search for: truncate_prepare

...uncate/truncate.c index 6408c35..bed5a03 100644 --- a/filters/truncate/truncate.c +++ b/filters/truncate/truncate.c @@ -38,6 +38,7 @@ #include <string.h> #include <limits.h> #include <errno.h> +#include <inttypes.h> #include <nbdkit-filter.h> @@ -172,8 +173,14 @@ truncate_prepare (struct nbdkit_next_ops *next_ops, void *nxdata, */ if (truncate_size >= 0) h->size = truncate_size; - if (round_up > 0) + if (round_up > 0) { + if (ROUND_UP (h->size, round_up) > INT64_MAX) { + nbdkit_error ("cannot round size %" PRId64 " up t...

...2k, that changes client > A's use of 'real_size' to be something different than client A was > expecting). > > So, I think we have to move 'size' and 'real_size' into the > per-connection handle, at which point, can't we just set them once > during truncate_prepare by moving the existing guts of > truncate_get_size there, and then letting truncate_get_size just > return h->size while all other truncate_* functions just access > h->real_size without a lock? Even storing fields in the handle is not necessarily safe when using the parallel thread...

...+} + +static void +truncate_close (void *handle) +{ + struct handle *h = handle; + + free (h); +} + +/* In prepare, force a call to next_ops->get_size in order to set + * per-connection real_size & size; these values are not changed + * during the life of the connection. */ static int truncate_prepare (struct nbdkit_next_ops *next_ops, void *nxdata, void *handle) { int64_t r; - - r = truncate_get_size (next_ops, nxdata, handle); - return r >= 0 ? 0 : -1; -} - -/* Get the size. As a side effect, calculate the size to serve. */ -static int64_t -truncate_get_size (struc...

..." \ + "round-down=<N> Round down to multiple of N." + +static int64_t truncate_get_size (struct nbdkit_next_ops *next_ops, void *nxdata, void *handle); + +/* In prepare, force a call to get_size which sets the real_size & size + * globals. + */ +static int +truncate_prepare (struct nbdkit_next_ops *next_ops, void *nxdata, + void *handle) +{ + int64_t r; + + r = truncate_get_size (next_ops, nxdata, handle); + return r >= 0 ? 0 : -1; +} + +/* Get the size. As a side effect, calculate the size to serve. */ +static int64_t +truncate_get_size (struc...

...e 1k, then client B sees size 2k, that changes client A's use of 'real_size' to be something different than client A was expecting). So, I think we have to move 'size' and 'real_size' into the per-connection handle, at which point, can't we just set them once during truncate_prepare by moving the existing guts of truncate_get_size there, and then letting truncate_get_size just return h->size while all other truncate_* functions just access h->real_size without a lock? I guess I should look to see what other global filter state is better tracked per-filter rather than gl...

..." \ + "round-down=<N> Round down to multiple of N." + +static int64_t truncate_get_size (struct nbdkit_next_ops *next_ops, void *nxdata, void *handle); + +/* In prepare, force a call to get_size which sets the real_size & size + * globals. + */ +static int +truncate_prepare (struct nbdkit_next_ops *next_ops, void *nxdata, + void *handle) +{ + int64_t r; + + r = truncate_get_size (next_ops, nxdata, handle); + return r >= 0 ? 0 : -1; +} + +/* Get the size. As a side effect, calculate the size to serve. */ +static int64_t +truncate_get_size (struc...

While working on adding assertions to pthread_mutex_lock calls, I noticed that the truncate filter's use of mutex didn't really protect us, and isn't really necessary. Cleaning that up also spotted a couple of other potential cleanups. Eric Blake (4): filters: Drop useless .open callbacks truncate: Fix corruption when plugin changes per-connection size truncate: Test for safe

I do have a question about whether patch 2 is right, or whether I've exposed a bigger problem in the truncate (and possibly other) filter, but the rest seem fairly straightforward. Eric Blake (4): server: Check for pthread lock failures truncate: Factor out reading real_size under mutex plugins: Check for mutex failures filters: Check for mutex failures filters/cache/cache.c

...e *h = handle; + + if (h->real_size < h->size) + return 1; + return next_ops->init_sparse (nxdata); +} + /* Read data. */ static int truncate_pread (struct nbdkit_next_ops *next_ops, void *nxdata, @@ -426,6 +438,7 @@ static struct nbdkit_filter filter = { .prepare = truncate_prepare, .get_size = truncate_get_size, .can_fast_zero = truncate_can_fast_zero, + .init_sparse = truncate_init_sparse, .pread = truncate_pread, .pwrite = truncate_pwrite, .trim = truncate_trim, diff --git a/include/nbdkit-filter.h b/in...

I have dropped the map filter from this series for now while I try to get it working. However I think the truncate filter is in a good shape. This incorporates all feedback from Eric's review. Also there are three small fixes to the filter code, all revealed when I was testing using multiple filters which we'd not done much of before. Rich.

This patch series proposes two new filters. * truncate: This can truncate, extend, round up or round down the size of a plugin/device. A typical usage is to fix the qemu problem that it can only handle devices which are a multiple of 512-bytes: nbdkit --filter=truncate random size=500 round-up=512 This will serve a virtual device with size 512 bytes. Reading from the last 12 bytes will

I will be following up to this email with four separate threads each addressed to the appropriate single list, with proposed changes to: - the NBD protocol - qemu: both server and client - libnbd: client - nbdkit: server The feature in question adds a new optional NBD_INFO_ packet to the NBD_OPT_GO portion of handshake, adding up to 16 bits of information that the server can advertise to the

This is a cover letter to a series of patches being proposed in tandem to four different projects: - nbd: Document a new NBD_CMD_FLAG_FAST_ZERO command flag - qemu: Implement the flag for both clients and server - libnbd: Implement the flag for clients - nbdkit: Implement the flag for servers, including the nbd passthrough client If you want to test the patches together, I've pushed a