* [Qemu-devel] [PULL v2 0/2] SCSI changes for 2013-12-22
@ 2013-12-22 14:02 Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 1/2] block/iscsi: use a bh to schedule co reentrance Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 2/2] scsi-disk: add UNMAP limits to block limits VPD page Paolo Bonzini
0 siblings, 2 replies; 3+ messages in thread
From: Paolo Bonzini @ 2013-12-22 14:02 UTC (permalink / raw)
To: qemu-devel
Anthony,
The following changes since commit e157b8fdd412d48eacfbb8c67d3d58780154faa3:
Merge remote-tracking branch 'bonzini/virtio' into staging (2013-12-13 11:10:33 -0800)
are available in the git repository at:
git://github.com/bonzini/qemu.git scsi-next
for you to fetch changes up to 8a1bd2973ed5f99a3c37c9afdff823c4a22152b1:
scsi-disk: add UNMAP limits to block limits VPD page (2013-12-22 14:59:33 +0100)
Includes a new version of the scsi-disk patch.
----------------------------------------------------------------
Paolo Bonzini (1):
scsi-disk: add UNMAP limits to block limits VPD page
Peter Lieven (1):
block/iscsi: use a bh to schedule co reentrance
block/iscsi.c | 11 ++++++++++-
hw/scsi/scsi-disk.c | 20 ++++++++++++++++++++
2 files changed, 30 insertions(+), 1 deletion(-)
--
1.8.4.2
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Qemu-devel] [PULL v2 1/2] block/iscsi: use a bh to schedule co reentrance
2013-12-22 14:02 [Qemu-devel] [PULL v2 0/2] SCSI changes for 2013-12-22 Paolo Bonzini
@ 2013-12-22 14:02 ` Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 2/2] scsi-disk: add UNMAP limits to block limits VPD page Paolo Bonzini
1 sibling, 0 replies; 3+ messages in thread
From: Paolo Bonzini @ 2013-12-22 14:02 UTC (permalink / raw)
To: qemu-devel; +Cc: Peter Lieven, qemu-stable
From: Peter Lieven <pl@kamp.de>
this fixes a potential segfault and performance regression.
If the coroutine is reentered directly in the iscsi_co_generic_cb
iscsi_process_{read,write} are interrupted and reentered any
time later. One the one hand this could happen after an iscsi_close
where the iscsi context is already gone (segfault). On the
other hand this limits the number of processed callbacks
in each aio_dispatch to one (potential performance regression).
Cc: qemu-stable@nongnu.org
Signed-off-by: Peter Lieven <pl@kamp.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
block/iscsi.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/block/iscsi.c b/block/iscsi.c
index fa69408..b0e6eea 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -68,6 +68,7 @@ typedef struct IscsiTask {
int do_retry;
struct scsi_task *task;
Coroutine *co;
+ QEMUBH *bh;
} IscsiTask;
typedef struct IscsiAIOCB {
@@ -123,6 +124,13 @@ iscsi_schedule_bh(IscsiAIOCB *acb)
qemu_bh_schedule(acb->bh);
}
+static void iscsi_co_generic_bh_cb(void *opaque)
+{
+ struct IscsiTask *iTask = opaque;
+ qemu_bh_delete(iTask->bh);
+ qemu_coroutine_enter(iTask->co, NULL);
+}
+
static void
iscsi_co_generic_cb(struct iscsi_context *iscsi, int status,
void *command_data, void *opaque)
@@ -147,7 +155,8 @@ iscsi_co_generic_cb(struct iscsi_context *iscsi, int status,
out:
if (iTask->co) {
- qemu_coroutine_enter(iTask->co, NULL);
+ iTask->bh = qemu_bh_new(iscsi_co_generic_bh_cb, iTask);
+ qemu_bh_schedule(iTask->bh);
}
}
--
1.8.4.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
* [Qemu-devel] [PULL v2 2/2] scsi-disk: add UNMAP limits to block limits VPD page
2013-12-22 14:02 [Qemu-devel] [PULL v2 0/2] SCSI changes for 2013-12-22 Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 1/2] block/iscsi: use a bh to schedule co reentrance Paolo Bonzini
@ 2013-12-22 14:02 ` Paolo Bonzini
1 sibling, 0 replies; 3+ messages in thread
From: Paolo Bonzini @ 2013-12-22 14:02 UTC (permalink / raw)
To: qemu-devel
Linux prefers WRITE SAME to UNMAP if the limits are zero, and WRITE
SAME does not discard anything unless the device can guarantee that
the resulting block is zero.
Setting the maximum unmap block and descriptor counts to non-zero
makes Linux choose UNMAP and fixes thin provisioning on glusterfs.
While the maximum unmap block count can have some effect on performance,
the (suggested) maximum number of descriptors is not particularly
important so I didn't add a customization option. SCSI drivers are
used to online firmware updates so I'm not yet adding versioning support
for SCSI, but we're probably getting close to the point when it's worth
thinking about it.
Reported-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/scsi/scsi-disk.c | 20 ++++++++++++++++++++
1 file changed, 20 insertions(+)
diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
index 7653411..f6ccccc 100644
--- a/hw/scsi/scsi-disk.c
+++ b/hw/scsi/scsi-disk.c
@@ -47,6 +47,7 @@ do { printf("scsi-disk: " fmt , ## __VA_ARGS__); } while (0)
#define SCSI_MAX_MODE_LEN 256
#define DEFAULT_DISCARD_GRANULARITY 4096
+#define DEFAULT_MAX_UNMAP_SIZE (1 << 30) /* 1 GB */
typedef struct SCSIDiskState SCSIDiskState;
@@ -74,6 +75,7 @@ struct SCSIDiskState
bool media_event;
bool eject_request;
uint64_t wwn;
+ uint64_t max_unmap_size;
QEMUBH *bh;
char *version;
char *serial;
@@ -625,6 +627,8 @@ static int scsi_disk_emulate_inquiry(SCSIRequest *req, uint8_t *outbuf)
s->qdev.conf.min_io_size / s->qdev.blocksize;
unsigned int opt_io_size =
s->qdev.conf.opt_io_size / s->qdev.blocksize;
+ unsigned int max_unmap_sectors =
+ s->max_unmap_size / s->qdev.blocksize;
if (s->qdev.type == TYPE_ROM) {
DPRINTF("Inquiry (EVPD[%02X] not supported for CDROM\n",
@@ -647,6 +651,18 @@ static int scsi_disk_emulate_inquiry(SCSIRequest *req, uint8_t *outbuf)
outbuf[14] = (opt_io_size >> 8) & 0xff;
outbuf[15] = opt_io_size & 0xff;
+ /* max unmap LBA count, default is 1GB */
+ outbuf[20] = (max_unmap_sectors >> 24) & 0xff;
+ outbuf[21] = (max_unmap_sectors >> 16) & 0xff;
+ outbuf[22] = (max_unmap_sectors >> 8) & 0xff;
+ outbuf[23] = max_unmap_sectors & 0xff;
+
+ /* max unmap descriptors, 255 fit in 4 kb with an 8-byte header. */
+ outbuf[24] = 0;
+ outbuf[25] = 0;
+ outbuf[26] = 0;
+ outbuf[27] = 255;
+
/* optimal unmap granularity */
outbuf[28] = (unmap_sectors >> 24) & 0xff;
outbuf[29] = (unmap_sectors >> 16) & 0xff;
@@ -2519,6 +2535,8 @@ static Property scsi_hd_properties[] = {
DEFINE_PROP_BIT("dpofua", SCSIDiskState, features,
SCSI_DISK_F_DPOFUA, false),
DEFINE_PROP_HEX64("wwn", SCSIDiskState, wwn, 0),
+ DEFINE_PROP_UINT64("max_unmap_size", SCSIDiskState, max_unmap_size,
+ DEFAULT_MAX_UNMAP_SIZE),
DEFINE_BLOCK_CHS_PROPERTIES(SCSIDiskState, qdev.conf),
DEFINE_PROP_END_OF_LIST(),
};
@@ -2628,6 +2646,8 @@ static Property scsi_disk_properties[] = {
DEFINE_PROP_BIT("dpofua", SCSIDiskState, features,
SCSI_DISK_F_DPOFUA, false),
DEFINE_PROP_HEX64("wwn", SCSIDiskState, wwn, 0),
+ DEFINE_PROP_UINT64("max_unmap_size", SCSIDiskState, max_unmap_size,
+ DEFAULT_MAX_UNMAP_SIZE),
DEFINE_PROP_END_OF_LIST(),
};
--
1.8.4.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2013-12-22 14:02 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-12-22 14:02 [Qemu-devel] [PULL v2 0/2] SCSI changes for 2013-12-22 Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 1/2] block/iscsi: use a bh to schedule co reentrance Paolo Bonzini
2013-12-22 14:02 ` [Qemu-devel] [PULL v2 2/2] scsi-disk: add UNMAP limits to block limits VPD page Paolo Bonzini
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).