qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: kwolf@redhat.com
Subject: [Qemu-devel] [PATCH 1/8] scsi: do not call transfer_data after canceling a request
Date: Tue, 25 Oct 2011 12:53:33 +0200	[thread overview]
Message-ID: <1319540020-32484-2-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1319540020-32484-1-git-send-email-pbonzini@redhat.com>

Otherwise, if cancellation is "faked" by the AIO layer and goes
through qemu_aio_flush, the whole request is completed synchronously
during scsi_req_cancel.

Using the enqueued flag would work here, but not in the next patches,
so I'm introducing a new io_canceled flag.  That's because scsi_req_data
is a synchronous callback and the enqueued flag might be reset by the
time it returns.  scsi-disk cannot unref the request until after calling
scsi_req_data.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 hw/scsi-bus.c |   23 +++++++++++++++++++----
 hw/scsi.h     |    1 +
 trace-events  |    1 +
 3 files changed, 21 insertions(+), 4 deletions(-)

diff --git a/hw/scsi-bus.c b/hw/scsi-bus.c
index 72c0dd2..cd0c2cd 100644
--- a/hw/scsi-bus.c
+++ b/hw/scsi-bus.c
@@ -1107,8 +1107,12 @@ void scsi_req_continue(SCSIRequest *req)
    Once it completes, calling scsi_req_continue will restart I/O.  */
 void scsi_req_data(SCSIRequest *req, int len)
 {
-    trace_scsi_req_data(req->dev->id, req->lun, req->tag, len);
-    req->bus->info->transfer_data(req, len);
+    if (req->io_canceled) {
+        trace_scsi_req_data_canceled(req->dev->id, req->lun, req->tag, len);
+    } else {
+        trace_scsi_req_data(req->dev->id, req->lun, req->tag, len);
+        req->bus->info->transfer_data(req, len);
+    }
 }
 
 void scsi_req_print(SCSIRequest *req)
@@ -1173,11 +1177,15 @@ void scsi_req_complete(SCSIRequest *req, int status)
 
 void scsi_req_cancel(SCSIRequest *req)
 {
-    if (req->ops->cancel_io) {
-        req->ops->cancel_io(req);
+    if (!req->enqueued) {
+        return;
     }
     scsi_req_ref(req);
     scsi_req_dequeue(req);
+    req->io_canceled = true;
+    if (req->ops->cancel_io) {
+        req->ops->cancel_io(req);
+    }
     if (req->bus->info->cancel) {
         req->bus->info->cancel(req);
     }
@@ -1186,10 +1194,17 @@ void scsi_req_cancel(SCSIRequest *req)
 
 void scsi_req_abort(SCSIRequest *req, int status)
 {
+    if (!req->enqueued) {
+        return;
+    }
+    scsi_req_ref(req);
+    scsi_req_dequeue(req);
+    req->io_canceled = true;
     if (req->ops->cancel_io) {
         req->ops->cancel_io(req);
     }
     scsi_req_complete(req, status);
+    scsi_req_unref(req);
 }
 
 void scsi_device_purge_requests(SCSIDevice *sdev, SCSISense sense)
diff --git a/hw/scsi.h b/hw/scsi.h
index 8ea744a..fcc3455 100644
--- a/hw/scsi.h
+++ b/hw/scsi.h
@@ -51,6 +51,7 @@ struct SCSIRequest {
     uint8_t sense[SCSI_SENSE_BUF_SIZE];
     uint32_t sense_len;
     bool enqueued;
+    bool io_canceled;
     void *hba_private;
     QTAILQ_ENTRY(SCSIRequest) next;
 };
diff --git a/trace-events b/trace-events
index 7f9cec4..148084e 100644
--- a/trace-events
+++ b/trace-events
@@ -277,6 +277,7 @@ usb_host_claim_port(int bus, int hub, int port) "bus %d, hub addr %d, port %d"
 # hw/scsi-bus.c
 scsi_req_alloc(int target, int lun, int tag) "target %d lun %d tag %d"
 scsi_req_data(int target, int lun, int tag, int len) "target %d lun %d tag %d len %d"
+scsi_req_data_canceled(int target, int lun, int tag, int len) "target %d lun %d tag %d len %d"
 scsi_req_dequeue(int target, int lun, int tag) "target %d lun %d tag %d"
 scsi_req_continue(int target, int lun, int tag) "target %d lun %d tag %d"
 scsi_req_parsed(int target, int lun, int tag, int cmd, int mode, int xfer) "target %d lun %d tag %d command %d dir %d length %d"
-- 
1.7.6

  reply	other threads:[~2011-10-25 10:53 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-10-25 10:53 [Qemu-devel] [PATCH 0/5] My remaining block/SCSI patches for 1.0 Paolo Bonzini
2011-10-25 10:53 ` Paolo Bonzini [this message]
2011-10-25 10:53 ` [Qemu-devel] [PATCH 2/8] scsi-disk: bump SCSIRequest reference count until aio completion runs Paolo Bonzini
2011-10-25 10:53 ` [Qemu-devel] [PATCH 3/8] scsi-generic: " Paolo Bonzini
2011-10-25 10:53 ` [Qemu-devel] [PATCH 4/8] scsi: push request restart to SCSIDevice Paolo Bonzini
2011-10-25 10:53 ` [Qemu-devel] [PATCH 5/8] scsi-disk: add scsi-block for device passthrough Paolo Bonzini
2011-10-28 17:04   ` Kevin Wolf
2011-10-25 10:53 ` [Qemu-devel] [PATCH 6/8] block: add eject request callback Paolo Bonzini
2011-10-28 17:21   ` Kevin Wolf
2011-10-29  7:46     ` Paolo Bonzini
2011-11-07 13:21   ` Markus Armbruster
2011-11-07 13:36     ` Paolo Bonzini
2011-11-07 13:49       ` Kevin Wolf
2011-11-07 13:56         ` Paolo Bonzini
2011-11-07 14:12           ` Kevin Wolf
2011-11-07 15:23       ` Markus Armbruster
2011-11-07 16:14         ` Paolo Bonzini
2011-11-07 16:50   ` [Qemu-devel] [PATCH 6/8 v2] " Paolo Bonzini
2011-11-08 13:18     ` Kevin Wolf
2011-10-25 10:53 ` [Qemu-devel] [PATCH 7/8] atapi: implement eject requests Paolo Bonzini
2011-10-25 10:53 ` [Qemu-devel] [PATCH 8/8] scsi-disk: " Paolo Bonzini
2011-10-27 11:45 ` [Qemu-devel] ping Re: [PATCH 0/5] My remaining block/SCSI patches for 1.0 Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1319540020-32484-2-git-send-email-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).