If the pagemap we read at some point covers two (or more) pagemaps
in parent images, we cannot just call the parent page read for all
those pages. Instead, we must call parent page read with shorter
requests so that it handles its own pagemaps properly.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When an image of a certian type is not found, CRIU sometimes
fails, sometimes ignores this fact. I propose to ignore this
fact always and treat absent images and those containing no
objects inside (i.e. -- empty). If the latter code flow will
_need_ objects, then criu will fail later.
Why object will be explicitly required? For example, due to
restoring code reading the image with pb_read_one, w/o the
_eof suffix thus required the object to be in the image.
Another example is objects dependencies. E.g. fdinfo objects
require various files objects. So missing image files will
result in non-resolved searches later.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When page-read fails to open the pagemap image it reports error.
One place (stacked page-reads) need to handle the absent images
case gracefully, so fix the return codes to make this check
work.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We want to have buffered images to speed up dump and,
slightly, restore. Right now we use plan file descriptors
to write and read images to/from. Making them buffered
cannot be gracefully done on plain fds, so introduce
a new class.
This will also help if (when?) we will want to do more
complex changes with images, e.g. store them all in one
file or send them directly to the network.
For now the cr_img just contains one int _fd variable.
This patch chages the prototype of open_image() to
return struct cr_img *, pb_(read|write)* to accept one
and fixes the compilation of the rest of the code :)
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
For example restore_shmem_content allocates the page_read structure on
stack.
Cc: Pavel Emelyanov <xemul@parallels.com>
Reported-by: Jenkins Criuovich
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
when decide that data is no longer needed, there are two cases:
-if data neighbours previous block of "no needed" data, extend bunch
block(it holds begining and size of concequent "no needed" data) by
length of curent block and go next.
-if data not neighbours bunch block(or bunch block size will be bigger
than MAX_BUNCH_SIZE), than we punch bunch block and set bunch block
to curent block.
in the end make cleanup to punch last bunch block.
changes in v1:
punch_hole takes whole page_read
make restriction more precise
Signed-off-by: Tikhomirov Pavel <snorcht@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
if option --auto-dedup is set on restore, then as soon as page is
restored it will be punched from the image.
open image in O_RDWR mode
Signed-off-by: Tikhomirov Pavel <snorcht@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
because eof in restore means that we don't have "must have" data, so
it mean error exit, but in dedup, it's ok not to have data we seek, so
we continue returning normally
also it's used that if off < pr-cvaddr then no eof happened and if eof
happened then off >= pr-cvaddr(then return of seek_pagemap_page is 0)
Signed-off-by: Tikhomirov Pavel <snorcht@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Modern tar and rsync can replace zero chunks with holes. So it's valid situation
and shouldn't abort the restore.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use the format specifier %zu instead of %lu to print a size_t integer.
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
1. add it instead of BUG_ON
we should print error if warn is assigned to true
2. correct use of read_pagemap_page_from_parent, provide warn = true
Signed-off-by: Tikhomirov Pavel <snorcht@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
These processes don't have image files in a parent snapshot and crtools
should not fail in this case.
https://bugzilla.openvz.org/show_bug.cgi?id=2636
v2: return NULL from mem_snap_init, if a parent image is absent.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Otherwise it contains trash and BUG at page-read.c:98 may be triggered.
https://bugzilla.openvz.org/show_bug.cgi?id=2633
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use the format specifier PRIx64 instead of %lx to print an off_t integer.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This one is a little bit more tricky that dump. On restore we
should open the chain of pagemap-s (by "parent" links). Then for
every pagemap with in_parent set we should go to parent and ask
for the page. Parent, in turn, should properly determine where
the respective page is in his pagemap and position the page.img's
position respectively.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Now we have 2 forms of storing pages -- legacy pages.img and
new pagemap + pages image. We'll have one more (ovz) and the
pagemap + pages will be stacked (snapshot restore). Thus it's
handy to have this as an page-reader object.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
