.uk> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-15block/scsi: correctly emulate the VPD block limits pageMaxim Levitsky1-2/+7
When the device doesn't support the VPD block limits page, we emulate it even for SCSI passthrough. As a part of the emulation we need to add it to the 'Supported VPD Pages' The code that does this adds it to the page, but it doesn't increase the length of the data to be copied to the guest, thus the guest never sees the VPD block limits page as supported. Bump the transfer size by 1 in this case. Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20201217165612.942849-6-mlevitsk@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-06-02docs: fix references to docs/devel/tracing.rstStefano Garzarella1-1/+1
Commit e50caf4a5c ("tracing: convert documentation to rST") converted docs/devel/tracing.txt to docs/devel/tracing.rst. We still have several references to the old file, so let's fix them with the following command: sed -i s/tracing.txt/tracing.rst/ $(git grep -l docs/devel/tracing.txt) Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20210517151702.109066-2-sgarzare@redhat.com> Signed-off-by: Thomas Huth <thuth@redhat.com>
2021-05-26hw/scsi: Fix sector translation bug in scsi_unmap_complete_noioKit Westneat1-5/+7
check_lba_range expects sectors to be expressed in original qdev blocksize, but scsi_unmap_complete_noio was translating them to 512 block sizes, which was causing sense errors in the larger LBAs in devices using a 4k block size. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/345 Signed-off-by: Kit Westneat <kit.westneat@gmail.com> Message-Id: <20210521142829.326217-1-kit.westneat@gmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2021-05-16Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into stagingPeter Maydell1-18/+38
pc,pci,virtio: bugfixes, improvements Fixes all over the place. Faster boot for virtio. ioeventfd support for mmio. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> # gpg: Signature made Fri 14 May 2021 15:27:13 BST # gpg: using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469 # gpg: issuer "mst@redhat.com" # gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full] # gpg: aka "Michael S. Tsirkin <mst@redhat.com>" [full] # Primary key fingerprint: 0270 606B 6F3C DF3D 0B17 0970 C350 3912 AFBE 8E67 # Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA 8A0D 281F 0DB8 D28D 5469 * remotes/mst/tags/for_upstream: Fix build with 64 bits time_t vhost-vdpa: Make vhost_vdpa_get_device_id() static hw/virtio: enable ioeventfd configuring for mmio hw/smbios: support for type 41 (onboard devices extended information) checkpatch: Fix use of uninitialized value virtio-scsi: Configure all host notifiers in a single MR transaction virtio-scsi: Set host notifiers and callbacks separately virtio-blk: Configure all host notifiers in a single MR transaction virtio-blk: Fix rollback path in virtio_blk_data_plane_start() pc-dimm: remove unnecessary get_vmstate_memory_region() method amd_iommu: fix wrong MMIO operations virtio-net: Constify VirtIOFeature feature_sizes[] virtio-blk: Constify VirtIOFeature feature_sizes[] hw/virtio: Pass virtio_feature_get_config_size() a const argument x86: acpi: use offset instead of pointer when using build_header() amd_iommu: Fix pte_override_page_mask() Signed-off-by: Peter Maydell <peter.maydell@linaro.org> # Conflicts: # hw/arm/virt.c
2021-05-14virtio-scsi: Configure all host notifiers in a single MR transactionGreg Kurz1-0/+16
This allows the virtio-scsi-pci device to batch the setup of all its host notifiers. This significantly improves boot time of VMs with a high number of vCPUs, e.g. from 6m5.563s down to 1m2.884s for a pseries machine with 384 vCPUs. Note that memory_region_transaction_commit() must be called before virtio_bus_cleanup_host_notifier() because the latter might close ioeventfds that the transaction still assumes to be around when it commits. Signed-off-by: Greg Kurz <groug@kaod.org> Message-Id: <20210407143501.244343-5-groug@kaod.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2021-05-14virtio-scsi: Set host notifiers and callbacks separatelyGreg Kurz1-18/+22
Host notifiers are guaranteed to be idle until the callbacks are hooked up with virtio_queue_aio_set_host_notifier_handler(). They thus don't need to be set or unset with the AioContext lock held. Do this outside the critical section, like virtio-blk already does : basically downgrading virtio_scsi_vring_init() to only setup the host notifier and set the callback in the caller. This will allow to batch addition/deletion of ioeventds in a single memory transaction, which is expected to greatly improve initialization time. Signed-off-by: Greg Kurz <groug@kaod.org> Message-Id: <20210407143501.244343-4-groug@kaod.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2021-05-02Do not include cpu.h if it's not really necessaryThomas Huth1-1/+0
Stop including cpu.h in files that don't need it. Signed-off-by: Thomas Huth <thuth@redhat.com> Message-Id: <20210416171314.2074665-4-thuth@redhat.com> Signed-off-by: Laurent Vivier <laurent@vivier.eu>
2021-04-19mptsas: Remove unused MPTSASState 'pending' field (CVE-2021-3392)Michael Tokarev2-7/+0
While processing SCSI i/o requests in mptsas_process_scsi_io_request(), the Megaraid emulator appends new MPTSASRequest object 'req' to the 's->pending' queue. In case of an error, this same object gets dequeued in mptsas_free_request() only if SCSIRequest object 'req->sreq' is initialised. This may lead to a use-after-free issue. Since s->pending is actually not used, simply remove it from MPTSASState. Cc: qemu-stable@nongnu.org Signed-off-by: Michael Tokarev <mjt@tls.msk.ru> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr> Message-id: 20210419134247.1467982-1-f4bug@amsat.org Message-Id: <20210416102243.1293871-1-mjt@msgid.tls.msk.ru> Suggested-by: Paolo Bonzini <pbonzini@redhat.com> Reported-by: Cheolwoo Myung <cwmyung@snu.ac.kr> BugLink: https://bugs.launchpad.net/qemu/+bug/1914236 (CVE-2021-3392) Fixes: e351b826112 ("hw: Add support for LSI SAS1068 (mptsas) device") [PMD: Reworded description, added more tags] Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
2021-04-12esp: ensure that do_cmd is set to zero before submitting an ESP select commandMark Cave-Ayland1-0/+2
When a CDB has been received and is about to be submitted to the SCSI layer via one of the ESP select commands, ensure that do_cmd is set to zero before executing the command. Otherwise a guest executing 2 valid CDBs in quick sequence can invoke the SCSI .transfer_data callback again before do_cmd is set to zero by the callback function triggering an assert at the start of esp_transfer_data(). Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-Id: <20210407195801.685-12-mark.cave-ayland@ilande.co.uk>
2021-04-12esp: don't reset async_len directly in esp_select() if cancelling requestMark Cave-Ayland1-1/+1
Instead let the SCSI layer invoke the .cancel callback itself to cancel and reset the request state. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Tested-by: Alexander Bulekov <alxndr@bu.edu> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Message-Id: <20210407195801.685-11-mark.cave-ayland@ilande.co.uk>
2021-04-12esp: don't overflow cmdfifo if TC is larger than the cmdfifo sizeMark Cave-Ayland1-0/+1
If a guest transfers the message out/command phase data using DMA with a TC that is larger than the cmdfifo size then the cmdfifo overflows triggering an assert. Limit the size of the transfer to the free space available in cmdfifo. Buglink: https://bugs.launchpad.net/qemu/+bug/1919036 Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Message-Id: <20210407195801.685-10-mark.cave-ayland@ilande.co.uk>
2021-04-12esp: don't overflow cmdfifo in get_cmd()Mark Cave-Ayland1-0/+2
If the guest tries to read a CDB using DMA and cmdfifo is not empty then it is possible to overflow cmdfifo. Since this can only occur by issuing deliberately incorrect instruction sequences, ensure that the maximum length of the CDB transferred to cmdfifo is limited to the available free space within cmdfifo. Buglink: https://bugs.launchpad.net/qemu/+bug/1909247 Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Message-Id: <20210407195801.685-9-mark.cave-ayland@ilande.co.uk>
2021-04-12esp: don't underflow cmdfifo in do_cmd()Mark Cave-Ayland1-2/+4
If the guest tries to execute a CDB when cmdfifo is not empty before the start of the message out phase then clearing the message out phase data will cause cmdfifo to underflow due to cmdfifo_cdb_offset being larger than the amount of data within. Since this can only occur by issuing deliberately incorrect instruction sequences, ensure that the maximum length of esp_fifo_pop_buf() is limited to the size of the data within cmdfifo. Buglink: https://bugs.launchpad.net/qemu/+bug/1909247 Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Tested-by: Alexander Bulekov <alxndr@bu.edu> Message-Id: <20210407195801.685-8-mark.cave-ayland@ilande.co.uk>
generated by cgit v1.1 at 2024-12-28 19:14:58 +0000