search for: nr_region

...variable vs global typedef). Also I got rid of the get_regions accessor method, replacing it everywhere with direct use of regions->ptr[]. Although the accessor method did bounds checking, my reasoning is this change is correct because in all cases we were iterating over the regions from 0 .. nr_regions-1. However an improvement would be to have a proper iterator, although that change is not so straightforward because of lack of closures in C. --- common/regions/regions.h | 37 +++++++++------------- common/regions/regions.c | 46 ++++++++++++---------------- plugins/par...

...> > Also I got rid of the get_regions accessor method, replacing it > everywhere with direct use of regions->ptr[]. Although the accessor > method did bounds checking, my reasoning is this change is correct > because in all cases we were iterating over the regions from > 0 .. nr_regions-1. However an improvement would be to have a proper > iterator, although that change is not so straightforward because of > lack of closures in C. > --- > @@ -70,38 +72,35 @@ struct region { > const char *description; > }; > > -/* Array of regions. */ > -struc...

...len, end; /* byte offsets; end = start + len - 1 */ + enum region_type type; + union { + size_t i; /* region_file: i'th file */ + const unsigned char *data; /* region_data: data (partition table) */ + } u; +}; + +static struct region *regions = NULL; +static size_t nr_regions = 0; + +/* Primary and secondary partition tables (secondary is only used for GPT). */ +static unsigned char *primary = NULL, *secondary = NULL; + +static void +partitioning_load (void) +{ + srandom (time (NULL)); +} + +static void +partitioning_unload (void) +{ + size_t i; + + for (i = 0; i &l...

nbdkit partitioning boot.img swap.img root.img ... creates a virtual disk by adding a partition table. In ancient times Xen used to do this. Rich.

The partitioning plugin patch is the same (except for rebasing). However I have changed the first two patches based on feedback received. In particular this fixes a very serious bug found by Eric Blake in the current truncate filter. Rich.

This was a rather longer trip around the houses than I anticipated! The basic purpose of the patch series is to set $nbdkit_stdio_safe to "0" or "1" in sh and eval plugin scripts. To do that, I ended up adding a nicer way to manipulate environ lists, and to do that, I ended up adding a whole generic vector implementation which is applicable in a lot of different places.

Should we use attribute((nonnull)) at all? There's a very interesting history of this in libvirt -- try looking at commit eefb881 plus the commits referencing eefb881 -- but it does seem to work for me using recent GCC and Clang. I only did a few functions because annotating them gets old quickly... Rich.

...t random_state *state) { const uint64_t result_starstar = rotl (state->s[1] * 5, 7) * 9; diff --git a/common/regions/regions.h b/common/regions/regions.h index ca9b3d5..4fcaf09 100644 --- a/common/regions/regions.h +++ b/common/regions/regions.h @@ -76,25 +76,30 @@ struct regions { size_t nr_regions; }; -extern void init_regions (struct regions *regions); -extern void free_regions (struct regions *regions); -extern const struct region *find_region (const struct regions *regions, uint64_t offset); -extern int append_region (struct regions *regions, struct region region); +extern void init_r...

v1 was here: https://www.redhat.com/archives/libguestfs/2019-January/msg00008.html In v2 I have provided two patches: The first patch extends attribute((nonnull)) to most internal functions, but not to the external API. The second patch uses a macro so that attribute((format)) is only used in the public API on GCC or Clang. At least in theory these headers could be used by a C compiler which

...c const struct region *find_file_region (size_t i, size_t *j); +static const struct region *find_ebr_region (size_t i, size_t *j); + +/* Create the MBR and optionally EBRs. */ void -create_mbr_partition_table (unsigned char *out) +create_mbr_layout (void) { - size_t i, j; - - for (j = 0; j < nr_regions (&regions); ++j) { - const struct region *region = get_region (&regions, j); - - if (region->type == region_file) { - i = region->u.i; - assert (i < 4); - create_mbr_partition_table_entry (region, i == 0, - files[i].mbr_i...

An evolution of the patch I posted yesterday to qemu-devel (https://www.mail-archive.com/qemu-devel@nongnu.org/msg588920.html) which (a) works and (b) has a test. Rich.

This is a revised version of the two series previously posted here: https://www.redhat.com/archives/libguestfs/2019-January/msg00137.html https://www.redhat.com/archives/libguestfs/2019-January/msg00139.html There have been many smaller changes but the highlights are: - Using SECTOR_SIZE instead of hard-coding 512 everywhere. - Additional safety checks that the EBR chain doesn't jump

...lba = htole64 (region->start / SECTOR_SIZE); + entry->last_lba = htole64 (region->end / SECTOR_SIZE); + entry->attributes = htole64 (bootable ? 4 : 0); +} + +static void +create_gpt_partition_table (struct virtual_disk *disk, unsigned char *out) +{ + size_t j; + + for (j = 0; j < nr_regions (&disk->regions); ++j) { + const struct region *region = get_region (&disk->regions, j); + + /* Find the (only) partition region, which has type region_file. */ + if (region->type == region_file) { + create_gpt_partition_table_entry (region, true, +...

..._zero; + region.description = "file padding"; + if (region.len > 0 && append_region (&floppy->regions, region) == -1) + return -1; + } + + nbdkit_debug ("floppy: %zu regions, " + "total disk size %" PRIi64, + nr_regions (&floppy->regions), + virtual_size (&floppy->regions)); + + return 0; +} diff --git a/plugins/floppy/virtual-floppy.h b/plugins/floppy/virtual-floppy.h new file mode 100644 index 0000000..66b5e22 --- /dev/null +++ b/plugins/floppy/virtual-floppy.h @@ -0,0 +1,233 @@ +/...

...lba = htole64 (region->start / SECTOR_SIZE); + entry->last_lba = htole64 (region->end / SECTOR_SIZE); + entry->attributes = htole64 (bootable ? 4 : 0); +} + +static void +create_gpt_partition_table (struct virtual_disk *disk, unsigned char *out) +{ + size_t j; + + for (j = 0; j < nr_regions (&disk->regions); ++j) { + const struct region *region = get_region (&disk->regions, j); + + /* Find the (only) partition region, which has type region_file. */ + if (region->type == region_file) { + create_gpt_partition_table_entry (region, true, +...

For v3 I reimplemented this using mke2fs -d. This obviously makes the implementation a whole lot simpler, but cannot support multiple directory merging. Patches 1-3 are the same as before. I've also reproduced the notes from v2 below. v2: - Fix inconsistent tab/space. - All 3 plugins now contain a block of text pointing to the other 2 plugins. - TMDIR -> TMPDIR - Unlink the

Add nbdkit-floppy-plugin, “inspired” by qemu's VVFAT driver, but without the ability to handle writes. The implementation is pretty complete, supporting FAT32, LFNs, volume labels, timestamps, etc, and it passes both ‘make check’ and ‘make check-valgrind’. Usage is simple; to serve the current directory: $ nbdkit floppy . Then using guestfish (or any NBD client): $ guestfish --ro

Another interesting thing you can do with this plugin: https://rwmj.wordpress.com/2019/02/19/nbdkit-linuxdisk-plugin/ v2: - Fix inconsistent tab/space. - All 3 plugins now contain a block of text pointing to the other 2 plugins. - TMDIR -> TMPDIR - Unlink the temporary file and other cleanups along error paths. - fclose -> pclose, and check the return value for errors. -

Turns out Japanese trains are good for coding! In supermin we have a bunch of code to create the libguestfs appliance. It creates it directly using libext2fs (part of e2fsprogs). We can use the same technique to create ext2 virtual disks in nbdkit, which is what this new plugin does. Why a new plugin instead of modifying the floppy plugin? See the 4/4 commit message for an explanation. The