qemu/scsi-bus-fix-incorrect-call-for-blk_error_retry_rese.patch

82 lines
2.7 KiB
Diff
Raw Normal View History

QEMU update to version 8.2.0-2 - block: bugfix: Don't pause vm when NOSPACE EIO happened - block: enable cache mode of empty cdrom - block/mirror: fix file-system went to read-only after block-mirror - scsi-bus: fix incorrect call for blk_error_retry_reset_timeout() - scsi-bus: fix unmatched object_unref() - block: Add sanity check when setting retry parameters - block-backend: Stop retrying when draining - scsi-disk: Add support for retry on errors - scsi-bus: Refactor the code that retries requests - virtio_blk: Add support for retry on errors - block: Add error retry param setting - block-backend: Add timeout support for retry - block-backend: Enable retry action on errors - block-backend: Add device specific retry callback - block-backend: Introduce retry timer - qapi/block-core: Add retry option for error action - scsi: bugfix: fix division by zero - scsi: cdrom: Fix crash after remote cdrom detached - qemu-pr: fixed ioctl failed for multipath disk - scsi-disk: define props in scsi_block_disk to avoid memleaks - bugfix: fix possible memory leak - bugfix: fix some illegal memory access and memory leak - util/log: add CONFIG_DISABLE_QEMU_LOG macro - log: Add some logs on VM runtime path - bugfix: fix eventfds may double free when vm_id reused in ivshmem - hw/display/macfb: Fix missing ERRP_GUARD() in macfb_nubus_realize() - hw/cxl/cxl-host: Fix missing ERRP_GUARD() in cxl_fixed_memory_window_config() - qemu-img create: 'cache' paramter only use for reg file image - qemu-img: add qemu-img direct create - qemu-img block: set zero flags only when discard_zeros of the block supported - Revert "file-posix: Remove unused s->discard_zeroes" - pcie_sriov: Validate NumVFs (CVE-2024-26327) - hw/nvme: Use pcie_sriov_num_vfs() (CVE-2024-26328) - hw/acpi/cpu: Use CPUState typedef - target/i386/sev: Fix missing ERRP_GUARD() for error_prepend() - virtio-gpu: remove needless condition - hw/i2c/smbus_slave: Add object path on error prints - vfio/pci: Ascend710 change to bar2 quirk - vfio/pci: Ascend910 need 4Bytes quirk in bar0 - vfio/pci: Ascend710 need 4Bytes quirk in bar0 - vfio/pci: Ascend310 need 4Bytes quirk in bar4 - chardev/char-socket: Fix TLS io channels sending too much data to the backend - i386/cpuid: Move leaf 7 to correct group - i386/cpuid: Decrease cpuid_i when skipping CPUID leaf 1F - i386/cpu: Mask with XCR0/XSS mask for FEAT_XSAVE_XCR0_HI and FEAT_XSAVE_XSS_HI leafs - i386/cpu: Clear FEAT_XSAVE_XSS_LO/HI leafs when CPUID_EXT_XSAVE is not available - blkio: Respect memory-alignment for bounce buffer allocations - virtio-gpu: Correct virgl_renderer_resource_get_info() error check - hw/usb: Style cleanup - tests/qemu-iotests: resolved the problem that the 108 test cases in the container fail Signed-off-by: Jiabo Feng <fengjiabo1@huawei.com> (cherry picked from commit 404d45bf9147058a475a8031c454a6c8e0acc123)
2024-03-23 09:20:40 +08:00
From 60181b02c77f533105f904ab9e023bc22f65ad48 Mon Sep 17 00:00:00 2001
From: Yan Wang <wangyan122@huawei.com>
Date: Tue, 29 Mar 2022 12:05:56 +0800
Subject: [PATCH] scsi-bus: fix incorrect call for
blk_error_retry_reset_timeout()
Fix commit 52115ca0("scsi-disk: Add support for retry on errors").
Call Stack:
...
scsi_read_data()
scsi_do_read(r, 0)
scsi_disk_req_check_error()
blk_error_retry_reset_timeout()
blk->retry_start_time = 0;
It will cause IO hang when storage network disconnected. Before the
storage network recovered, the upper call stack will reset the
retry_start_time, and cause the next IO operation not returned immediately.
Signed-off-by: Yan Wang <wangyan122@huawei.com>
Signed-off-by: shaodenghui <shaodenghui@huawei.com>
---
hw/scsi/scsi-disk.c | 20 ++++++++++++++++----
1 file changed, 16 insertions(+), 4 deletions(-)
diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
index 97d8c5bb30..845a2a7d5d 100644
--- a/hw/scsi/scsi-disk.c
+++ b/hw/scsi/scsi-disk.c
@@ -258,10 +258,8 @@ static bool scsi_handle_rw_error(SCSIDiskReq *r, int ret, bool acct_failed)
}
}
-static bool scsi_disk_req_check_error(SCSIDiskReq *r, int ret, bool acct_failed)
+static bool scsi_disk_req_handle_error(SCSIDiskReq *r, int ret, bool acct_failed)
{
- SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev);
-
if (r->req.io_canceled) {
scsi_req_cancel_complete(&r->req);
return true;
@@ -271,6 +269,17 @@ static bool scsi_disk_req_check_error(SCSIDiskReq *r, int ret, bool acct_failed)
return scsi_handle_rw_error(r, ret, acct_failed);
}
+ return false;
+}
+
+static bool scsi_disk_req_check_error(SCSIDiskReq *r, int ret, bool acct_failed)
+{
+ SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev);
+
+ if (r->req.io_canceled || ret < 0) {
+ return scsi_disk_req_handle_error(r, ret, acct_failed);
+ }
+
blk_error_retry_reset_timeout(s->qdev.conf.blk);
return false;
}
@@ -423,7 +432,7 @@ static void scsi_do_read(SCSIDiskReq *r, int ret)
SCSIDiskClass *sdc = (SCSIDiskClass *) object_get_class(OBJECT(s));
assert (r->req.aiocb == NULL);
- if (scsi_disk_req_check_error(r, ret, false)) {
+ if (scsi_disk_req_handle_error(r, ret, false)) {
goto done;
}
@@ -464,6 +473,9 @@ static void scsi_do_read_cb(void *opaque, int ret)
block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct);
} else {
block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct);
+ if (!r->req.io_canceled) {
+ blk_error_retry_reset_timeout(s->qdev.conf.blk);
+ }
}
scsi_do_read(opaque, ret);
aio_context_release(blk_get_aio_context(s->qdev.conf.blk));
--
2.27.0