From: Hanna Reitz <hreitz@redhat.com>
To: Emanuele Giuseppe Esposito <eesposit@redhat.com>, qemu-block@nongnu.org
Cc: "Kevin Wolf" <kwolf@redhat.com>, "Fam Zheng" <fam@euphon.net>,
"Vladimir Sementsov-Ogievskiy" <vsementsov@virtuozzo.com>,
"Daniel P. Berrangé" <berrange@redhat.com>,
"Eduardo Habkost" <ehabkost@redhat.com>,
"Juan Quintela" <quintela@redhat.com>,
qemu-devel@nongnu.org, "John Snow" <jsnow@redhat.com>,
"Richard Henderson" <richard.henderson@linaro.org>,
"Markus Armbruster" <armbru@redhat.com>,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
"Stefan Hajnoczi" <stefanha@redhat.com>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Eric Blake" <eblake@redhat.com>
Subject: Re: [PATCH v5 04/31] include/sysemu/block-backend: split header into I/O and global state (GS) API
Date: Fri, 10 Dec 2021 15:21:31 +0100 [thread overview]
Message-ID: <40961f22-ca91-29fa-0488-f5549a5d314f@redhat.com> (raw)
In-Reply-To: <20211124064418.3120601-5-eesposit@redhat.com>
On 24.11.21 07:43, Emanuele Giuseppe Esposito wrote:
> Similarly to the previous patches, split block-backend.h
> in block-backend-io.h and block-backend-global-state.h
>
> In addition, remove "block/block.h" include as it seems
> it is not necessary anymore, together with "qemu/iov.h"
>
> block-backend-common.h contains the structures shared between
> the two headers, and the functions that can't be categorized as
> I/O or global state.
>
> Assertions are added in the next patch.
>
> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
> ---
> include/sysemu/block-backend-common.h | 84 ++++++
> include/sysemu/block-backend-global-state.h | 121 +++++++++
> include/sysemu/block-backend-io.h | 139 ++++++++++
> include/sysemu/block-backend.h | 269 +-------------------
> block/block-backend.c | 9 +-
> 5 files changed, 353 insertions(+), 269 deletions(-)
> create mode 100644 include/sysemu/block-backend-common.h
> create mode 100644 include/sysemu/block-backend-global-state.h
> create mode 100644 include/sysemu/block-backend-io.h
[...]
> diff --git a/include/sysemu/block-backend-global-state.h b/include/sysemu/block-backend-global-state.h
> new file mode 100644
> index 0000000000..77009bf7a2
> --- /dev/null
> +++ b/include/sysemu/block-backend-global-state.h
[...]
> +int blk_ioctl(BlockBackend *blk, unsigned long int req, void *buf);
I don’t quite follow what makes blk_ioctl() a GS function. (It’s
possible that I forgot something about this from v4, though, and can’t
find it anymore...)
I suppose it can be said that in the context of the block layer, ioctl
functions are generally used outside of I/O threads as something that
kind of affects the global data state, but... bdrv_co_ioctl() is in
block-io.h, and internally we definitely do use ioctl in I/O threads
(various uses in file-posix.c, e.g. to zero out areas). I also don’t
understand why blk_ioctl() is GS, and blk_aio_ioctl() is I/O.
And if you have a virtio-scsi device in an iothread, and then create a
scsi-generic device on it, will bdrv_co_ioctl() not run in the I/O
thread still...? Hm, testing (on patch 6), it appears that not – I
think that’s because it mostly uses blk_aio_ioctl(), apart from
*_realize() and scsi_SG_IO_FROM_DEV() (whatever that is; and this makes
me a bit cautious, too). Still, I’m not quite sure how it’s more global
state than e.g. writing zeroes or, well, blk_aio_ioctl().
However, testing brought some other problems to light:
$ ls /dev/sg?
/dev/sg0
$ sudo modprobe scsi_debug max_luns=1 num_tgts=1 add_host=1
$ ls /dev/sg?
/dev/sg0 /dev/sg1
$ ./qemu-system-x86_64 \
-object iothread,id=foo \
-device virtio-scsi,iothread=foo \
-blockdev host_device,filename=/dev/sg1,node-name=bar \
-device scsi-generic,drive=bar
qemu-system-x86_64: ../block/block-backend.c:2038:
blk_set_guest_block_size: Assertion `qemu_in_main_thread()' failed.
[1] 121353 IOT instruction (core dumped) ./qemu-system-x86_64
-object iothread,id=foo -device virtio-scsi,iothread=foo
And:
$ ./qemu-system-x86_64 \
-object iothread,id=foo \
-device virtio-scsi,iothread=foo \
-blockdev null-co,read-zeroes=true,node-name=bar \
-device scsi-hd,drive=bar \
-cdrom ~/tmp/arch.iso \
-enable-kvm \
-m 1g
qemu-system-x86_64: ../block/block-backend.c:1918:
blk_enable_write_cache: Assertion `qemu_in_main_thread()' failed.
[1] 127203 IOT instruction (core dumped) ./qemu-system-x86_64
-object iothread,id=foo -device virtio-scsi,iothread=foo
If I boot from a CD under virtio-scsi (in an iothread), I also get an
error for blk_lock_medium(), and if I eject such a CD, for blk_eject()
and blk_get_attached_dev_id() (called from blk_eject()).
It appears like scsi_disk_emulate_command() is always run in the BB’s
AioContext, and so everything it (indirectly) calls cannot assume to be
in the main thread. As for blk_set_guest_block_size(), that one’s
called from scsi-generic’s scsi_read_complete().
Hanna
next prev parent reply other threads:[~2021-12-10 14:22 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-24 6:43 [PATCH v5 00/31] block layer: split block APIs in global state and I/O Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 01/31] main-loop.h: introduce qemu_in_main_thread() Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 02/31] include/block/block: split header into I/O and global state API Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 03/31] assertions for block " Emanuele Giuseppe Esposito
2021-12-16 15:17 ` Hanna Reitz
2021-11-24 6:43 ` [PATCH v5 04/31] include/sysemu/block-backend: split header into I/O and global state (GS) API Emanuele Giuseppe Esposito
2021-12-10 14:21 ` Hanna Reitz [this message]
2021-11-24 6:43 ` [PATCH v5 05/31] block-backend: special comments for blk_set/get_perm due to fuse Emanuele Giuseppe Esposito
2021-12-10 14:38 ` Hanna Reitz
2021-12-15 8:57 ` Emanuele Giuseppe Esposito
2021-12-15 10:13 ` Emanuele Giuseppe Esposito
2021-12-16 14:00 ` Hanna Reitz
2021-11-24 6:43 ` [PATCH v5 06/31] block/block-backend.c: assertions for block-backend Emanuele Giuseppe Esposito
2021-12-10 15:37 ` Hanna Reitz
2021-11-24 6:43 ` [PATCH v5 07/31] include/block/block_int: split header into I/O and global state API Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 08/31] assertions for block_int " Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 09/31] block: introduce assert_bdrv_graph_writable Emanuele Giuseppe Esposito
2021-12-10 17:43 ` Hanna Reitz
2021-12-14 19:48 ` Emanuele Giuseppe Esposito
2021-12-16 14:01 ` Hanna Reitz
2021-11-24 6:43 ` [PATCH v5 10/31] block.c: modify .attach and .detach callbacks of child_of_bds Emanuele Giuseppe Esposito
2021-12-16 14:57 ` Hanna Reitz
2021-12-16 16:05 ` Emanuele Giuseppe Esposito
2021-12-16 16:12 ` Hanna Reitz
2021-11-24 6:43 ` [PATCH v5 11/31] include/block/blockjob_int.h: split header into I/O and GS API Emanuele Giuseppe Esposito
2021-11-24 6:43 ` [PATCH v5 12/31] assertions for blockjob_int.h Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 13/31] block.c: add assertions to static functions Emanuele Giuseppe Esposito
2021-12-16 16:08 ` Hanna Reitz
2021-12-16 16:39 ` Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 14/31] include/block/blockjob.h: global state API Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 15/31] assertions for blockjob.h " Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 16/31] include/sysemu/blockdev.h: " Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 17/31] assertions for blockdev.h " Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 18/31] include/block/snapshot: global state API + assertions Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 19/31] block/copy-before-write.h: " Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 20/31] block/coroutines: I/O API Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 21/31] block_int-common.h: split function pointers in BlockDriver Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 22/31] block_int-common.h: assertion in the callers of BlockDriver function pointers Emanuele Giuseppe Esposito
2021-12-16 18:43 ` Hanna Reitz
2021-12-17 15:53 ` Emanuele Giuseppe Esposito
2021-12-17 16:00 ` Hanna Reitz
2021-11-24 6:44 ` [PATCH v5 23/31] block_int-common.h: split function pointers in BdrvChildClass Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 24/31] block_int-common.h: assertions in the callers of BdrvChildClass function pointers Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 25/31] block-backend-common.h: split function pointers in BlockDevOps Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 26/31] job.h: split function pointers in JobDriver Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 27/31] job.h: assertions in the callers of JobDriver funcion pointers Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 28/31] block.c: assert BQL lock held in bdrv_co_invalidate_cache Emanuele Giuseppe Esposito
2021-12-17 11:04 ` Hanna Reitz
2021-12-17 16:38 ` Emanuele Giuseppe Esposito
2021-12-20 12:20 ` Emanuele Giuseppe Esposito
2021-12-23 17:11 ` Hanna Reitz
2022-01-19 15:57 ` Hanna Reitz
2022-01-19 17:44 ` Hanna Reitz
2022-01-19 18:34 ` Kevin Wolf
2022-01-20 13:22 ` Paolo Bonzini
2022-01-20 13:48 ` Kevin Wolf
2022-01-21 10:22 ` Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 29/31] jobs: introduce pre_run function in JobDriver Emanuele Giuseppe Esposito
2021-11-24 6:44 ` [PATCH v5 30/31] crypto: delegate permission functions to JobDriver .pre_run Emanuele Giuseppe Esposito
2021-12-17 12:29 ` Hanna Reitz
2021-12-17 14:32 ` Hanna Reitz
2021-12-20 15:47 ` Emanuele Giuseppe Esposito
2021-12-23 17:15 ` Hanna Reitz
2021-11-24 6:44 ` [PATCH v5 31/31] block.c: assertions to the block layer permissions API Emanuele Giuseppe Esposito
2021-11-29 13:32 ` [PATCH v5 00/31] block layer: split block APIs in global state and I/O Stefan Hajnoczi
2021-11-30 8:39 ` Emanuele Giuseppe Esposito
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=40961f22-ca91-29fa-0488-f5549a5d314f@redhat.com \
--to=hreitz@redhat.com \
--cc=armbru@redhat.com \
--cc=berrange@redhat.com \
--cc=dgilbert@redhat.com \
--cc=eblake@redhat.com \
--cc=eesposit@redhat.com \
--cc=ehabkost@redhat.com \
--cc=fam@euphon.net \
--cc=jsnow@redhat.com \
--cc=kwolf@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
--cc=richard.henderson@linaro.org \
--cc=stefanha@redhat.com \
--cc=vsementsov@virtuozzo.com \
/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).