Re: [Qemu-devel] [PATCH v12 00/27] IOMMU: Enable interrupt remapping for Intel IOMMU

[Igor Mammedov <imammedo@redhat.com>]

On Thu, 14 Jul 2016 13:56:09 +0800
Peter Xu <peterx@redhat.com> wrote:

> This is v12 for Intel IR. Rebased to lastest master, with some tiny
> tweaks from v11. Please check changelog below.

Hi Peter,

While testing current master 9b8595bc with merged IR remapping
and default IDE disk as following:

qemu-system-x86_64 --enable-kvm -m 2G -snapshot -M q35 \
  -smp 1,sockets=7,cores=32,threads=1,maxcpus=240 \
  -device qemu64-x86_64-cpu,socket-id=1,core-id=30,thread-id=0 \
  -device intel-iommu,intremap=on -machine kernel_irqchip=split \
  rhel73_disk.img

I see errors, on QEMU stderr:

KVM: injection failed, MSI lost (Operation not permitted)
[...]

and in guest:

[   48.377370] ata1.00: exception Emask 0x0 SAct 0x7c0fffff SErr 0x0 action 0x6 frozen
[   48.378069] ata1.00: failed command: WRITE FPDMA QUEUED
[   48.378493] ata1.00: cmd 61/08:00:90:0a:c0/00:00:00:00:00/40 tag 0 ncq 4096 out
[   48.378493]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[...]
[   48.465957] ata1: hard resetting link
[   48.772096] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[   53.773067] ata1.00: qc timeout (cmd 0xec)
[   53.774054] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[   53.775126] ata1.00: revalidation failed (errno=-5)
[...]
[   94.394215] ata1.00: qc timeout (cmd 0xec)
[   94.395163] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[   94.396213] ata1.00: revalidation failed (errno=-5)
[   94.397184] ata1.00: disabled
[   94.398021] ata1.00: device reported invalid CHS sector 0
[...]
[   94.725462] sd 0:0:0:0: [sda] FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
[   94.726559] sd 0:0:0:0: [sda] CDB: Write(10) 2a 00 00 c0 09 00 00 00 08 00
[   94.727462] blk_update_request: I/O error, dev sda, sector 12585216
[   94.728320] Buffer I/O error on dev sda1, logical block 1572896, lost async page write
[...]
[   94.771975] Buffer I/O error on dev sda1, logical block 2097257, lost async page write
[   94.773458] EXT4-fs warning (device sda1): ext4_end_bio:316: I/O error -5 writing to inode 524365 (offset 0 size 0 starting block 2130264)
[   94.773540] EXT4-fs warning (device sda1): ext4_end_bio:316: I/O error -5 writing to inode 525096 (offset 0 size 8192 starting block 2130596)
[   94.773541] Buffer I/O error on device sda1, logical block 2130338
[...]
[   94.783649] EXT4-fs error (device sda1): ext4_wait_block_bitmap:494: comm kworker/u480:5: Cannot read block bitmap - block_group = 11, block_bitmap = 1038
[   94.786408] EXT4-fs (sda1): Delayed block allocation failed for inode 524366 at logical offset 1280 with max blocks 1 with error 5
[   94.788973] EXT4-fs (sda1): This should not happen!! Data will be lost