* [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status()
@ 2014-09-22 15:36 Max Reitz
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
` (2 more replies)
0 siblings, 3 replies; 9+ messages in thread
From: Max Reitz @ 2014-09-22 15:36 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Stefan Hajnoczi, Max Reitz
raw_co_get_block_status() should return 0 and set *pnum to 0 after the
EOF; currently it does this merely by accident, so implement it
directly. Also, nb_sectors should be clamped against the image end.
While doing that, centralize the generation of
raw_co_get_block_status()'s return value along the way.
v2:
- Patch 1: Clamp nb_sectors against image end
- Patch 2: Fix alignment issue
Max Reitz (2):
raw-posix: Fix raw_co_get_block_status() after EOF
raw-posix: raw_co_get_block_status() return value
block/raw-posix.c | 36 ++++++++++++++++++++++--------------
1 file changed, 22 insertions(+), 14 deletions(-)
--
2.1.0
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF
2014-09-22 15:36 [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
@ 2014-09-22 15:36 ` Max Reitz
2014-10-09 4:13 ` Eric Blake
2014-10-16 8:30 ` Kevin Wolf
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value Max Reitz
2014-10-08 19:43 ` [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2 siblings, 2 replies; 9+ messages in thread
From: Max Reitz @ 2014-09-22 15:36 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Stefan Hajnoczi, Max Reitz
As its comment states, raw_co_get_block_status() should unconditionally
return 0 and set *pnum to 0 for after EOF.
An assertion after lseek(..., SEEK_HOLE) tried to catch this case by
asserting that errno != -ENXIO (which would indicate a position after
the EOF); but it should be errno != ENXIO instead. Fix this, too.
Additionally, nb_sectors should be clamped against the image end. This
was probably not an issue if FIEMAP or SEEK_HOLE/SEEK_DATA worked, but
the fallback did not take this case into account.
Reported-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/raw-posix.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/block/raw-posix.c b/block/raw-posix.c
index a253697..dd57992 100644
--- a/block/raw-posix.c
+++ b/block/raw-posix.c
@@ -1509,9 +1509,9 @@ static int64_t try_seek_hole(BlockDriverState *bs, off_t start, off_t *data,
*hole = lseek(s->fd, start, SEEK_HOLE);
if (*hole == -1) {
- /* -ENXIO indicates that sector_num was past the end of the file.
+ /* ENXIO indicates that sector_num was past the end of the file.
* There is a virtual hole there. */
- assert(errno != -ENXIO);
+ assert(errno != ENXIO);
return -errno;
}
@@ -1552,6 +1552,7 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
int nb_sectors, int *pnum)
{
off_t start, data = 0, hole = 0;
+ int64_t total_size;
int64_t ret;
ret = fd_open(bs);
@@ -1560,6 +1561,13 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
}
start = sector_num * BDRV_SECTOR_SIZE;
+ total_size = bdrv_getlength(bs);
+ if (start >= total_size) {
+ *pnum = 0;
+ return 0;
+ } else if (start + nb_sectors * BDRV_SECTOR_SIZE > total_size) {
+ nb_sectors = (total_size - start) / BDRV_SECTOR_SIZE;
+ }
ret = try_fiemap(bs, start, &data, &hole, nb_sectors, pnum);
if (ret < 0) {
--
2.1.0
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value
2014-09-22 15:36 [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
@ 2014-09-22 15:36 ` Max Reitz
2014-10-09 4:16 ` Eric Blake
2014-10-08 19:43 ` [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2 siblings, 1 reply; 9+ messages in thread
From: Max Reitz @ 2014-09-22 15:36 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Stefan Hajnoczi, Max Reitz
Instead of generating the full return value thrice in try_fiemap(),
try_seek_hole() and as a fall-back in raw_co_get_block_status() itself,
generate the value only in raw_co_get_block_status().
Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/raw-posix.c | 24 ++++++++++++------------
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/block/raw-posix.c b/block/raw-posix.c
index dd57992..be030b2 100644
--- a/block/raw-posix.c
+++ b/block/raw-posix.c
@@ -1455,12 +1455,12 @@ out:
return result;
}
-static int64_t try_fiemap(BlockDriverState *bs, off_t start, off_t *data,
- off_t *hole, int nb_sectors, int *pnum)
+static int try_fiemap(BlockDriverState *bs, off_t start, off_t *data,
+ off_t *hole, int nb_sectors, int *pnum)
{
#ifdef CONFIG_FIEMAP
BDRVRawState *s = bs->opaque;
- int64_t ret = BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | start;
+ int ret = 0;
struct {
struct fiemap fm;
struct fiemap_extent fe;
@@ -1501,8 +1501,8 @@ static int64_t try_fiemap(BlockDriverState *bs, off_t start, off_t *data,
#endif
}
-static int64_t try_seek_hole(BlockDriverState *bs, off_t start, off_t *data,
- off_t *hole, int *pnum)
+static int try_seek_hole(BlockDriverState *bs, off_t start, off_t *data,
+ off_t *hole, int *pnum)
{
#if defined SEEK_HOLE && defined SEEK_DATA
BDRVRawState *s = bs->opaque;
@@ -1526,7 +1526,7 @@ static int64_t try_seek_hole(BlockDriverState *bs, off_t start, off_t *data,
}
}
- return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | start;
+ return 0;
#else
return -ENOTSUP;
#endif
@@ -1553,7 +1553,7 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
{
off_t start, data = 0, hole = 0;
int64_t total_size;
- int64_t ret;
+ int ret;
ret = fd_open(bs);
if (ret < 0) {
@@ -1576,21 +1576,21 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
/* Assume everything is allocated. */
data = 0;
hole = start + nb_sectors * BDRV_SECTOR_SIZE;
- ret = BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | start;
+ ret = 0;
}
}
+ assert(ret >= 0);
+
if (data <= start) {
/* On a data extent, compute sectors to the end of the extent. */
*pnum = MIN(nb_sectors, (hole - start) / BDRV_SECTOR_SIZE);
+ return ret | BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID | start;
} else {
/* On a hole, compute sectors to the beginning of the next extent. */
*pnum = MIN(nb_sectors, (data - start) / BDRV_SECTOR_SIZE);
- ret &= ~BDRV_BLOCK_DATA;
- ret |= BDRV_BLOCK_ZERO;
+ return ret | BDRV_BLOCK_ZERO | BDRV_BLOCK_OFFSET_VALID | start;
}
-
- return ret;
}
static coroutine_fn BlockDriverAIOCB *raw_aio_discard(BlockDriverState *bs,
--
2.1.0
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status()
2014-09-22 15:36 [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value Max Reitz
@ 2014-10-08 19:43 ` Max Reitz
2014-10-09 18:58 ` Benoît Canet
2 siblings, 1 reply; 9+ messages in thread
From: Max Reitz @ 2014-10-08 19:43 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Benoît Canet, Stefan Hajnoczi
On 22.09.2014 17:36, Max Reitz wrote:
> raw_co_get_block_status() should return 0 and set *pnum to 0 after the
> EOF; currently it does this merely by accident, so implement it
> directly. Also, nb_sectors should be clamped against the image end.
>
> While doing that, centralize the generation of
> raw_co_get_block_status()'s return value along the way.
>
>
> v2:
> - Patch 1: Clamp nb_sectors against image end
> - Patch 2: Fix alignment issue
>
>
> Max Reitz (2):
> raw-posix: Fix raw_co_get_block_status() after EOF
> raw-posix: raw_co_get_block_status() return value
>
> block/raw-posix.c | 36 ++++++++++++++++++++++--------------
> 1 file changed, 22 insertions(+), 14 deletions(-)
Ping. (This should be rather simple to review)
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
@ 2014-10-09 4:13 ` Eric Blake
2014-10-16 8:30 ` Kevin Wolf
1 sibling, 0 replies; 9+ messages in thread
From: Eric Blake @ 2014-10-09 4:13 UTC (permalink / raw)
To: Max Reitz, qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Stefan Hajnoczi
[-- Attachment #1: Type: text/plain, Size: 946 bytes --]
On 09/22/2014 09:36 AM, Max Reitz wrote:
> As its comment states, raw_co_get_block_status() should unconditionally
> return 0 and set *pnum to 0 for after EOF.
>
> An assertion after lseek(..., SEEK_HOLE) tried to catch this case by
> asserting that errno != -ENXIO (which would indicate a position after
> the EOF); but it should be errno != ENXIO instead. Fix this, too.
>
> Additionally, nb_sectors should be clamped against the image end. This
> was probably not an issue if FIEMAP or SEEK_HOLE/SEEK_DATA worked, but
> the fallback did not take this case into account.
>
> Reported-by: Kevin Wolf <kwolf@redhat.com>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/raw-posix.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 539 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value Max Reitz
@ 2014-10-09 4:16 ` Eric Blake
0 siblings, 0 replies; 9+ messages in thread
From: Eric Blake @ 2014-10-09 4:16 UTC (permalink / raw)
To: Max Reitz, qemu-devel; +Cc: Kevin Wolf, Paolo Bonzini, Stefan Hajnoczi
[-- Attachment #1: Type: text/plain, Size: 619 bytes --]
On 09/22/2014 09:36 AM, Max Reitz wrote:
> Instead of generating the full return value thrice in try_fiemap(),
> try_seek_hole() and as a fall-back in raw_co_get_block_status() itself,
> generate the value only in raw_co_get_block_status().
>
> Suggested-by: Kevin Wolf <kwolf@redhat.com>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/raw-posix.c | 24 ++++++++++++------------
> 1 file changed, 12 insertions(+), 12 deletions(-)
>
Reviewed-by: Eric Blake <eblake@redhat.com>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 539 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status()
2014-10-08 19:43 ` [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
@ 2014-10-09 18:58 ` Benoît Canet
2014-10-11 8:55 ` Max Reitz
0 siblings, 1 reply; 9+ messages in thread
From: Benoît Canet @ 2014-10-09 18:58 UTC (permalink / raw)
To: Max Reitz
Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi, Paolo Bonzini,
Benoît Canet
On Wed, Oct 08, 2014 at 09:43:19PM +0200, Max Reitz wrote:
> On 22.09.2014 17:36, Max Reitz wrote:
> >raw_co_get_block_status() should return 0 and set *pnum to 0 after the
> >EOF; currently it does this merely by accident, so implement it
> >directly. Also, nb_sectors should be clamped against the image end.
> >
> >While doing that, centralize the generation of
> >raw_co_get_block_status()'s return value along the way.
> >
> >
> >v2:
> >- Patch 1: Clamp nb_sectors against image end
> >- Patch 2: Fix alignment issue
> >
> >
> >Max Reitz (2):
> > raw-posix: Fix raw_co_get_block_status() after EOF
> > raw-posix: raw_co_get_block_status() return value
> >
> > block/raw-posix.c | 36 ++++++++++++++++++++++--------------
> > 1 file changed, 22 insertions(+), 14 deletions(-)
>
> Ping. (This should be rather simple to review)
Hi Max,
I will review these tomorow.
Best regards
Benoît
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status()
2014-10-09 18:58 ` Benoît Canet
@ 2014-10-11 8:55 ` Max Reitz
0 siblings, 0 replies; 9+ messages in thread
From: Max Reitz @ 2014-10-11 8:55 UTC (permalink / raw)
To: Benoît Canet; +Cc: Kevin Wolf, Paolo Bonzini, qemu-devel, Stefan Hajnoczi
Am 09.10.2014 um 20:58 schrieb Benoît Canet:
> On Wed, Oct 08, 2014 at 09:43:19PM +0200, Max Reitz wrote:
>> On 22.09.2014 17:36, Max Reitz wrote:
>>> raw_co_get_block_status() should return 0 and set *pnum to 0 after the
>>> EOF; currently it does this merely by accident, so implement it
>>> directly. Also, nb_sectors should be clamped against the image end.
>>>
>>> While doing that, centralize the generation of
>>> raw_co_get_block_status()'s return value along the way.
>>>
>>>
>>> v2:
>>> - Patch 1: Clamp nb_sectors against image end
>>> - Patch 2: Fix alignment issue
>>>
>>>
>>> Max Reitz (2):
>>> raw-posix: Fix raw_co_get_block_status() after EOF
>>> raw-posix: raw_co_get_block_status() return value
>>>
>>> block/raw-posix.c | 36 ++++++++++++++++++++++--------------
>>> 1 file changed, 22 insertions(+), 14 deletions(-)
>> Ping. (This should be rather simple to review)
> Hi Max,
>
> I will review these tomorow.
Thanks a lot for all of your reviews!
Max
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
2014-10-09 4:13 ` Eric Blake
@ 2014-10-16 8:30 ` Kevin Wolf
1 sibling, 0 replies; 9+ messages in thread
From: Kevin Wolf @ 2014-10-16 8:30 UTC (permalink / raw)
To: Max Reitz; +Cc: Paolo Bonzini, qemu-devel, Stefan Hajnoczi
Am 22.09.2014 um 17:36 hat Max Reitz geschrieben:
> As its comment states, raw_co_get_block_status() should unconditionally
> return 0 and set *pnum to 0 for after EOF.
>
> An assertion after lseek(..., SEEK_HOLE) tried to catch this case by
> asserting that errno != -ENXIO (which would indicate a position after
> the EOF); but it should be errno != ENXIO instead. Fix this, too.
>
> Additionally, nb_sectors should be clamped against the image end. This
> was probably not an issue if FIEMAP or SEEK_HOLE/SEEK_DATA worked, but
> the fallback did not take this case into account.
>
> Reported-by: Kevin Wolf <kwolf@redhat.com>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/raw-posix.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/block/raw-posix.c b/block/raw-posix.c
> index a253697..dd57992 100644
> --- a/block/raw-posix.c
> +++ b/block/raw-posix.c
> @@ -1509,9 +1509,9 @@ static int64_t try_seek_hole(BlockDriverState *bs, off_t start, off_t *data,
>
> *hole = lseek(s->fd, start, SEEK_HOLE);
> if (*hole == -1) {
> - /* -ENXIO indicates that sector_num was past the end of the file.
> + /* ENXIO indicates that sector_num was past the end of the file.
> * There is a virtual hole there. */
> - assert(errno != -ENXIO);
> + assert(errno != ENXIO);
This assertion can be triggered if another process truncates the file in
the background after it has been opened (bdrv_getlength() usually uses
the cached value, so this race condition isn't even hard to reproduce).
Kevin
> return -errno;
> }
> @@ -1552,6 +1552,7 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
> int nb_sectors, int *pnum)
> {
> off_t start, data = 0, hole = 0;
> + int64_t total_size;
> int64_t ret;
>
> ret = fd_open(bs);
> @@ -1560,6 +1561,13 @@ static int64_t coroutine_fn raw_co_get_block_status(BlockDriverState *bs,
> }
>
> start = sector_num * BDRV_SECTOR_SIZE;
> + total_size = bdrv_getlength(bs);
bdrv_getlength() can fail.
> + if (start >= total_size) {
> + *pnum = 0;
> + return 0;
> + } else if (start + nb_sectors * BDRV_SECTOR_SIZE > total_size) {
> + nb_sectors = (total_size - start) / BDRV_SECTOR_SIZE;
> + }
Kevin
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2014-10-16 8:30 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-09-22 15:36 [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 1/2] raw-posix: Fix raw_co_get_block_status() after EOF Max Reitz
2014-10-09 4:13 ` Eric Blake
2014-10-16 8:30 ` Kevin Wolf
2014-09-22 15:36 ` [Qemu-devel] [PATCH v2 2/2] raw-posix: raw_co_get_block_status() return value Max Reitz
2014-10-09 4:16 ` Eric Blake
2014-10-08 19:43 ` [Qemu-devel] [PATCH v2 0/2] raw-posix: Fix raw_co_get_block_status() Max Reitz
2014-10-09 18:58 ` Benoît Canet
2014-10-11 8:55 ` Max Reitz
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.