From: Jeff Cody <jcody@redhat.com>
To: Supriya Kannery <supriyak@linux.vnet.ibm.com>
Cc: Kevin Wolf <kwolf@redhat.com>,
Shrinidhi Joshi <spjoshi31@gmail.com>,
Stefan Hajnoczi <stefanha@gmail.com>,
qemu-devel@nongnu.org, Luiz Capitulino <lcapitulino@redhat.com>,
Christoph Hellwig <hch@lst.de>
Subject: Re: [Qemu-devel] [v2 Patch 2/9]block: raw-posix image file reopen
Date: Thu, 09 Aug 2012 00:26:21 -0400 [thread overview]
Message-ID: <50233BED.7010009@redhat.com> (raw)
In-Reply-To: <20120730213437.21536.87232.sendpatchset@skannery.in.ibm.com>
On 07/30/2012 05:34 PM, Supriya Kannery wrote:
> raw-posix driver changes for bdrv_reopen_xx functions to
> safely reopen image files. Reopening of image files while
> changing hostcache dynamically is handled here.
>
> Signed-off-by: Supriya Kannery <supriyak@linux.vnet.ibm.com>
>
> ---
> Index: qemu/block/raw.c
> ===================================================================
> --- qemu.orig/block/raw.c
> +++ qemu/block/raw.c
> @@ -9,6 +9,22 @@ static int raw_open(BlockDriverState *bs
> return 0;
> }
>
> +static int raw_reopen_prepare(BlockDriverState *bs, BDRVReopenState **prs,
> + int flags)
> +{
> + return bdrv_reopen_prepare(bs->file, prs, flags);
> +}
> +
> +static void raw_reopen_commit(BlockDriverState *bs, BDRVReopenState *rs)
> +{
> + bdrv_reopen_commit(bs->file, rs);
> +}
> +
> +static void raw_reopen_abort(BlockDriverState *bs, BDRVReopenState *rs)
> +{
> + bdrv_reopen_abort(bs->file, rs);
> +}
> +
I think the above should be stubs that do nothing (or, alternatively,
are not present at all), if we have the reopen() do the bs->file
reopen() (which is needed for qcow2 - see my comments for patch 1/9)
> static int coroutine_fn raw_co_readv(BlockDriverState *bs, int64_t sector_num,
> int nb_sectors, QEMUIOVector *qiov)
> {
> @@ -113,6 +129,10 @@ static BlockDriver bdrv_raw = {
> .instance_size = 1,
>
> .bdrv_open = raw_open,
> + .bdrv_reopen_prepare
> + = raw_reopen_prepare,
> + .bdrv_reopen_commit = raw_reopen_commit,
> + .bdrv_reopen_abort = raw_reopen_abort,
> .bdrv_close = raw_close,
>
> .bdrv_co_readv = raw_co_readv,
> Index: qemu/block/raw-posix.c
> ===================================================================
> --- qemu.orig/block/raw-posix.c
> +++ qemu/block/raw-posix.c
> @@ -140,8 +140,15 @@ typedef struct BDRVRawState {
> #endif
> } BDRVRawState;
>
> +typedef struct BDRVRawReopenState {
> + BDRVReopenState reopen_state;
> + BDRVRawState *stash_s;
> +} BDRVRawReopenState;
> +
> static int fd_open(BlockDriverState *bs);
> static int64_t raw_getlength(BlockDriverState *bs);
> +static void raw_stash_state(BDRVRawState *stashed_state, BDRVRawState *s);
> +static void raw_revert_state(BDRVRawState *s, BDRVRawState *stashed_state);
>
> #if defined(__FreeBSD__) || defined(__FreeBSD_kernel__)
> static int cdrom_reopen(BlockDriverState *bs);
> @@ -283,6 +290,117 @@ static int raw_open(BlockDriverState *bs
> return raw_open_common(bs, filename, flags, 0);
> }
>
> +static int raw_reopen_prepare(BlockDriverState *bs, BDRVReopenState **prs,
> + int flags)
> +{
> + BDRVRawReopenState *raw_rs = g_malloc0(sizeof(BDRVRawReopenState));
> + BDRVRawState *s = bs->opaque;
> + int ret = 0;
> +
> + raw_rs->reopen_state.bs = bs;
> +
> + /* stash state before reopen */
> + raw_rs->stash_s = g_malloc0(sizeof(BDRVRawState));
> + raw_stash_state(raw_rs->stash_s, s);
> + s->fd = dup3(raw_rs->stash_s->fd, s->fd, O_CLOEXEC);
> +
> + *prs = &(raw_rs->reopen_state);
> +
> + /* Flags that can be set using fcntl */
> + int fcntl_flags = BDRV_O_NOCACHE;
> +
> + if ((bs->open_flags & ~fcntl_flags) == (flags & ~fcntl_flags)) {
> + if ((flags & BDRV_O_NOCACHE)) {
> + s->open_flags |= O_DIRECT;
> + } else {
> + s->open_flags &= ~O_DIRECT;
> + }
> + ret = fcntl_setfl(s->fd, s->open_flags);
> + } else {
> +
> + /* close and reopen using new flags */
> + bs->drv->bdrv_close(bs);
> + ret = bs->drv->bdrv_file_open(bs, bs->filename, flags);
This worries me, in that we close the BDS and then try
to reopen it. At this point, if the bdrv_file_open() fails, we have no
recovery mechanism.
Maybe we could try to do something like a bdrv_file_open() into a new BDS
before we close, followed by something akin to a bdrv_swap() /
bdrv_append() if the new bdrv_file_open() was successful, and then do
the close on the original. On failure, just abandon the new BDS.
> + }
> + return ret;
> +}
> +
> +static void raw_reopen_commit(BlockDriverState *bs, BDRVReopenState *rs)
> +{
> + BDRVRawReopenState *raw_rs;
> +
> + raw_rs = container_of(rs, BDRVRawReopenState, reopen_state);
> +
> + /* clean up stashed state */
> + close(raw_rs->stash_s->fd);
> + g_free(raw_rs->stash_s);
> + g_free(raw_rs);
> +}
> +
> +static void raw_reopen_abort(BlockDriverState *bs, BDRVReopenState *rs)
> +{
> + BDRVRawReopenState *raw_rs;
> + BDRVRawState *s = bs->opaque;
> +
> + raw_rs = container_of(rs, BDRVRawReopenState, reopen_state);
> +
> + /* revert to stashed state */
> + if (s->fd != -1) {
> + close(s->fd);
> + }
> + raw_revert_state(s, raw_rs->stash_s);
> + g_free(raw_rs->stash_s);
> + g_free(raw_rs);
> +}
> +
> +static void raw_stash_state(BDRVRawState *stashed_s, BDRVRawState *s)
> +{
> + stashed_s->fd = -1;
> + stashed_s->type = s->type;
> + stashed_s->open_flags = s->open_flags;
> +#if defined(__linux__)
> + /* linux floppy specific */
> + stashed_s->fd_open_time = s->fd_open_time;
> + stashed_s->fd_error_time = s->fd_error_time;
> + stashed_s->fd_got_error = s->fd_got_error;
> + stashed_s->fd_media_changed = s->fd_media_changed;
> +#endif
> +#ifdef CONFIG_LINUX_AIO
> + stashed_s->use_aio = s->use_aio;
> + stashed_s->aio_ctx = s->aio_ctx;
> +#endif
> + stashed_s->aligned_buf = s->aligned_buf;
> + stashed_s->aligned_buf_size = s->aligned_buf_size;
> +#ifdef CONFIG_XFS
> + stashed_s->is_xfs = s->is_xfs;
> +#endif
> +
> +}
> +
> +static void raw_revert_state(BDRVRawState *s, BDRVRawState *stashed_s)
> +{
> +
> + s->fd = stashed_s->fd;
> + s->type = stashed_s->type;
> + s->open_flags = stashed_s->open_flags;
> +#if defined(__linux__)
> + /* linux floppy specific */
> + s->fd_open_time = stashed_s->fd_open_time;
> + s->fd_error_time = stashed_s->fd_error_time;
> + s->fd_got_error = stashed_s->fd_got_error;
> + s->fd_media_changed = stashed_s->fd_media_changed;
> +#endif
> +#ifdef CONFIG_LINUX_AIO
> + s->use_aio = stashed_s->use_aio;
> + s->aio_ctx = stashed_s->aio_ctx;
> +#endif
> + s->aligned_buf = stashed_s->aligned_buf;
> + s->aligned_buf_size = stashed_s->aligned_buf_size;
> +#ifdef CONFIG_XFS
> + s->is_xfs = stashed_s->is_xfs;
> +#endif
> +}
> +
> /* XXX: use host sector size if necessary with:
> #ifdef DIOCGSECTORSIZE
> {
> @@ -735,6 +853,9 @@ static BlockDriver bdrv_file = {
> .instance_size = sizeof(BDRVRawState),
> .bdrv_probe = NULL, /* no probe for protocols */
> .bdrv_file_open = raw_open,
> + .bdrv_reopen_prepare = raw_reopen_prepare,
> + .bdrv_reopen_commit = raw_reopen_commit,
> + .bdrv_reopen_abort = raw_reopen_abort,
> .bdrv_close = raw_close,
> .bdrv_create = raw_create,
> .bdrv_co_discard = raw_co_discard,
>
>
next prev parent reply other threads:[~2012-08-09 4:26 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-07-30 21:34 [Qemu-devel] [v2 Patch 0/9]block: Image file reopen and dynamic host pagecache change Supriya Kannery
2012-07-30 21:34 ` [Qemu-devel] [v2 Patch 1/9]block: Framework for reopening image files safely Supriya Kannery
2012-08-01 15:51 ` Stefan Hajnoczi
2012-08-02 20:19 ` Luiz Capitulino
2012-08-14 8:54 ` Supriya Kannery
2012-08-08 21:13 ` Jeff Cody
2012-08-14 9:21 ` Supriya Kannery
2012-08-09 4:26 ` Jeff Cody
2012-08-09 9:20 ` Kevin Wolf
2012-08-09 13:02 ` Jeff Cody
2012-08-14 10:24 ` Supriya Kannery
2012-07-30 21:34 ` [Qemu-devel] [v2 Patch 2/9]block: raw-posix image file reopen Supriya Kannery
2012-07-31 17:17 ` Eric Blake
2012-08-01 6:18 ` Kevin Wolf
2012-08-03 22:32 ` Jeff Cody
2012-08-14 10:53 ` Supriya Kannery
2012-08-09 4:26 ` Jeff Cody [this message]
2012-08-10 13:45 ` Corey Bryant
2012-08-14 11:13 ` Supriya Kannery
2012-08-14 11:35 ` Kevin Wolf
2012-07-30 21:34 ` [Qemu-devel] [v2 Patch 3/9]block: raw-win32 " Supriya Kannery
2012-07-30 21:35 ` [Qemu-devel] [v2 Patch 4/9]block: vmdk " Supriya Kannery
2012-07-31 17:43 ` Eric Blake
2012-07-30 21:35 ` [Qemu-devel] [v2 Patch 5/9]block: qcow2 " Supriya Kannery
2012-08-09 4:26 ` Jeff Cody
2012-08-09 9:37 ` Kevin Wolf
2012-07-30 21:35 ` [Qemu-devel] [v2 Patch 7/9]block: qed " Supriya Kannery
2012-07-30 21:36 ` [Qemu-devel] [v2 Patch 9/9]block: Enhance "info block" to display host cache setting Supriya Kannery
2012-07-31 17:47 ` Eric Blake
2012-07-31 20:33 ` [Qemu-devel] [v2 Patch 0/9]block: Image file reopen and dynamic host pagecache change Jeff Cody
2012-08-01 17:11 ` Supriya Kannery
2012-08-01 18:36 ` [Qemu-devel] [v2 Patch 6/9]block: qcow image file reopen Supriya Kannery
2012-08-01 18:44 ` [Qemu-devel] [v2 Patch 8/9]block: Cmd "block_set_hostcache" for dynamic cache change Supriya Kannery
2012-08-01 19:21 ` Eric Blake
2012-08-02 20:36 ` Luiz Capitulino
2012-08-31 19:32 ` Jeff Cody
2012-08-09 4:26 ` [Qemu-devel] [v2 Patch 0/9]block: Image file reopen and dynamic host pagecache change Jeff Cody
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=50233BED.7010009@redhat.com \
--to=jcody@redhat.com \
--cc=hch@lst.de \
--cc=kwolf@redhat.com \
--cc=lcapitulino@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=spjoshi31@gmail.com \
--cc=stefanha@gmail.com \
--cc=supriyak@linux.vnet.ibm.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).