* [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor
@ 2013-05-28 15:27 Kevin Wolf
2013-05-28 15:27 ` [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command Kevin Wolf
` (16 more replies)
0 siblings, 17 replies; 45+ messages in thread
From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw)
To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino
This is a prerequisite for some kind of tests. It involves reorganising the
qemu-io code so that the command part can be separated and doesn't pollute the
global namespace any more, so we can link it with qemu.
Kevin Wolf (16):
qemu-io: Remove unused args_command
cutils: Support 'P' and 'E' suffixes in strtosz()
qemu-io: Make cvtnum() a wrapper around strtosz_suffix()
qemu-io: Handle cvtnum() errors in 'alloc'
qemu-io: Don't use global bs in command implementations
qemu-io: Split off commands to qemu-io-cmds.c
qemu-io: Factor out qemuio_command
qemu-io: Move 'help' function
qemu-io: Move 'quit' function
qemu-io: Move qemu_strsep() to cutils.c
qemu-io: Move functions for registering and running commands
qemu-io: Move command_loop() and friends
qemu-io: Move remaining helpers from cmd.c
qemu-io: Interface cleanup
qemu-io: Use the qemu version for -V
Make qemu-io commands available in the monitor
Makefile | 2 +-
Makefile.objs | 1 +
blockdev.c | 15 +
cmd.c | 612 -------------
cmd.h | 79 --
hmp-commands.hx | 16 +
hmp.c | 10 +
hmp.h | 1 +
include/qemu-common.h | 3 +
include/qemu-io.h | 46 +
monitor.c | 8 +-
qapi-schema.json | 16 +
qemu-img.c | 10 +-
qemu-io-cmds.c | 2118 ++++++++++++++++++++++++++++++++++++++++++++
qemu-io.c | 1988 ++++-------------------------------------
qmp-commands.hx | 28 +
tests/qemu-iotests/049.out | 8 +-
util/cutils.c | 25 +
18 files changed, 2466 insertions(+), 2520 deletions(-)
delete mode 100644 cmd.c
delete mode 100644 cmd.h
create mode 100644 include/qemu-io.h
create mode 100644 qemu-io-cmds.c
--
1.8.1.4
^ permalink raw reply [flat|nested] 45+ messages in thread* [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 16:18 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() Kevin Wolf ` (15 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino The original intention seems to be something with handling multiple images at once, but this has never been implemented and the only function ever registered is implemented to make everything behave like a "global" command. Just do that unconditionally now. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 28 ++-------------------------- cmd.h | 2 -- qemu-io.c | 10 ---------- 3 files changed, 2 insertions(+), 38 deletions(-) diff --git a/cmd.c b/cmd.c index 10a8688..4e7579b 100644 --- a/cmd.c +++ b/cmd.c @@ -34,7 +34,6 @@ cmdinfo_t *cmdtab; int ncmds; -static argsfunc_t args_func; static checkfunc_t check_func; static int ncmdline; static char **cmdline; @@ -127,22 +126,6 @@ void add_user_command(char *optarg) cmdline[ncmdline-1] = optarg; } -static int -args_command( - int index) -{ - if (args_func) - return args_func(index); - return 0; -} - -void -add_args_command( - argsfunc_t af) -{ - args_func = af; -} - static void prep_fetchline(void *opaque) { int *fetchable = opaque; @@ -155,7 +138,7 @@ static char *get_prompt(void); void command_loop(void) { - int c, i, j = 0, done = 0, fetchable = 0, prompted = 0; + int c, i, done = 0, fetchable = 0, prompted = 0; char *input; char **v; const cmdinfo_t *ct; @@ -171,14 +154,7 @@ void command_loop(void) if (c) { ct = find_command(v[0]); if (ct) { - if (ct->flags & CMD_FLAG_GLOBAL) { - done = command(ct, c, v); - } else { - j = 0; - while (!done && (j = args_command(j))) { - done = command(ct, c, v); - } - } + done = command(ct, c, v); } else { fprintf(stderr, _("command \"%s\" not found\n"), v[0]); } diff --git a/cmd.h b/cmd.h index b763b19..8e6f753 100644 --- a/cmd.h +++ b/cmd.h @@ -41,12 +41,10 @@ extern int ncmds; void help_init(void); void quit_init(void); -typedef int (*argsfunc_t)(int index); typedef int (*checkfunc_t)(const cmdinfo_t *ci); void add_command(const cmdinfo_t *ci); void add_user_command(char *optarg); -void add_args_command(argsfunc_t af); void add_check_command(checkfunc_t cf); const cmdinfo_t *find_command(const char *cmd); diff --git a/qemu-io.c b/qemu-io.c index 5e6680b..4288b8c 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -1888,15 +1888,6 @@ static int open_f(int argc, char **argv) return openfile(argv[optind], flags, growable); } -static int init_args_command(int index) -{ - /* only one device allowed so far */ - if (index >= 1) { - return 0; - } - return ++index; -} - static int init_check_command(const cmdinfo_t *ct) { if (ct->flags & CMD_FLAG_GLOBAL) { @@ -2043,7 +2034,6 @@ int main(int argc, char **argv) add_command(&wait_break_cmd); add_command(&abort_cmd); - add_args_command(init_args_command); add_check_command(init_check_command); /* open the device */ -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command 2013-05-28 15:27 ` [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command Kevin Wolf @ 2013-05-28 16:18 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-28 16:18 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 684 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > The original intention seems to be something with handling multiple > images at once, but this has never been implemented and the only > function ever registered is implemented to make everything behave like a > "global" command. Just do that unconditionally now. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 28 ++-------------------------- > cmd.h | 2 -- > qemu-io.c | 10 ---------- > 3 files changed, 2 insertions(+), 38 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 16:29 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() Kevin Wolf ` (14 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu-common.h | 2 ++ monitor.c | 8 ++++---- qemu-img.c | 10 ++++++---- tests/qemu-iotests/049.out | 8 ++++---- util/cutils.c | 4 ++++ 5 files changed, 20 insertions(+), 12 deletions(-) diff --git a/include/qemu-common.h b/include/qemu-common.h index cb82ef3..d95ea1e 100644 --- a/include/qemu-common.h +++ b/include/qemu-common.h @@ -191,6 +191,8 @@ int parse_uint_full(const char *s, unsigned long long *value, int base); * A-Z, as strtosz() will use qemu_toupper() on the given argument * prior to comparison. */ +#define STRTOSZ_DEFSUFFIX_EB 'E' +#define STRTOSZ_DEFSUFFIX_PB 'P' #define STRTOSZ_DEFSUFFIX_TB 'T' #define STRTOSZ_DEFSUFFIX_GB 'G' #define STRTOSZ_DEFSUFFIX_MB 'M' diff --git a/monitor.c b/monitor.c index 6ce2a4e..cfb93b8 100644 --- a/monitor.c +++ b/monitor.c @@ -93,10 +93,10 @@ * 'M' Non-negative target long (32 or 64 bit), in user mode the * value is multiplied by 2^20 (think Mebibyte) * 'o' octets (aka bytes) - * user mode accepts an optional T, t, G, g, M, m, K, k - * suffix, which multiplies the value by 2^40 for - * suffixes T and t, 2^30 for suffixes G and g, 2^20 for - * M and m, 2^10 for K and k + * user mode accepts an optional E, e, P, p, T, t, G, g, M, m, + * K, k suffix, which multiplies the value by 2^60 for suffixes E + * and e, 2^50 for suffixes P and p, 2^40 for suffixes T and t, + * 2^30 for suffixes G and g, 2^20 for M and m, 2^10 for K and k * 'T' double * user mode accepts an optional ms, us, ns suffix, * which divides the value by 1e3, 1e6, 1e9, respectively diff --git a/qemu-img.c b/qemu-img.c index a88428d..dd08990 100644 --- a/qemu-img.c +++ b/qemu-img.c @@ -84,8 +84,9 @@ static void help(void) " options are: 'none', 'writeback' (default, except for convert), 'writethrough',\n" " 'directsync' and 'unsafe' (default for convert)\n" " 'size' is the disk image size in bytes. Optional suffixes\n" - " 'k' or 'K' (kilobyte, 1024), 'M' (megabyte, 1024k), 'G' (gigabyte, 1024M)\n" - " and T (terabyte, 1024G) are supported. 'b' is ignored.\n" + " 'k' or 'K' (kilobyte, 1024), 'M' (megabyte, 1024k), 'G' (gigabyte, 1024M),\n" + " 'T' (terabyte, 1024G), 'P' (petabyte, 1024T) and 'E' (exabyte, 1024P) are\n" + " supported. 'b' is ignored.\n" " 'output_filename' is the destination disk image filename\n" " 'output_fmt' is the destination format\n" " 'options' is a comma separated list of format specific options in a\n" @@ -386,8 +387,9 @@ static int img_create(int argc, char **argv) error_report("Image size must be less than 8 EiB!"); } else { error_report("Invalid image size specified! You may use k, M, " - "G or T suffixes for "); - error_report("kilobytes, megabytes, gigabytes and terabytes."); + "G, T, P or E suffixes for "); + error_report("kilobytes, megabytes, gigabytes, terabytes, " + "petabytes and exabytes."); } return 1; } diff --git a/tests/qemu-iotests/049.out b/tests/qemu-iotests/049.out index 72db13f..d2f0efe 100644 --- a/tests/qemu-iotests/049.out +++ b/tests/qemu-iotests/049.out @@ -108,15 +108,15 @@ qemu-img: Formatting or formatting option not supported for file format 'qcow2' Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=-1024 encryption=off cluster_size=65536 lazy_refcounts=off qemu-img create -f qcow2 TEST_DIR/t.qcow2 -- 1kilobyte -qemu-img: Invalid image size specified! You may use k, M, G or T suffixes for -qemu-img: kilobytes, megabytes, gigabytes and terabytes. +qemu-img: Invalid image size specified! You may use k, M, G, T, P or E suffixes for +qemu-img: kilobytes, megabytes, gigabytes, terabytes, petabytes and exabytes. qemu-img create -f qcow2 -o size=1kilobyte TEST_DIR/t.qcow2 Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=1024 encryption=off cluster_size=65536 lazy_refcounts=off qemu-img create -f qcow2 TEST_DIR/t.qcow2 -- foobar -qemu-img: Invalid image size specified! You may use k, M, G or T suffixes for -qemu-img: kilobytes, megabytes, gigabytes and terabytes. +qemu-img: Invalid image size specified! You may use k, M, G, T, P or E suffixes for +qemu-img: kilobytes, megabytes, gigabytes, terabytes, petabytes and exabytes. qemu-img create -f qcow2 -o size=foobar TEST_DIR/t.qcow2 qemu-img: Parameter 'size' expects a size diff --git a/util/cutils.c b/util/cutils.c index a165819..8f28896 100644 --- a/util/cutils.c +++ b/util/cutils.c @@ -267,6 +267,10 @@ static int64_t suffix_mul(char suffix, int64_t unit) return unit * unit * unit; case STRTOSZ_DEFSUFFIX_TB: return unit * unit * unit * unit; + case STRTOSZ_DEFSUFFIX_PB: + return unit * unit * unit * unit * unit; + case STRTOSZ_DEFSUFFIX_EB: + return unit * unit * unit * unit * unit * unit; } return -1; } -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() 2013-05-28 15:27 ` [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() Kevin Wolf @ 2013-05-28 16:29 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-28 16:29 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 2076 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > include/qemu-common.h | 2 ++ > monitor.c | 8 ++++---- > qemu-img.c | 10 ++++++---- > tests/qemu-iotests/049.out | 8 ++++---- > util/cutils.c | 4 ++++ > 5 files changed, 20 insertions(+), 12 deletions(-) > Reviewed-by: Eric Blake <eblake@redhat.com> > +++ b/qemu-img.c > @@ -84,8 +84,9 @@ static void help(void) > " options are: 'none', 'writeback' (default, except for convert), 'writethrough',\n" > " 'directsync' and 'unsafe' (default for convert)\n" > " 'size' is the disk image size in bytes. Optional suffixes\n" > - " 'k' or 'K' (kilobyte, 1024), 'M' (megabyte, 1024k), 'G' (gigabyte, 1024M)\n" > - " and T (terabyte, 1024G) are supported. 'b' is ignored.\n" > + " 'k' or 'K' (kilobyte, 1024), 'M' (megabyte, 1024k), 'G' (gigabyte, 1024M),\n" > + " 'T' (terabyte, 1024G), 'P' (petabyte, 1024T) and 'E' (exabyte, 1024P) are\n" Pre-existing, but should we be favoring the technically correct kibibyte, mebibyte, gibibyte, ... terms, since we are treating it as powers of 1024, rather than the potentially confusing kilobyte term (since disk vendors favor powers of 1000)? > +++ b/util/cutils.c > @@ -267,6 +267,10 @@ static int64_t suffix_mul(char suffix, int64_t unit) > return unit * unit * unit; > case STRTOSZ_DEFSUFFIX_TB: > return unit * unit * unit * unit; > + case STRTOSZ_DEFSUFFIX_PB: > + return unit * unit * unit * unit * unit; > + case STRTOSZ_DEFSUFFIX_EB: > + return unit * unit * unit * unit * unit * unit; Pre-existing, but this code is not robust against overflow. And now that you are adding E, it's 6 fewer decimal characters that I have to type to pass in a value that will exceed even int64_t. -- 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 16:42 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' Kevin Wolf ` (13 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino No reason to implement the same thing multiple times. A nice side effect is that fractional numbers like 0.5M can be used in qemu-io now. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 37 ------------------------------------- cmd.h | 1 - qemu-io.c | 6 ++++++ 3 files changed, 6 insertions(+), 38 deletions(-) diff --git a/cmd.c b/cmd.c index 4e7579b..214c6f7 100644 --- a/cmd.c +++ b/cmd.c @@ -344,43 +344,6 @@ doneline( #define MEGABYTES(x) ((long long)(x) << 20) #define KILOBYTES(x) ((long long)(x) << 10) -long long -cvtnum( - char *s) -{ - long long i; - char *sp; - int c; - - i = strtoll(s, &sp, 0); - if (i == 0 && sp == s) - return -1LL; - if (*sp == '\0') - return i; - - if (sp[1] != '\0') - return -1LL; - - c = qemu_tolower(*sp); - switch (c) { - default: - return i; - case 'k': - return KILOBYTES(i); - case 'm': - return MEGABYTES(i); - case 'g': - return GIGABYTES(i); - case 't': - return TERABYTES(i); - case 'p': - return PETABYTES(i); - case 'e': - return EXABYTES(i); - } - return -1LL; -} - #define TO_EXABYTES(x) ((x) / EXABYTES(1)) #define TO_PETABYTES(x) ((x) / PETABYTES(1)) #define TO_TERABYTES(x) ((x) / TERABYTES(1)) diff --git a/cmd.h b/cmd.h index 8e6f753..4dcfe88 100644 --- a/cmd.h +++ b/cmd.h @@ -58,7 +58,6 @@ char **breakline(char *input, int *count); void doneline(char *input, char **vec); char *fetchline(void); -long long cvtnum(char *s); void cvtstr(double value, char *str, size_t sz); struct timeval tsub(struct timeval t1, struct timeval t2); diff --git a/qemu-io.c b/qemu-io.c index 4288b8c..8a719a8 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -29,6 +29,12 @@ static BlockDriverState *bs; static int misalign; +static int64_t cvtnum(const char *s) +{ + char *end; + return strtosz_suffix(s, &end, STRTOSZ_DEFSUFFIX_B); +} + /* * Parse the pattern argument to various sub-commands. * -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() 2013-05-28 15:27 ` [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() Kevin Wolf @ 2013-05-28 16:42 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-28 16:42 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 1508 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > No reason to implement the same thing multiple times. A nice side effect > is that fractional numbers like 0.5M can be used in qemu-io now. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 37 ------------------------------------- > cmd.h | 1 - > qemu-io.c | 6 ++++++ > 3 files changed, 6 insertions(+), 38 deletions(-) > - > - i = strtoll(s, &sp, 0); > - if (i == 0 && sp == s) > - return -1LL; > - if (*sp == '\0') > - return i; Another nice side effect: you are also getting rid of poor handling that fails to check for errors after strtoll() (such as clamping at INT64_MAX on EOVERFLOW). On the other hand, strtosz_suffix uses strtod() which may inadvertently cause rounding of low-order bits of large input numbers (that is, parsing with only 53 significant bits instead of 64 may cause us to round a value that was previously exact); but I'd rather deal with consistent behavior even if it rounds some values, than question which of multiple slightly different implementations will be parsing my values (we can also argue that perhaps strtosz_suffix could be improved to guarantee 64 bits of precision if there is no '.' in the value being parsed, while still providing the flexibility of 0.5G parsing when precision isn't important). 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (2 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 16:53 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations Kevin Wolf ` (12 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-io.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/qemu-io.c b/qemu-io.c index 8a719a8..b4f56fc 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -1596,7 +1596,10 @@ static int alloc_f(int argc, char **argv) int ret; offset = cvtnum(argv[1]); - if (offset & 0x1ff) { + if (offset < 0) { + printf("non-numeric offset argument -- %s\n", argv[1]); + return 0; + } else if (offset & 0x1ff) { printf("offset %" PRId64 " is not sector aligned\n", offset); return 0; @@ -1604,6 +1607,10 @@ static int alloc_f(int argc, char **argv) if (argc == 3) { nb_sectors = cvtnum(argv[2]); + if (nb_sectors < 0) { + printf("non-numeric length argument -- %s\n", argv[2]); + return 0; + } } else { nb_sectors = 1; } -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' 2013-05-28 15:27 ` [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' Kevin Wolf @ 2013-05-28 16:53 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-28 16:53 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 339 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > qemu-io.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (3 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 15:37 ` Kevin Wolf 2013-06-05 12:28 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c Kevin Wolf ` (11 subsequent siblings) 16 siblings, 2 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Pass in the BlockDriverState to the command handlers instead of using the global variable. This is an important step to make the commands usable outside of qemu-io. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 6 ++- cmd.h | 8 ++- qemu-io.c | 165 ++++++++++++++++++++++++++++++++++---------------------------- 3 files changed, 100 insertions(+), 79 deletions(-) diff --git a/cmd.c b/cmd.c index 214c6f7..d501aab 100644 --- a/cmd.c +++ b/cmd.c @@ -57,7 +57,7 @@ check_command( const cmdinfo_t *ci) { if (check_func) - return check_func(ci); + return check_func(qemuio_bs, ci); return 1; } @@ -103,7 +103,7 @@ command( return 0; } optind = 0; - return ct->cfunc(argc, argv); + return ct->cfunc(qemuio_bs, argc, argv); } const cmdinfo_t * @@ -452,6 +452,7 @@ static cmdinfo_t quit_cmd; /* ARGSUSED */ static int quit_f( + BlockDriverState *bs, int argc, char **argv) { @@ -490,6 +491,7 @@ help_all(void) static int help_f( + BlockDriverState *bs, int argc, char **argv) { diff --git a/cmd.h b/cmd.h index 4dcfe88..ccf6336 100644 --- a/cmd.h +++ b/cmd.h @@ -17,9 +17,13 @@ #ifndef __COMMAND_H__ #define __COMMAND_H__ +#include "qemu-common.h" + #define CMD_FLAG_GLOBAL ((int)0x80000000) /* don't iterate "args" */ -typedef int (*cfunc_t)(int argc, char **argv); +extern BlockDriverState *qemuio_bs; + +typedef int (*cfunc_t)(BlockDriverState *bs, int argc, char **argv); typedef void (*helpfunc_t)(void); typedef struct cmdinfo { @@ -41,7 +45,7 @@ extern int ncmds; void help_init(void); void quit_init(void); -typedef int (*checkfunc_t)(const cmdinfo_t *ci); +typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); void add_command(const cmdinfo_t *ci); void add_user_command(char *optarg); diff --git a/qemu-io.c b/qemu-io.c index b4f56fc..3abf9ff 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -25,8 +25,8 @@ #define CMD_NOFILE_OK 0x01 char *progname; -static BlockDriverState *bs; +BlockDriverState *qemuio_bs; static int misalign; static int64_t cvtnum(const char *s) @@ -63,7 +63,7 @@ static int parse_pattern(const char *arg) */ #define MISALIGN_OFFSET 16 -static void *qemu_io_alloc(size_t len, int pattern) +static void *qemu_io_alloc(BlockDriverState *bs, size_t len, int pattern) { void *buf; @@ -136,7 +136,8 @@ static void print_report(const char *op, struct timeval *t, int64_t offset, * vector matching it. */ static void * -create_iovec(QEMUIOVector *qiov, char **argv, int nr_iov, int pattern) +create_iovec(BlockDriverState *bs, QEMUIOVector *qiov, char **argv, int nr_iov, + int pattern) { size_t *sizes = g_new0(size_t, nr_iov); size_t count = 0; @@ -172,7 +173,7 @@ create_iovec(QEMUIOVector *qiov, char **argv, int nr_iov, int pattern) qemu_iovec_init(qiov, nr_iov); - buf = p = qemu_io_alloc(count, pattern); + buf = p = qemu_io_alloc(bs, count, pattern); for (i = 0; i < nr_iov; i++) { qemu_iovec_add(qiov, p, sizes[i]); @@ -184,7 +185,8 @@ fail: return buf; } -static int do_read(char *buf, int64_t offset, int count, int *total) +static int do_read(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) { int ret; @@ -196,7 +198,8 @@ static int do_read(char *buf, int64_t offset, int count, int *total) return 1; } -static int do_write(char *buf, int64_t offset, int count, int *total) +static int do_write(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) { int ret; @@ -208,7 +211,8 @@ static int do_write(char *buf, int64_t offset, int count, int *total) return 1; } -static int do_pread(char *buf, int64_t offset, int count, int *total) +static int do_pread(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) { *total = bdrv_pread(bs, offset, (uint8_t *)buf, count); if (*total < 0) { @@ -217,7 +221,8 @@ static int do_pread(char *buf, int64_t offset, int count, int *total) return 1; } -static int do_pwrite(char *buf, int64_t offset, int count, int *total) +static int do_pwrite(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) { *total = bdrv_pwrite(bs, offset, (uint8_t *)buf, count); if (*total < 0) { @@ -227,6 +232,7 @@ static int do_pwrite(char *buf, int64_t offset, int count, int *total) } typedef struct { + BlockDriverState *bs; int64_t offset; int count; int *total; @@ -238,7 +244,7 @@ static void coroutine_fn co_write_zeroes_entry(void *opaque) { CoWriteZeroes *data = opaque; - data->ret = bdrv_co_write_zeroes(bs, data->offset / BDRV_SECTOR_SIZE, + data->ret = bdrv_co_write_zeroes(data->bs, data->offset / BDRV_SECTOR_SIZE, data->count / BDRV_SECTOR_SIZE); data->done = true; if (data->ret < 0) { @@ -249,10 +255,12 @@ static void coroutine_fn co_write_zeroes_entry(void *opaque) *data->total = data->count; } -static int do_co_write_zeroes(int64_t offset, int count, int *total) +static int do_co_write_zeroes(BlockDriverState *bs, int64_t offset, int count, + int *total) { Coroutine *co; CoWriteZeroes data = { + .bs = bs, .offset = offset, .count = count, .total = total, @@ -271,7 +279,8 @@ static int do_co_write_zeroes(int64_t offset, int count, int *total) } } -static int do_write_compressed(char *buf, int64_t offset, int count, int *total) +static int do_write_compressed(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) { int ret; @@ -283,7 +292,8 @@ static int do_write_compressed(char *buf, int64_t offset, int count, int *total) return 1; } -static int do_load_vmstate(char *buf, int64_t offset, int count, int *total) +static int do_load_vmstate(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) { *total = bdrv_load_vmstate(bs, (uint8_t *)buf, offset, count); if (*total < 0) { @@ -292,7 +302,8 @@ static int do_load_vmstate(char *buf, int64_t offset, int count, int *total) return 1; } -static int do_save_vmstate(char *buf, int64_t offset, int count, int *total) +static int do_save_vmstate(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) { *total = bdrv_save_vmstate(bs, (uint8_t *)buf, offset, count); if (*total < 0) { @@ -307,7 +318,8 @@ static void aio_rw_done(void *opaque, int ret) *(int *)opaque = ret; } -static int do_aio_readv(QEMUIOVector *qiov, int64_t offset, int *total) +static int do_aio_readv(BlockDriverState *bs, QEMUIOVector *qiov, + int64_t offset, int *total) { int async_ret = NOT_DONE; @@ -321,7 +333,8 @@ static int do_aio_readv(QEMUIOVector *qiov, int64_t offset, int *total) return async_ret < 0 ? async_ret : 1; } -static int do_aio_writev(QEMUIOVector *qiov, int64_t offset, int *total) +static int do_aio_writev(BlockDriverState *bs, QEMUIOVector *qiov, + int64_t offset, int *total) { int async_ret = NOT_DONE; @@ -350,7 +363,8 @@ static void multiwrite_cb(void *opaque, int ret) } } -static int do_aio_multiwrite(BlockRequest* reqs, int num_reqs, int *total) +static int do_aio_multiwrite(BlockDriverState *bs, BlockRequest* reqs, + int num_reqs, int *total) { int i, ret; struct multiwrite_async_ret async_ret = { @@ -399,7 +413,7 @@ static void read_help(void) "\n"); } -static int read_f(int argc, char **argv); +static int read_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t read_cmd = { .name = "read", @@ -412,7 +426,7 @@ static const cmdinfo_t read_cmd = { .help = read_help, }; -static int read_f(int argc, char **argv) +static int read_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, pflag = 0, qflag = 0, vflag = 0; @@ -518,15 +532,15 @@ static int read_f(int argc, char **argv) } } - buf = qemu_io_alloc(count, 0xab); + buf = qemu_io_alloc(bs, count, 0xab); gettimeofday(&t1, NULL); if (pflag) { - cnt = do_pread(buf, offset, count, &total); + cnt = do_pread(bs, buf, offset, count, &total); } else if (bflag) { - cnt = do_load_vmstate(buf, offset, count, &total); + cnt = do_load_vmstate(bs, buf, offset, count, &total); } else { - cnt = do_read(buf, offset, count, &total); + cnt = do_read(bs, buf, offset, count, &total); } gettimeofday(&t2, NULL); @@ -583,7 +597,7 @@ static void readv_help(void) "\n"); } -static int readv_f(int argc, char **argv); +static int readv_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t readv_cmd = { .name = "readv", @@ -595,7 +609,7 @@ static const cmdinfo_t readv_cmd = { .help = readv_help, }; -static int readv_f(int argc, char **argv) +static int readv_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, qflag = 0, vflag = 0; @@ -651,13 +665,13 @@ static int readv_f(int argc, char **argv) } nr_iov = argc - optind; - buf = create_iovec(&qiov, &argv[optind], nr_iov, 0xab); + buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, 0xab); if (buf == NULL) { return 0; } gettimeofday(&t1, NULL); - cnt = do_aio_readv(&qiov, offset, &total); + cnt = do_aio_readv(bs, &qiov, offset, &total); gettimeofday(&t2, NULL); if (cnt < 0) { @@ -714,7 +728,7 @@ static void write_help(void) "\n"); } -static int write_f(int argc, char **argv); +static int write_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t write_cmd = { .name = "write", @@ -727,7 +741,7 @@ static const cmdinfo_t write_cmd = { .help = write_help, }; -static int write_f(int argc, char **argv) +static int write_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, pflag = 0, qflag = 0, bflag = 0, Pflag = 0, zflag = 0; @@ -814,20 +828,20 @@ static int write_f(int argc, char **argv) } if (!zflag) { - buf = qemu_io_alloc(count, pattern); + buf = qemu_io_alloc(bs, count, pattern); } gettimeofday(&t1, NULL); if (pflag) { - cnt = do_pwrite(buf, offset, count, &total); + cnt = do_pwrite(bs, buf, offset, count, &total); } else if (bflag) { - cnt = do_save_vmstate(buf, offset, count, &total); + cnt = do_save_vmstate(bs, buf, offset, count, &total); } else if (zflag) { - cnt = do_co_write_zeroes(offset, count, &total); + cnt = do_co_write_zeroes(bs, offset, count, &total); } else if (cflag) { - cnt = do_write_compressed(buf, offset, count, &total); + cnt = do_write_compressed(bs, buf, offset, count, &total); } else { - cnt = do_write(buf, offset, count, &total); + cnt = do_write(bs, buf, offset, count, &total); } gettimeofday(&t2, NULL); @@ -870,7 +884,7 @@ writev_help(void) "\n"); } -static int writev_f(int argc, char **argv); +static int writev_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t writev_cmd = { .name = "writev", @@ -882,7 +896,7 @@ static const cmdinfo_t writev_cmd = { .help = writev_help, }; -static int writev_f(int argc, char **argv) +static int writev_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, qflag = 0; @@ -932,13 +946,13 @@ static int writev_f(int argc, char **argv) } nr_iov = argc - optind; - buf = create_iovec(&qiov, &argv[optind], nr_iov, pattern); + buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, pattern); if (buf == NULL) { return 0; } gettimeofday(&t1, NULL); - cnt = do_aio_writev(&qiov, offset, &total); + cnt = do_aio_writev(bs, &qiov, offset, &total); gettimeofday(&t2, NULL); if (cnt < 0) { @@ -979,7 +993,7 @@ static void multiwrite_help(void) "\n"); } -static int multiwrite_f(int argc, char **argv); +static int multiwrite_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t multiwrite_cmd = { .name = "multiwrite", @@ -991,7 +1005,7 @@ static const cmdinfo_t multiwrite_cmd = { .help = multiwrite_help, }; -static int multiwrite_f(int argc, char **argv) +static int multiwrite_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, qflag = 0; @@ -1072,7 +1086,7 @@ static int multiwrite_f(int argc, char **argv) nr_iov = j - optind; /* Build request */ - buf[i] = create_iovec(&qiovs[i], &argv[optind], nr_iov, pattern); + buf[i] = create_iovec(bs, &qiovs[i], &argv[optind], nr_iov, pattern); if (buf[i] == NULL) { goto out; } @@ -1090,7 +1104,7 @@ static int multiwrite_f(int argc, char **argv) nr_reqs = i; gettimeofday(&t1, NULL); - cnt = do_aio_multiwrite(reqs, nr_reqs, &total); + cnt = do_aio_multiwrite(bs, reqs, nr_reqs, &total); gettimeofday(&t2, NULL); if (cnt < 0) { @@ -1218,7 +1232,7 @@ static void aio_read_help(void) "\n"); } -static int aio_read_f(int argc, char **argv); +static int aio_read_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t aio_read_cmd = { .name = "aio_read", @@ -1230,7 +1244,7 @@ static const cmdinfo_t aio_read_cmd = { .help = aio_read_help, }; -static int aio_read_f(int argc, char **argv) +static int aio_read_f(BlockDriverState *bs, int argc, char **argv) { int nr_iov, c; struct aio_ctx *ctx = g_new0(struct aio_ctx, 1); @@ -1281,7 +1295,7 @@ static int aio_read_f(int argc, char **argv) } nr_iov = argc - optind; - ctx->buf = create_iovec(&ctx->qiov, &argv[optind], nr_iov, 0xab); + ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, 0xab); if (ctx->buf == NULL) { g_free(ctx); return 0; @@ -1313,7 +1327,7 @@ static void aio_write_help(void) "\n"); } -static int aio_write_f(int argc, char **argv); +static int aio_write_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t aio_write_cmd = { .name = "aio_write", @@ -1325,7 +1339,7 @@ static const cmdinfo_t aio_write_cmd = { .help = aio_write_help, }; -static int aio_write_f(int argc, char **argv) +static int aio_write_f(BlockDriverState *bs, int argc, char **argv) { int nr_iov, c; int pattern = 0xcd; @@ -1373,7 +1387,7 @@ static int aio_write_f(int argc, char **argv) } nr_iov = argc - optind; - ctx->buf = create_iovec(&ctx->qiov, &argv[optind], nr_iov, pattern); + ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, pattern); if (ctx->buf == NULL) { g_free(ctx); return 0; @@ -1385,7 +1399,7 @@ static int aio_write_f(int argc, char **argv) return 0; } -static int aio_flush_f(int argc, char **argv) +static int aio_flush_f(BlockDriverState *bs, int argc, char **argv) { bdrv_drain_all(); return 0; @@ -1397,7 +1411,7 @@ static const cmdinfo_t aio_flush_cmd = { .oneline = "completes all outstanding aio requests" }; -static int flush_f(int argc, char **argv) +static int flush_f(BlockDriverState *bs, int argc, char **argv) { bdrv_flush(bs); return 0; @@ -1410,7 +1424,7 @@ static const cmdinfo_t flush_cmd = { .oneline = "flush all in-core file state to disk", }; -static int truncate_f(int argc, char **argv) +static int truncate_f(BlockDriverState *bs, int argc, char **argv) { int64_t offset; int ret; @@ -1440,7 +1454,7 @@ static const cmdinfo_t truncate_cmd = { .oneline = "truncates the current file at the given offset", }; -static int length_f(int argc, char **argv) +static int length_f(BlockDriverState *bs, int argc, char **argv) { int64_t size; char s1[64]; @@ -1465,7 +1479,7 @@ static const cmdinfo_t length_cmd = { }; -static int info_f(int argc, char **argv) +static int info_f(BlockDriverState *bs, int argc, char **argv) { BlockDriverInfo bdi; char s1[64], s2[64]; @@ -1516,7 +1530,7 @@ static void discard_help(void) "\n"); } -static int discard_f(int argc, char **argv); +static int discard_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t discard_cmd = { .name = "discard", @@ -1529,7 +1543,7 @@ static const cmdinfo_t discard_cmd = { .help = discard_help, }; -static int discard_f(int argc, char **argv) +static int discard_f(BlockDriverState *bs, int argc, char **argv) { struct timeval t1, t2; int Cflag = 0, qflag = 0; @@ -1587,7 +1601,7 @@ out: return 0; } -static int alloc_f(int argc, char **argv) +static int alloc_f(BlockDriverState *bs, int argc, char **argv) { int64_t offset, sector_num; int nb_sectors, remaining; @@ -1649,7 +1663,8 @@ static const cmdinfo_t alloc_cmd = { }; -static int map_is_allocated(int64_t sector_num, int64_t nb_sectors, int64_t *pnum) +static int map_is_allocated(BlockDriverState *bs, int64_t sector_num, + int64_t nb_sectors, int64_t *pnum) { int num, num_checked; int ret, firstret; @@ -1679,7 +1694,7 @@ static int map_is_allocated(int64_t sector_num, int64_t nb_sectors, int64_t *pnu return firstret; } -static int map_f(int argc, char **argv) +static int map_f(BlockDriverState *bs, int argc, char **argv) { int64_t offset; int64_t nb_sectors; @@ -1692,7 +1707,7 @@ static int map_f(int argc, char **argv) nb_sectors = bs->total_sectors; do { - ret = map_is_allocated(offset, nb_sectors, &num); + ret = map_is_allocated(bs, offset, nb_sectors, &num); if (ret < 0) { error_report("Failed to get allocation status: %s", strerror(-ret)); return 0; @@ -1720,7 +1735,7 @@ static const cmdinfo_t map_cmd = { .oneline = "prints the allocated areas of a file", }; -static int break_f(int argc, char **argv) +static int break_f(BlockDriverState *bs, int argc, char **argv) { int ret; @@ -1742,7 +1757,7 @@ static const cmdinfo_t break_cmd = { "request as tag", }; -static int resume_f(int argc, char **argv) +static int resume_f(BlockDriverState *bs, int argc, char **argv) { int ret; @@ -1763,7 +1778,7 @@ static const cmdinfo_t resume_cmd = { .oneline = "resumes the request tagged as tag", }; -static int wait_break_f(int argc, char **argv) +static int wait_break_f(BlockDriverState *bs, int argc, char **argv) { while (!bdrv_debug_is_suspended(bs, argv[1])) { qemu_aio_wait(); @@ -1781,7 +1796,7 @@ static const cmdinfo_t wait_break_cmd = { .oneline = "waits for the suspension of a request", }; -static int abort_f(int argc, char **argv) +static int abort_f(BlockDriverState *bs, int argc, char **argv) { abort(); } @@ -1793,7 +1808,7 @@ static const cmdinfo_t abort_cmd = { .oneline = "simulate a program crash using abort(3)", }; -static int close_f(int argc, char **argv) +static int close_f(BlockDriverState *bs, int argc, char **argv) { bdrv_delete(bs); bs = NULL; @@ -1809,23 +1824,23 @@ static const cmdinfo_t close_cmd = { static int openfile(char *name, int flags, int growable) { - if (bs) { + if (qemuio_bs) { fprintf(stderr, "file open already, try 'help close'\n"); return 1; } if (growable) { - if (bdrv_file_open(&bs, name, NULL, flags)) { + if (bdrv_file_open(&qemuio_bs, name, NULL, flags)) { fprintf(stderr, "%s: can't open device %s\n", progname, name); return 1; } } else { - bs = bdrv_new("hda"); + qemuio_bs = bdrv_new("hda"); - if (bdrv_open(bs, name, NULL, flags, NULL) < 0) { + if (bdrv_open(qemuio_bs, name, NULL, flags, NULL) < 0) { fprintf(stderr, "%s: can't open device %s\n", progname, name); - bdrv_delete(bs); - bs = NULL; + bdrv_delete(qemuio_bs); + qemuio_bs = NULL; return 1; } } @@ -1850,7 +1865,7 @@ static void open_help(void) "\n"); } -static int open_f(int argc, char **argv); +static int open_f(BlockDriverState *bs, int argc, char **argv); static const cmdinfo_t open_cmd = { .name = "open", @@ -1864,7 +1879,7 @@ static const cmdinfo_t open_cmd = { .help = open_help, }; -static int open_f(int argc, char **argv) +static int open_f(BlockDriverState *bs, int argc, char **argv) { int flags = 0; int readonly = 0; @@ -1901,7 +1916,7 @@ static int open_f(int argc, char **argv) return openfile(argv[optind], flags, growable); } -static int init_check_command(const cmdinfo_t *ct) +static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) { if (ct->flags & CMD_FLAG_GLOBAL) { return 1; @@ -2064,8 +2079,8 @@ int main(int argc, char **argv) */ bdrv_drain_all(); - if (bs) { - bdrv_delete(bs); + if (qemuio_bs) { + bdrv_delete(qemuio_bs); } return 0; } -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations 2013-05-28 15:27 ` [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations Kevin Wolf @ 2013-05-28 15:37 ` Kevin Wolf 2013-05-28 17:02 ` Eric Blake 2013-06-05 12:28 ` Stefan Hajnoczi 1 sibling, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:37 UTC (permalink / raw) To: qemu-devel; +Cc: pbonzini, stefanha, lcapitulino Am 28.05.2013 um 17:27 hat Kevin Wolf geschrieben: > Pass in the BlockDriverState to the command handlers instead of using > the global variable. This is an important step to make the commands > usable outside of qemu-io. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > @@ -1793,7 +1808,7 @@ static const cmdinfo_t abort_cmd = { > .oneline = "simulate a program crash using abort(3)", > }; > > -static int close_f(int argc, char **argv) > +static int close_f(BlockDriverState *bs, int argc, char **argv) > { > bdrv_delete(bs); > bs = NULL; Here I missed the following hunk that I've applied locally now: --- a/qemu-io.c +++ b/qemu-io.c @@ -1811,7 +1811,7 @@ static const cmdinfo_t abort_cmd = { static int close_f(BlockDriverState *bs, int argc, char **argv) { bdrv_delete(bs); - bs = NULL; + qemuio_bs = NULL; return 0; } ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations 2013-05-28 15:37 ` Kevin Wolf @ 2013-05-28 17:02 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-28 17:02 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 1127 bytes --] On 05/28/2013 09:37 AM, Kevin Wolf wrote: > Am 28.05.2013 um 17:27 hat Kevin Wolf geschrieben: >> Pass in the BlockDriverState to the command handlers instead of using >> the global variable. This is an important step to make the commands >> usable outside of qemu-io. >> >> Signed-off-by: Kevin Wolf <kwolf@redhat.com> > >> @@ -1793,7 +1808,7 @@ static const cmdinfo_t abort_cmd = { >> .oneline = "simulate a program crash using abort(3)", >> }; >> >> -static int close_f(int argc, char **argv) >> +static int close_f(BlockDriverState *bs, int argc, char **argv) >> { >> bdrv_delete(bs); >> bs = NULL; > > Here I missed the following hunk that I've applied locally now: > > --- a/qemu-io.c > +++ b/qemu-io.c > @@ -1811,7 +1811,7 @@ static const cmdinfo_t abort_cmd = { > static int close_f(BlockDriverState *bs, int argc, char **argv) > { > bdrv_delete(bs); > - bs = NULL; > + qemuio_bs = NULL; 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations 2013-05-28 15:27 ` [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations Kevin Wolf 2013-05-28 15:37 ` Kevin Wolf @ 2013-06-05 12:28 ` Stefan Hajnoczi 2013-06-05 12:39 ` Kevin Wolf 1 sibling, 1 reply; 45+ messages in thread From: Stefan Hajnoczi @ 2013-06-05 12:28 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino On Tue, May 28, 2013 at 05:27:25PM +0200, Kevin Wolf wrote: > Pass in the BlockDriverState to the command handlers instead of using > the global variable. This is an important step to make the commands > usable outside of qemu-io. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 6 ++- > cmd.h | 8 ++- > qemu-io.c | 165 ++++++++++++++++++++++++++++++++++---------------------------- > 3 files changed, 100 insertions(+), 79 deletions(-) > > diff --git a/cmd.c b/cmd.c > index 214c6f7..d501aab 100644 > --- a/cmd.c > +++ b/cmd.c > @@ -57,7 +57,7 @@ check_command( > const cmdinfo_t *ci) > { > if (check_func) > - return check_func(ci); > + return check_func(qemuio_bs, ci); > return 1; > } > > @@ -103,7 +103,7 @@ command( > return 0; > } > optind = 0; > - return ct->cfunc(argc, argv); > + return ct->cfunc(qemuio_bs, argc, argv); > } > > const cmdinfo_t * > @@ -452,6 +452,7 @@ static cmdinfo_t quit_cmd; > /* ARGSUSED */ > static int > quit_f( > + BlockDriverState *bs, > int argc, tabs vs spaces. I try to keep the existing style unless I decide to reformat the entire section of code. Not trying to start a flamewar but this file appears to use tabs and IMO you should stick to that instead of mixing spaces :-). > char **argv) > { > @@ -490,6 +491,7 @@ help_all(void) > > static int > help_f( > + BlockDriverState *bs, > int argc, > char **argv) > { > diff --git a/cmd.h b/cmd.h > index 4dcfe88..ccf6336 100644 > --- a/cmd.h > +++ b/cmd.h > @@ -17,9 +17,13 @@ > #ifndef __COMMAND_H__ > #define __COMMAND_H__ > > +#include "qemu-common.h" > + > #define CMD_FLAG_GLOBAL ((int)0x80000000) /* don't iterate "args" */ > > -typedef int (*cfunc_t)(int argc, char **argv); > +extern BlockDriverState *qemuio_bs; > + > +typedef int (*cfunc_t)(BlockDriverState *bs, int argc, char **argv); > typedef void (*helpfunc_t)(void); > > typedef struct cmdinfo { > @@ -41,7 +45,7 @@ extern int ncmds; > void help_init(void); > void quit_init(void); > > -typedef int (*checkfunc_t)(const cmdinfo_t *ci); > +typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); > > void add_command(const cmdinfo_t *ci); > void add_user_command(char *optarg); cmd.h does not know about the block layer. I would use void *opaque instead of BlockDriverState *bs. That way the file stays generic and can be used in other command-line tools. ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations 2013-06-05 12:28 ` Stefan Hajnoczi @ 2013-06-05 12:39 ` Kevin Wolf 0 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-06-05 12:39 UTC (permalink / raw) To: Stefan Hajnoczi; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino Am 05.06.2013 um 14:28 hat Stefan Hajnoczi geschrieben: > On Tue, May 28, 2013 at 05:27:25PM +0200, Kevin Wolf wrote: > > Pass in the BlockDriverState to the command handlers instead of using > > the global variable. This is an important step to make the commands > > usable outside of qemu-io. > > > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > > --- > > cmd.c | 6 ++- > > cmd.h | 8 ++- > > qemu-io.c | 165 ++++++++++++++++++++++++++++++++++---------------------------- > > 3 files changed, 100 insertions(+), 79 deletions(-) > > > > diff --git a/cmd.c b/cmd.c > > index 214c6f7..d501aab 100644 > > --- a/cmd.c > > +++ b/cmd.c > > @@ -57,7 +57,7 @@ check_command( > > const cmdinfo_t *ci) > > { > > if (check_func) > > - return check_func(ci); > > + return check_func(qemuio_bs, ci); > > return 1; > > } > > > > @@ -103,7 +103,7 @@ command( > > return 0; > > } > > optind = 0; > > - return ct->cfunc(argc, argv); > > + return ct->cfunc(qemuio_bs, argc, argv); > > } > > > > const cmdinfo_t * > > @@ -452,6 +452,7 @@ static cmdinfo_t quit_cmd; > > /* ARGSUSED */ > > static int > > quit_f( > > + BlockDriverState *bs, > > int argc, > > tabs vs spaces. I try to keep the existing style unless I decide to > reformat the entire section of code. > > Not trying to start a flamewar but this file appears to use tabs and IMO > you should stick to that instead of mixing spaces :-). Ah yes, didn't notice that. Doesn't really matter though, at the end of the series cmd.c is gone. > > --- a/cmd.h > > +++ b/cmd.h > > @@ -17,9 +17,13 @@ > > #ifndef __COMMAND_H__ > > #define __COMMAND_H__ > > > > +#include "qemu-common.h" > > + > > #define CMD_FLAG_GLOBAL ((int)0x80000000) /* don't iterate "args" */ > > > > -typedef int (*cfunc_t)(int argc, char **argv); > > +extern BlockDriverState *qemuio_bs; > > + > > +typedef int (*cfunc_t)(BlockDriverState *bs, int argc, char **argv); > > typedef void (*helpfunc_t)(void); > > > > typedef struct cmdinfo { > > @@ -41,7 +45,7 @@ extern int ncmds; > > void help_init(void); > > void quit_init(void); > > > > -typedef int (*checkfunc_t)(const cmdinfo_t *ci); > > +typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); > > > > void add_command(const cmdinfo_t *ci); > > void add_user_command(char *optarg); > > cmd.h does not know about the block layer. I would use void *opaque > instead of BlockDriverState *bs. That way the file stays generic and > can be used in other command-line tools. Do you plan to use this in different context? Because this series is exactly the opposite of keeping it generic. It moves everything directly into qemu-io. Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (4 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-29 20:29 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command Kevin Wolf ` (10 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino This is the implementation of all qemu-io commands that make sense to be called from the qemu monitor, i.e. everything except open, close and quit. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- Makefile | 2 +- qemu-io-cmds.c | 1835 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ qemu-io.c | 1817 +------------------------------------------------------ 3 files changed, 1838 insertions(+), 1816 deletions(-) create mode 100644 qemu-io-cmds.c diff --git a/Makefile b/Makefile index 9695c9d..df2212b 100644 --- a/Makefile +++ b/Makefile @@ -186,7 +186,7 @@ qemu-img.o: qemu-img-cmds.h qemu-img$(EXESUF): qemu-img.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-nbd$(EXESUF): qemu-nbd.o $(block-obj-y) libqemuutil.a libqemustub.a -qemu-io$(EXESUF): qemu-io.o cmd.o $(block-obj-y) libqemuutil.a libqemustub.a +qemu-io$(EXESUF): qemu-io.o qemu-io-cmds.o cmd.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-bridge-helper$(EXESUF): qemu-bridge-helper.o diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c new file mode 100644 index 0000000..57b751e --- /dev/null +++ b/qemu-io-cmds.c @@ -0,0 +1,1835 @@ +/* + * Command line utility to exercise the QEMU I/O path. + * + * Copyright (C) 2009 Red Hat, Inc. + * Copyright (c) 2003-2005 Silicon Graphics, Inc. + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#include "qemu-common.h" +#include "block/block_int.h" +#include "cmd.h" + +#define CMD_NOFILE_OK 0x01 + +int qemuio_misalign; + +static int64_t cvtnum(const char *s) +{ + char *end; + return strtosz_suffix(s, &end, STRTOSZ_DEFSUFFIX_B); +} + +/* + * Parse the pattern argument to various sub-commands. + * + * Because the pattern is used as an argument to memset it must evaluate + * to an unsigned integer that fits into a single byte. + */ +static int parse_pattern(const char *arg) +{ + char *endptr = NULL; + long pattern; + + pattern = strtol(arg, &endptr, 0); + if (pattern < 0 || pattern > UCHAR_MAX || *endptr != '\0') { + printf("%s is not a valid pattern byte\n", arg); + return -1; + } + + return pattern; +} + +/* + * Memory allocation helpers. + * + * Make sure memory is aligned by default, or purposefully qemuio_misaligned if + * that is specified on the command line. + */ + +#define MISALIGN_OFFSET 16 +static void *qemu_io_alloc(BlockDriverState *bs, size_t len, int pattern) +{ + void *buf; + + if (qemuio_misalign) { + len += MISALIGN_OFFSET; + } + buf = qemu_blockalign(bs, len); + memset(buf, pattern, len); + if (qemuio_misalign) { + buf += MISALIGN_OFFSET; + } + return buf; +} + +static void qemu_io_free(void *p) +{ + if (qemuio_misalign) { + p -= MISALIGN_OFFSET; + } + qemu_vfree(p); +} + +static void dump_buffer(const void *buffer, int64_t offset, int len) +{ + int i, j; + const uint8_t *p; + + for (i = 0, p = buffer; i < len; i += 16) { + const uint8_t *s = p; + + printf("%08" PRIx64 ": ", offset + i); + for (j = 0; j < 16 && i + j < len; j++, p++) { + printf("%02x ", *p); + } + printf(" "); + for (j = 0; j < 16 && i + j < len; j++, s++) { + if (isalnum(*s)) { + printf("%c", *s); + } else { + printf("."); + } + } + printf("\n"); + } +} + +static void print_report(const char *op, struct timeval *t, int64_t offset, + int count, int total, int cnt, int Cflag) +{ + char s1[64], s2[64], ts[64]; + + timestr(t, ts, sizeof(ts), Cflag ? VERBOSE_FIXED_TIME : 0); + if (!Cflag) { + cvtstr((double)total, s1, sizeof(s1)); + cvtstr(tdiv((double)total, *t), s2, sizeof(s2)); + printf("%s %d/%d bytes at offset %" PRId64 "\n", + op, total, count, offset); + printf("%s, %d ops; %s (%s/sec and %.4f ops/sec)\n", + s1, cnt, ts, s2, tdiv((double)cnt, *t)); + } else {/* bytes,ops,time,bytes/sec,ops/sec */ + printf("%d,%d,%s,%.3f,%.3f\n", + total, cnt, ts, + tdiv((double)total, *t), + tdiv((double)cnt, *t)); + } +} + +/* + * Parse multiple length statements for vectored I/O, and construct an I/O + * vector matching it. + */ +static void * +create_iovec(BlockDriverState *bs, QEMUIOVector *qiov, char **argv, int nr_iov, + int pattern) +{ + size_t *sizes = g_new0(size_t, nr_iov); + size_t count = 0; + void *buf = NULL; + void *p; + int i; + + for (i = 0; i < nr_iov; i++) { + char *arg = argv[i]; + int64_t len; + + len = cvtnum(arg); + if (len < 0) { + printf("non-numeric length argument -- %s\n", arg); + goto fail; + } + + /* should be SIZE_T_MAX, but that doesn't exist */ + if (len > INT_MAX) { + printf("too large length argument -- %s\n", arg); + goto fail; + } + + if (len & 0x1ff) { + printf("length argument %" PRId64 + " is not sector aligned\n", len); + goto fail; + } + + sizes[i] = len; + count += len; + } + + qemu_iovec_init(qiov, nr_iov); + + buf = p = qemu_io_alloc(bs, count, pattern); + + for (i = 0; i < nr_iov; i++) { + qemu_iovec_add(qiov, p, sizes[i]); + p += sizes[i]; + } + +fail: + g_free(sizes); + return buf; +} + +static int do_read(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) +{ + int ret; + + ret = bdrv_read(bs, offset >> 9, (uint8_t *)buf, count >> 9); + if (ret < 0) { + return ret; + } + *total = count; + return 1; +} + +static int do_write(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) +{ + int ret; + + ret = bdrv_write(bs, offset >> 9, (uint8_t *)buf, count >> 9); + if (ret < 0) { + return ret; + } + *total = count; + return 1; +} + +static int do_pread(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) +{ + *total = bdrv_pread(bs, offset, (uint8_t *)buf, count); + if (*total < 0) { + return *total; + } + return 1; +} + +static int do_pwrite(BlockDriverState *bs, char *buf, int64_t offset, int count, + int *total) +{ + *total = bdrv_pwrite(bs, offset, (uint8_t *)buf, count); + if (*total < 0) { + return *total; + } + return 1; +} + +typedef struct { + BlockDriverState *bs; + int64_t offset; + int count; + int *total; + int ret; + bool done; +} CoWriteZeroes; + +static void coroutine_fn co_write_zeroes_entry(void *opaque) +{ + CoWriteZeroes *data = opaque; + + data->ret = bdrv_co_write_zeroes(data->bs, data->offset / BDRV_SECTOR_SIZE, + data->count / BDRV_SECTOR_SIZE); + data->done = true; + if (data->ret < 0) { + *data->total = data->ret; + return; + } + + *data->total = data->count; +} + +static int do_co_write_zeroes(BlockDriverState *bs, int64_t offset, int count, + int *total) +{ + Coroutine *co; + CoWriteZeroes data = { + .bs = bs, + .offset = offset, + .count = count, + .total = total, + .done = false, + }; + + co = qemu_coroutine_create(co_write_zeroes_entry); + qemu_coroutine_enter(co, &data); + while (!data.done) { + qemu_aio_wait(); + } + if (data.ret < 0) { + return data.ret; + } else { + return 1; + } +} + +static int do_write_compressed(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) +{ + int ret; + + ret = bdrv_write_compressed(bs, offset >> 9, (uint8_t *)buf, count >> 9); + if (ret < 0) { + return ret; + } + *total = count; + return 1; +} + +static int do_load_vmstate(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) +{ + *total = bdrv_load_vmstate(bs, (uint8_t *)buf, offset, count); + if (*total < 0) { + return *total; + } + return 1; +} + +static int do_save_vmstate(BlockDriverState *bs, char *buf, int64_t offset, + int count, int *total) +{ + *total = bdrv_save_vmstate(bs, (uint8_t *)buf, offset, count); + if (*total < 0) { + return *total; + } + return 1; +} + +#define NOT_DONE 0x7fffffff +static void aio_rw_done(void *opaque, int ret) +{ + *(int *)opaque = ret; +} + +static int do_aio_readv(BlockDriverState *bs, QEMUIOVector *qiov, + int64_t offset, int *total) +{ + int async_ret = NOT_DONE; + + bdrv_aio_readv(bs, offset >> 9, qiov, qiov->size >> 9, + aio_rw_done, &async_ret); + while (async_ret == NOT_DONE) { + main_loop_wait(false); + } + + *total = qiov->size; + return async_ret < 0 ? async_ret : 1; +} + +static int do_aio_writev(BlockDriverState *bs, QEMUIOVector *qiov, + int64_t offset, int *total) +{ + int async_ret = NOT_DONE; + + bdrv_aio_writev(bs, offset >> 9, qiov, qiov->size >> 9, + aio_rw_done, &async_ret); + while (async_ret == NOT_DONE) { + main_loop_wait(false); + } + + *total = qiov->size; + return async_ret < 0 ? async_ret : 1; +} + +struct multiwrite_async_ret { + int num_done; + int error; +}; + +static void multiwrite_cb(void *opaque, int ret) +{ + struct multiwrite_async_ret *async_ret = opaque; + + async_ret->num_done++; + if (ret < 0) { + async_ret->error = ret; + } +} + +static int do_aio_multiwrite(BlockDriverState *bs, BlockRequest* reqs, + int num_reqs, int *total) +{ + int i, ret; + struct multiwrite_async_ret async_ret = { + .num_done = 0, + .error = 0, + }; + + *total = 0; + for (i = 0; i < num_reqs; i++) { + reqs[i].cb = multiwrite_cb; + reqs[i].opaque = &async_ret; + *total += reqs[i].qiov->size; + } + + ret = bdrv_aio_multiwrite(bs, reqs, num_reqs); + if (ret < 0) { + return ret; + } + + while (async_ret.num_done < num_reqs) { + main_loop_wait(false); + } + + return async_ret.error < 0 ? async_ret.error : 1; +} + +static void read_help(void) +{ + printf( +"\n" +" reads a range of bytes from the given offset\n" +"\n" +" Example:\n" +" 'read -v 512 1k' - dumps 1 kilobyte read from 512 bytes into the file\n" +"\n" +" Reads a segment of the currently open file, optionally dumping it to the\n" +" standard output stream (with -v option) for subsequent inspection.\n" +" -b, -- read from the VM state rather than the virtual disk\n" +" -C, -- report statistics in a machine parsable format\n" +" -l, -- length for pattern verification (only with -P)\n" +" -p, -- use bdrv_pread to read the file\n" +" -P, -- use a pattern to verify read data\n" +" -q, -- quiet mode, do not show I/O statistics\n" +" -s, -- start offset for pattern verification (only with -P)\n" +" -v, -- dump buffer to standard output\n" +"\n"); +} + +static int read_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t read_cmd = { + .name = "read", + .altname = "r", + .cfunc = read_f, + .argmin = 2, + .argmax = -1, + .args = "[-abCpqv] [-P pattern [-s off] [-l len]] off len", + .oneline = "reads a number of bytes at a specified offset", + .help = read_help, +}; + +static int read_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, pflag = 0, qflag = 0, vflag = 0; + int Pflag = 0, sflag = 0, lflag = 0, bflag = 0; + int c, cnt; + char *buf; + int64_t offset; + int count; + /* Some compilers get confused and warn if this is not initialized. */ + int total = 0; + int pattern = 0, pattern_offset = 0, pattern_count = 0; + + while ((c = getopt(argc, argv, "bCl:pP:qs:v")) != EOF) { + switch (c) { + case 'b': + bflag = 1; + break; + case 'C': + Cflag = 1; + break; + case 'l': + lflag = 1; + pattern_count = cvtnum(optarg); + if (pattern_count < 0) { + printf("non-numeric length argument -- %s\n", optarg); + return 0; + } + break; + case 'p': + pflag = 1; + break; + case 'P': + Pflag = 1; + pattern = parse_pattern(optarg); + if (pattern < 0) { + return 0; + } + break; + case 'q': + qflag = 1; + break; + case 's': + sflag = 1; + pattern_offset = cvtnum(optarg); + if (pattern_offset < 0) { + printf("non-numeric length argument -- %s\n", optarg); + return 0; + } + break; + case 'v': + vflag = 1; + break; + default: + return command_usage(&read_cmd); + } + } + + if (optind != argc - 2) { + return command_usage(&read_cmd); + } + + if (bflag && pflag) { + printf("-b and -p cannot be specified at the same time\n"); + return 0; + } + + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + optind++; + count = cvtnum(argv[optind]); + if (count < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + if (!Pflag && (lflag || sflag)) { + return command_usage(&read_cmd); + } + + if (!lflag) { + pattern_count = count - pattern_offset; + } + + if ((pattern_count < 0) || (pattern_count + pattern_offset > count)) { + printf("pattern verification range exceeds end of read data\n"); + return 0; + } + + if (!pflag) { + if (offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + offset); + return 0; + } + if (count & 0x1ff) { + printf("count %d is not sector aligned\n", + count); + return 0; + } + } + + buf = qemu_io_alloc(bs, count, 0xab); + + gettimeofday(&t1, NULL); + if (pflag) { + cnt = do_pread(bs, buf, offset, count, &total); + } else if (bflag) { + cnt = do_load_vmstate(bs, buf, offset, count, &total); + } else { + cnt = do_read(bs, buf, offset, count, &total); + } + gettimeofday(&t2, NULL); + + if (cnt < 0) { + printf("read failed: %s\n", strerror(-cnt)); + goto out; + } + + if (Pflag) { + void *cmp_buf = g_malloc(pattern_count); + memset(cmp_buf, pattern, pattern_count); + if (memcmp(buf + pattern_offset, cmp_buf, pattern_count)) { + printf("Pattern verification failed at offset %" + PRId64 ", %d bytes\n", + offset + pattern_offset, pattern_count); + } + g_free(cmp_buf); + } + + if (qflag) { + goto out; + } + + if (vflag) { + dump_buffer(buf, offset, count); + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, t1); + print_report("read", &t2, offset, count, total, cnt, Cflag); + +out: + qemu_io_free(buf); + + return 0; +} + +static void readv_help(void) +{ + printf( +"\n" +" reads a range of bytes from the given offset into multiple buffers\n" +"\n" +" Example:\n" +" 'readv -v 512 1k 1k ' - dumps 2 kilobytes read from 512 bytes into the file\n" +"\n" +" Reads a segment of the currently open file, optionally dumping it to the\n" +" standard output stream (with -v option) for subsequent inspection.\n" +" Uses multiple iovec buffers if more than one byte range is specified.\n" +" -C, -- report statistics in a machine parsable format\n" +" -P, -- use a pattern to verify read data\n" +" -v, -- dump buffer to standard output\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int readv_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t readv_cmd = { + .name = "readv", + .cfunc = readv_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cqv] [-P pattern ] off len [len..]", + .oneline = "reads a number of bytes at a specified offset", + .help = readv_help, +}; + +static int readv_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, qflag = 0, vflag = 0; + int c, cnt; + char *buf; + int64_t offset; + /* Some compilers get confused and warn if this is not initialized. */ + int total = 0; + int nr_iov; + QEMUIOVector qiov; + int pattern = 0; + int Pflag = 0; + + while ((c = getopt(argc, argv, "CP:qv")) != EOF) { + switch (c) { + case 'C': + Cflag = 1; + break; + case 'P': + Pflag = 1; + pattern = parse_pattern(optarg); + if (pattern < 0) { + return 0; + } + break; + case 'q': + qflag = 1; + break; + case 'v': + vflag = 1; + break; + default: + return command_usage(&readv_cmd); + } + } + + if (optind > argc - 2) { + return command_usage(&readv_cmd); + } + + + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + optind++; + + if (offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + offset); + return 0; + } + + nr_iov = argc - optind; + buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, 0xab); + if (buf == NULL) { + return 0; + } + + gettimeofday(&t1, NULL); + cnt = do_aio_readv(bs, &qiov, offset, &total); + gettimeofday(&t2, NULL); + + if (cnt < 0) { + printf("readv failed: %s\n", strerror(-cnt)); + goto out; + } + + if (Pflag) { + void *cmp_buf = g_malloc(qiov.size); + memset(cmp_buf, pattern, qiov.size); + if (memcmp(buf, cmp_buf, qiov.size)) { + printf("Pattern verification failed at offset %" + PRId64 ", %zd bytes\n", offset, qiov.size); + } + g_free(cmp_buf); + } + + if (qflag) { + goto out; + } + + if (vflag) { + dump_buffer(buf, offset, qiov.size); + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, t1); + print_report("read", &t2, offset, qiov.size, total, cnt, Cflag); + +out: + qemu_iovec_destroy(&qiov); + qemu_io_free(buf); + return 0; +} + +static void write_help(void) +{ + printf( +"\n" +" writes a range of bytes from the given offset\n" +"\n" +" Example:\n" +" 'write 512 1k' - writes 1 kilobyte at 512 bytes into the open file\n" +"\n" +" Writes into a segment of the currently open file, using a buffer\n" +" filled with a set pattern (0xcdcdcdcd).\n" +" -b, -- write to the VM state rather than the virtual disk\n" +" -c, -- write compressed data with bdrv_write_compressed\n" +" -p, -- use bdrv_pwrite to write the file\n" +" -P, -- use different pattern to fill file\n" +" -C, -- report statistics in a machine parsable format\n" +" -q, -- quiet mode, do not show I/O statistics\n" +" -z, -- write zeroes using bdrv_co_write_zeroes\n" +"\n"); +} + +static int write_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t write_cmd = { + .name = "write", + .altname = "w", + .cfunc = write_f, + .argmin = 2, + .argmax = -1, + .args = "[-bcCpqz] [-P pattern ] off len", + .oneline = "writes a number of bytes at a specified offset", + .help = write_help, +}; + +static int write_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, pflag = 0, qflag = 0, bflag = 0, Pflag = 0, zflag = 0; + int cflag = 0; + int c, cnt; + char *buf = NULL; + int64_t offset; + int count; + /* Some compilers get confused and warn if this is not initialized. */ + int total = 0; + int pattern = 0xcd; + + while ((c = getopt(argc, argv, "bcCpP:qz")) != EOF) { + switch (c) { + case 'b': + bflag = 1; + break; + case 'c': + cflag = 1; + break; + case 'C': + Cflag = 1; + break; + case 'p': + pflag = 1; + break; + case 'P': + Pflag = 1; + pattern = parse_pattern(optarg); + if (pattern < 0) { + return 0; + } + break; + case 'q': + qflag = 1; + break; + case 'z': + zflag = 1; + break; + default: + return command_usage(&write_cmd); + } + } + + if (optind != argc - 2) { + return command_usage(&write_cmd); + } + + if (bflag + pflag + zflag > 1) { + printf("-b, -p, or -z cannot be specified at the same time\n"); + return 0; + } + + if (zflag && Pflag) { + printf("-z and -P cannot be specified at the same time\n"); + return 0; + } + + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + optind++; + count = cvtnum(argv[optind]); + if (count < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + if (!pflag) { + if (offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + offset); + return 0; + } + + if (count & 0x1ff) { + printf("count %d is not sector aligned\n", + count); + return 0; + } + } + + if (!zflag) { + buf = qemu_io_alloc(bs, count, pattern); + } + + gettimeofday(&t1, NULL); + if (pflag) { + cnt = do_pwrite(bs, buf, offset, count, &total); + } else if (bflag) { + cnt = do_save_vmstate(bs, buf, offset, count, &total); + } else if (zflag) { + cnt = do_co_write_zeroes(bs, offset, count, &total); + } else if (cflag) { + cnt = do_write_compressed(bs, buf, offset, count, &total); + } else { + cnt = do_write(bs, buf, offset, count, &total); + } + gettimeofday(&t2, NULL); + + if (cnt < 0) { + printf("write failed: %s\n", strerror(-cnt)); + goto out; + } + + if (qflag) { + goto out; + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, t1); + print_report("wrote", &t2, offset, count, total, cnt, Cflag); + +out: + if (!zflag) { + qemu_io_free(buf); + } + + return 0; +} + +static void +writev_help(void) +{ + printf( +"\n" +" writes a range of bytes from the given offset source from multiple buffers\n" +"\n" +" Example:\n" +" 'write 512 1k 1k' - writes 2 kilobytes at 512 bytes into the open file\n" +"\n" +" Writes into a segment of the currently open file, using a buffer\n" +" filled with a set pattern (0xcdcdcdcd).\n" +" -P, -- use different pattern to fill file\n" +" -C, -- report statistics in a machine parsable format\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int writev_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t writev_cmd = { + .name = "writev", + .cfunc = writev_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cq] [-P pattern ] off len [len..]", + .oneline = "writes a number of bytes at a specified offset", + .help = writev_help, +}; + +static int writev_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, qflag = 0; + int c, cnt; + char *buf; + int64_t offset; + /* Some compilers get confused and warn if this is not initialized. */ + int total = 0; + int nr_iov; + int pattern = 0xcd; + QEMUIOVector qiov; + + while ((c = getopt(argc, argv, "CqP:")) != EOF) { + switch (c) { + case 'C': + Cflag = 1; + break; + case 'q': + qflag = 1; + break; + case 'P': + pattern = parse_pattern(optarg); + if (pattern < 0) { + return 0; + } + break; + default: + return command_usage(&writev_cmd); + } + } + + if (optind > argc - 2) { + return command_usage(&writev_cmd); + } + + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + optind++; + + if (offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + offset); + return 0; + } + + nr_iov = argc - optind; + buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, pattern); + if (buf == NULL) { + return 0; + } + + gettimeofday(&t1, NULL); + cnt = do_aio_writev(bs, &qiov, offset, &total); + gettimeofday(&t2, NULL); + + if (cnt < 0) { + printf("writev failed: %s\n", strerror(-cnt)); + goto out; + } + + if (qflag) { + goto out; + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, t1); + print_report("wrote", &t2, offset, qiov.size, total, cnt, Cflag); +out: + qemu_iovec_destroy(&qiov); + qemu_io_free(buf); + return 0; +} + +static void multiwrite_help(void) +{ + printf( +"\n" +" writes a range of bytes from the given offset source from multiple buffers,\n" +" in a batch of requests that may be merged by qemu\n" +"\n" +" Example:\n" +" 'multiwrite 512 1k 1k ; 4k 1k'\n" +" writes 2 kB at 512 bytes and 1 kB at 4 kB into the open file\n" +"\n" +" Writes into a segment of the currently open file, using a buffer\n" +" filled with a set pattern (0xcdcdcdcd). The pattern byte is increased\n" +" by one for each request contained in the multiwrite command.\n" +" -P, -- use different pattern to fill file\n" +" -C, -- report statistics in a machine parsable format\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int multiwrite_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t multiwrite_cmd = { + .name = "multiwrite", + .cfunc = multiwrite_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cq] [-P pattern ] off len [len..] [; off len [len..]..]", + .oneline = "issues multiple write requests at once", + .help = multiwrite_help, +}; + +static int multiwrite_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, qflag = 0; + int c, cnt; + char **buf; + int64_t offset, first_offset = 0; + /* Some compilers get confused and warn if this is not initialized. */ + int total = 0; + int nr_iov; + int nr_reqs; + int pattern = 0xcd; + QEMUIOVector *qiovs; + int i; + BlockRequest *reqs; + + while ((c = getopt(argc, argv, "CqP:")) != EOF) { + switch (c) { + case 'C': + Cflag = 1; + break; + case 'q': + qflag = 1; + break; + case 'P': + pattern = parse_pattern(optarg); + if (pattern < 0) { + return 0; + } + break; + default: + return command_usage(&writev_cmd); + } + } + + if (optind > argc - 2) { + return command_usage(&writev_cmd); + } + + nr_reqs = 1; + for (i = optind; i < argc; i++) { + if (!strcmp(argv[i], ";")) { + nr_reqs++; + } + } + + reqs = g_malloc0(nr_reqs * sizeof(*reqs)); + buf = g_malloc0(nr_reqs * sizeof(*buf)); + qiovs = g_malloc(nr_reqs * sizeof(*qiovs)); + + for (i = 0; i < nr_reqs && optind < argc; i++) { + int j; + + /* Read the offset of the request */ + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric offset argument -- %s\n", argv[optind]); + goto out; + } + optind++; + + if (offset & 0x1ff) { + printf("offset %lld is not sector aligned\n", + (long long)offset); + goto out; + } + + if (i == 0) { + first_offset = offset; + } + + /* Read lengths for qiov entries */ + for (j = optind; j < argc; j++) { + if (!strcmp(argv[j], ";")) { + break; + } + } + + nr_iov = j - optind; + + /* Build request */ + buf[i] = create_iovec(bs, &qiovs[i], &argv[optind], nr_iov, pattern); + if (buf[i] == NULL) { + goto out; + } + + reqs[i].qiov = &qiovs[i]; + reqs[i].sector = offset >> 9; + reqs[i].nb_sectors = reqs[i].qiov->size >> 9; + + optind = j + 1; + + pattern++; + } + + /* If there were empty requests at the end, ignore them */ + nr_reqs = i; + + gettimeofday(&t1, NULL); + cnt = do_aio_multiwrite(bs, reqs, nr_reqs, &total); + gettimeofday(&t2, NULL); + + if (cnt < 0) { + printf("aio_multiwrite failed: %s\n", strerror(-cnt)); + goto out; + } + + if (qflag) { + goto out; + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, t1); + print_report("wrote", &t2, first_offset, total, total, cnt, Cflag); +out: + for (i = 0; i < nr_reqs; i++) { + qemu_io_free(buf[i]); + if (reqs[i].qiov != NULL) { + qemu_iovec_destroy(&qiovs[i]); + } + } + g_free(buf); + g_free(reqs); + g_free(qiovs); + return 0; +} + +struct aio_ctx { + QEMUIOVector qiov; + int64_t offset; + char *buf; + int qflag; + int vflag; + int Cflag; + int Pflag; + int pattern; + struct timeval t1; +}; + +static void aio_write_done(void *opaque, int ret) +{ + struct aio_ctx *ctx = opaque; + struct timeval t2; + + gettimeofday(&t2, NULL); + + + if (ret < 0) { + printf("aio_write failed: %s\n", strerror(-ret)); + goto out; + } + + if (ctx->qflag) { + goto out; + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, ctx->t1); + print_report("wrote", &t2, ctx->offset, ctx->qiov.size, + ctx->qiov.size, 1, ctx->Cflag); +out: + qemu_io_free(ctx->buf); + qemu_iovec_destroy(&ctx->qiov); + g_free(ctx); +} + +static void aio_read_done(void *opaque, int ret) +{ + struct aio_ctx *ctx = opaque; + struct timeval t2; + + gettimeofday(&t2, NULL); + + if (ret < 0) { + printf("readv failed: %s\n", strerror(-ret)); + goto out; + } + + if (ctx->Pflag) { + void *cmp_buf = g_malloc(ctx->qiov.size); + + memset(cmp_buf, ctx->pattern, ctx->qiov.size); + if (memcmp(ctx->buf, cmp_buf, ctx->qiov.size)) { + printf("Pattern verification failed at offset %" + PRId64 ", %zd bytes\n", ctx->offset, ctx->qiov.size); + } + g_free(cmp_buf); + } + + if (ctx->qflag) { + goto out; + } + + if (ctx->vflag) { + dump_buffer(ctx->buf, ctx->offset, ctx->qiov.size); + } + + /* Finally, report back -- -C gives a parsable format */ + t2 = tsub(t2, ctx->t1); + print_report("read", &t2, ctx->offset, ctx->qiov.size, + ctx->qiov.size, 1, ctx->Cflag); +out: + qemu_io_free(ctx->buf); + qemu_iovec_destroy(&ctx->qiov); + g_free(ctx); +} + +static void aio_read_help(void) +{ + printf( +"\n" +" asynchronously reads a range of bytes from the given offset\n" +"\n" +" Example:\n" +" 'aio_read -v 512 1k 1k ' - dumps 2 kilobytes read from 512 bytes into the file\n" +"\n" +" Reads a segment of the currently open file, optionally dumping it to the\n" +" standard output stream (with -v option) for subsequent inspection.\n" +" The read is performed asynchronously and the aio_flush command must be\n" +" used to ensure all outstanding aio requests have been completed.\n" +" -C, -- report statistics in a machine parsable format\n" +" -P, -- use a pattern to verify read data\n" +" -v, -- dump buffer to standard output\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int aio_read_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t aio_read_cmd = { + .name = "aio_read", + .cfunc = aio_read_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cqv] [-P pattern ] off len [len..]", + .oneline = "asynchronously reads a number of bytes", + .help = aio_read_help, +}; + +static int aio_read_f(BlockDriverState *bs, int argc, char **argv) +{ + int nr_iov, c; + struct aio_ctx *ctx = g_new0(struct aio_ctx, 1); + + while ((c = getopt(argc, argv, "CP:qv")) != EOF) { + switch (c) { + case 'C': + ctx->Cflag = 1; + break; + case 'P': + ctx->Pflag = 1; + ctx->pattern = parse_pattern(optarg); + if (ctx->pattern < 0) { + g_free(ctx); + return 0; + } + break; + case 'q': + ctx->qflag = 1; + break; + case 'v': + ctx->vflag = 1; + break; + default: + g_free(ctx); + return command_usage(&aio_read_cmd); + } + } + + if (optind > argc - 2) { + g_free(ctx); + return command_usage(&aio_read_cmd); + } + + ctx->offset = cvtnum(argv[optind]); + if (ctx->offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + g_free(ctx); + return 0; + } + optind++; + + if (ctx->offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + ctx->offset); + g_free(ctx); + return 0; + } + + nr_iov = argc - optind; + ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, 0xab); + if (ctx->buf == NULL) { + g_free(ctx); + return 0; + } + + gettimeofday(&ctx->t1, NULL); + bdrv_aio_readv(bs, ctx->offset >> 9, &ctx->qiov, + ctx->qiov.size >> 9, aio_read_done, ctx); + return 0; +} + +static void aio_write_help(void) +{ + printf( +"\n" +" asynchronously writes a range of bytes from the given offset source\n" +" from multiple buffers\n" +"\n" +" Example:\n" +" 'aio_write 512 1k 1k' - writes 2 kilobytes at 512 bytes into the open file\n" +"\n" +" Writes into a segment of the currently open file, using a buffer\n" +" filled with a set pattern (0xcdcdcdcd).\n" +" The write is performed asynchronously and the aio_flush command must be\n" +" used to ensure all outstanding aio requests have been completed.\n" +" -P, -- use different pattern to fill file\n" +" -C, -- report statistics in a machine parsable format\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int aio_write_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t aio_write_cmd = { + .name = "aio_write", + .cfunc = aio_write_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cq] [-P pattern ] off len [len..]", + .oneline = "asynchronously writes a number of bytes", + .help = aio_write_help, +}; + +static int aio_write_f(BlockDriverState *bs, int argc, char **argv) +{ + int nr_iov, c; + int pattern = 0xcd; + struct aio_ctx *ctx = g_new0(struct aio_ctx, 1); + + while ((c = getopt(argc, argv, "CqP:")) != EOF) { + switch (c) { + case 'C': + ctx->Cflag = 1; + break; + case 'q': + ctx->qflag = 1; + break; + case 'P': + pattern = parse_pattern(optarg); + if (pattern < 0) { + g_free(ctx); + return 0; + } + break; + default: + g_free(ctx); + return command_usage(&aio_write_cmd); + } + } + + if (optind > argc - 2) { + g_free(ctx); + return command_usage(&aio_write_cmd); + } + + ctx->offset = cvtnum(argv[optind]); + if (ctx->offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + g_free(ctx); + return 0; + } + optind++; + + if (ctx->offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + ctx->offset); + g_free(ctx); + return 0; + } + + nr_iov = argc - optind; + ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, pattern); + if (ctx->buf == NULL) { + g_free(ctx); + return 0; + } + + gettimeofday(&ctx->t1, NULL); + bdrv_aio_writev(bs, ctx->offset >> 9, &ctx->qiov, + ctx->qiov.size >> 9, aio_write_done, ctx); + return 0; +} + +static int aio_flush_f(BlockDriverState *bs, int argc, char **argv) +{ + bdrv_drain_all(); + return 0; +} + +static const cmdinfo_t aio_flush_cmd = { + .name = "aio_flush", + .cfunc = aio_flush_f, + .oneline = "completes all outstanding aio requests" +}; + +static int flush_f(BlockDriverState *bs, int argc, char **argv) +{ + bdrv_flush(bs); + return 0; +} + +static const cmdinfo_t flush_cmd = { + .name = "flush", + .altname = "f", + .cfunc = flush_f, + .oneline = "flush all in-core file state to disk", +}; + +static int truncate_f(BlockDriverState *bs, int argc, char **argv) +{ + int64_t offset; + int ret; + + offset = cvtnum(argv[1]); + if (offset < 0) { + printf("non-numeric truncate argument -- %s\n", argv[1]); + return 0; + } + + ret = bdrv_truncate(bs, offset); + if (ret < 0) { + printf("truncate: %s\n", strerror(-ret)); + return 0; + } + + return 0; +} + +static const cmdinfo_t truncate_cmd = { + .name = "truncate", + .altname = "t", + .cfunc = truncate_f, + .argmin = 1, + .argmax = 1, + .args = "off", + .oneline = "truncates the current file at the given offset", +}; + +static int length_f(BlockDriverState *bs, int argc, char **argv) +{ + int64_t size; + char s1[64]; + + size = bdrv_getlength(bs); + if (size < 0) { + printf("getlength: %s\n", strerror(-size)); + return 0; + } + + cvtstr(size, s1, sizeof(s1)); + printf("%s\n", s1); + return 0; +} + + +static const cmdinfo_t length_cmd = { + .name = "length", + .altname = "l", + .cfunc = length_f, + .oneline = "gets the length of the current file", +}; + + +static int info_f(BlockDriverState *bs, int argc, char **argv) +{ + BlockDriverInfo bdi; + char s1[64], s2[64]; + int ret; + + if (bs->drv && bs->drv->format_name) { + printf("format name: %s\n", bs->drv->format_name); + } + if (bs->drv && bs->drv->protocol_name) { + printf("format name: %s\n", bs->drv->protocol_name); + } + + ret = bdrv_get_info(bs, &bdi); + if (ret) { + return 0; + } + + cvtstr(bdi.cluster_size, s1, sizeof(s1)); + cvtstr(bdi.vm_state_offset, s2, sizeof(s2)); + + printf("cluster size: %s\n", s1); + printf("vm state offset: %s\n", s2); + + return 0; +} + + + +static const cmdinfo_t info_cmd = { + .name = "info", + .altname = "i", + .cfunc = info_f, + .oneline = "prints information about the current file", +}; + +static void discard_help(void) +{ + printf( +"\n" +" discards a range of bytes from the given offset\n" +"\n" +" Example:\n" +" 'discard 512 1k' - discards 1 kilobyte from 512 bytes into the file\n" +"\n" +" Discards a segment of the currently open file.\n" +" -C, -- report statistics in a machine parsable format\n" +" -q, -- quiet mode, do not show I/O statistics\n" +"\n"); +} + +static int discard_f(BlockDriverState *bs, int argc, char **argv); + +static const cmdinfo_t discard_cmd = { + .name = "discard", + .altname = "d", + .cfunc = discard_f, + .argmin = 2, + .argmax = -1, + .args = "[-Cq] off len", + .oneline = "discards a number of bytes at a specified offset", + .help = discard_help, +}; + +static int discard_f(BlockDriverState *bs, int argc, char **argv) +{ + struct timeval t1, t2; + int Cflag = 0, qflag = 0; + int c, ret; + int64_t offset; + int count; + + while ((c = getopt(argc, argv, "Cq")) != EOF) { + switch (c) { + case 'C': + Cflag = 1; + break; + case 'q': + qflag = 1; + break; + default: + return command_usage(&discard_cmd); + } + } + + if (optind != argc - 2) { + return command_usage(&discard_cmd); + } + + offset = cvtnum(argv[optind]); + if (offset < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + optind++; + count = cvtnum(argv[optind]); + if (count < 0) { + printf("non-numeric length argument -- %s\n", argv[optind]); + return 0; + } + + gettimeofday(&t1, NULL); + ret = bdrv_discard(bs, offset >> BDRV_SECTOR_BITS, + count >> BDRV_SECTOR_BITS); + gettimeofday(&t2, NULL); + + if (ret < 0) { + printf("discard failed: %s\n", strerror(-ret)); + goto out; + } + + /* Finally, report back -- -C gives a parsable format */ + if (!qflag) { + t2 = tsub(t2, t1); + print_report("discard", &t2, offset, count, count, 1, Cflag); + } + +out: + return 0; +} + +static int alloc_f(BlockDriverState *bs, int argc, char **argv) +{ + int64_t offset, sector_num; + int nb_sectors, remaining; + char s1[64]; + int num, sum_alloc; + int ret; + + offset = cvtnum(argv[1]); + if (offset < 0) { + printf("non-numeric offset argument -- %s\n", argv[1]); + return 0; + } else if (offset & 0x1ff) { + printf("offset %" PRId64 " is not sector aligned\n", + offset); + return 0; + } + + if (argc == 3) { + nb_sectors = cvtnum(argv[2]); + if (nb_sectors < 0) { + printf("non-numeric length argument -- %s\n", argv[2]); + return 0; + } + } else { + nb_sectors = 1; + } + + remaining = nb_sectors; + sum_alloc = 0; + sector_num = offset >> 9; + while (remaining) { + ret = bdrv_is_allocated(bs, sector_num, remaining, &num); + sector_num += num; + remaining -= num; + if (ret) { + sum_alloc += num; + } + if (num == 0) { + nb_sectors -= remaining; + remaining = 0; + } + } + + cvtstr(offset, s1, sizeof(s1)); + + printf("%d/%d sectors allocated at offset %s\n", + sum_alloc, nb_sectors, s1); + return 0; +} + +static const cmdinfo_t alloc_cmd = { + .name = "alloc", + .altname = "a", + .argmin = 1, + .argmax = 2, + .cfunc = alloc_f, + .args = "off [sectors]", + .oneline = "checks if a sector is present in the file", +}; + + +static int map_is_allocated(BlockDriverState *bs, int64_t sector_num, + int64_t nb_sectors, int64_t *pnum) +{ + int num, num_checked; + int ret, firstret; + + num_checked = MIN(nb_sectors, INT_MAX); + ret = bdrv_is_allocated(bs, sector_num, num_checked, &num); + if (ret < 0) { + return ret; + } + + firstret = ret; + *pnum = num; + + while (nb_sectors > 0 && ret == firstret) { + sector_num += num; + nb_sectors -= num; + + num_checked = MIN(nb_sectors, INT_MAX); + ret = bdrv_is_allocated(bs, sector_num, num_checked, &num); + if (ret == firstret) { + *pnum += num; + } else { + break; + } + } + + return firstret; +} + +static int map_f(BlockDriverState *bs, int argc, char **argv) +{ + int64_t offset; + int64_t nb_sectors; + char s1[64]; + int64_t num; + int ret; + const char *retstr; + + offset = 0; + nb_sectors = bs->total_sectors; + + do { + ret = map_is_allocated(bs, offset, nb_sectors, &num); + if (ret < 0) { + error_report("Failed to get allocation status: %s", strerror(-ret)); + return 0; + } + + retstr = ret ? " allocated" : "not allocated"; + cvtstr(offset << 9ULL, s1, sizeof(s1)); + printf("[% 24" PRId64 "] % 8" PRId64 "/% 8" PRId64 " sectors %s " + "at offset %s (%d)\n", + offset << 9ULL, num, nb_sectors, retstr, s1, ret); + + offset += num; + nb_sectors -= num; + } while (offset < bs->total_sectors); + + return 0; +} + +static const cmdinfo_t map_cmd = { + .name = "map", + .argmin = 0, + .argmax = 0, + .cfunc = map_f, + .args = "", + .oneline = "prints the allocated areas of a file", +}; + +static int break_f(BlockDriverState *bs, int argc, char **argv) +{ + int ret; + + ret = bdrv_debug_breakpoint(bs, argv[1], argv[2]); + if (ret < 0) { + printf("Could not set breakpoint: %s\n", strerror(-ret)); + } + + return 0; +} + +static const cmdinfo_t break_cmd = { + .name = "break", + .argmin = 2, + .argmax = 2, + .cfunc = break_f, + .args = "event tag", + .oneline = "sets a breakpoint on event and tags the stopped " + "request as tag", +}; + +static int resume_f(BlockDriverState *bs, int argc, char **argv) +{ + int ret; + + ret = bdrv_debug_resume(bs, argv[1]); + if (ret < 0) { + printf("Could not resume request: %s\n", strerror(-ret)); + } + + return 0; +} + +static const cmdinfo_t resume_cmd = { + .name = "resume", + .argmin = 1, + .argmax = 1, + .cfunc = resume_f, + .args = "tag", + .oneline = "resumes the request tagged as tag", +}; + +static int wait_break_f(BlockDriverState *bs, int argc, char **argv) +{ + while (!bdrv_debug_is_suspended(bs, argv[1])) { + qemu_aio_wait(); + } + + return 0; +} + +static const cmdinfo_t wait_break_cmd = { + .name = "wait_break", + .argmin = 1, + .argmax = 1, + .cfunc = wait_break_f, + .args = "tag", + .oneline = "waits for the suspension of a request", +}; + +static int abort_f(BlockDriverState *bs, int argc, char **argv) +{ + abort(); +} + +static const cmdinfo_t abort_cmd = { + .name = "abort", + .cfunc = abort_f, + .flags = CMD_NOFILE_OK, + .oneline = "simulate a program crash using abort(3)", +}; + +static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) +{ + if (ct->flags & CMD_FLAG_GLOBAL) { + return 1; + } + if (!(ct->flags & CMD_NOFILE_OK) && !bs) { + fprintf(stderr, "no file open, try 'help open'\n"); + return 0; + } + return 1; +} + +static void __attribute((constructor)) init_qemuio_commands(void) +{ + /* initialize commands */ + help_init(); + add_command(&read_cmd); + add_command(&readv_cmd); + add_command(&write_cmd); + add_command(&writev_cmd); + add_command(&multiwrite_cmd); + add_command(&aio_read_cmd); + add_command(&aio_write_cmd); + add_command(&aio_flush_cmd); + add_command(&flush_cmd); + add_command(&truncate_cmd); + add_command(&length_cmd); + add_command(&info_cmd); + add_command(&discard_cmd); + add_command(&alloc_cmd); + add_command(&map_cmd); + add_command(&break_cmd); + add_command(&resume_cmd); + add_command(&wait_break_cmd); + add_command(&abort_cmd); + + add_check_command(init_check_command); +} diff --git a/qemu-io.c b/qemu-io.c index 3abf9ff..cfee840 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -27,1786 +27,7 @@ char *progname; BlockDriverState *qemuio_bs; -static int misalign; - -static int64_t cvtnum(const char *s) -{ - char *end; - return strtosz_suffix(s, &end, STRTOSZ_DEFSUFFIX_B); -} - -/* - * Parse the pattern argument to various sub-commands. - * - * Because the pattern is used as an argument to memset it must evaluate - * to an unsigned integer that fits into a single byte. - */ -static int parse_pattern(const char *arg) -{ - char *endptr = NULL; - long pattern; - - pattern = strtol(arg, &endptr, 0); - if (pattern < 0 || pattern > UCHAR_MAX || *endptr != '\0') { - printf("%s is not a valid pattern byte\n", arg); - return -1; - } - - return pattern; -} - -/* - * Memory allocation helpers. - * - * Make sure memory is aligned by default, or purposefully misaligned if - * that is specified on the command line. - */ - -#define MISALIGN_OFFSET 16 -static void *qemu_io_alloc(BlockDriverState *bs, size_t len, int pattern) -{ - void *buf; - - if (misalign) { - len += MISALIGN_OFFSET; - } - buf = qemu_blockalign(bs, len); - memset(buf, pattern, len); - if (misalign) { - buf += MISALIGN_OFFSET; - } - return buf; -} - -static void qemu_io_free(void *p) -{ - if (misalign) { - p -= MISALIGN_OFFSET; - } - qemu_vfree(p); -} - -static void dump_buffer(const void *buffer, int64_t offset, int len) -{ - int i, j; - const uint8_t *p; - - for (i = 0, p = buffer; i < len; i += 16) { - const uint8_t *s = p; - - printf("%08" PRIx64 ": ", offset + i); - for (j = 0; j < 16 && i + j < len; j++, p++) { - printf("%02x ", *p); - } - printf(" "); - for (j = 0; j < 16 && i + j < len; j++, s++) { - if (isalnum(*s)) { - printf("%c", *s); - } else { - printf("."); - } - } - printf("\n"); - } -} - -static void print_report(const char *op, struct timeval *t, int64_t offset, - int count, int total, int cnt, int Cflag) -{ - char s1[64], s2[64], ts[64]; - - timestr(t, ts, sizeof(ts), Cflag ? VERBOSE_FIXED_TIME : 0); - if (!Cflag) { - cvtstr((double)total, s1, sizeof(s1)); - cvtstr(tdiv((double)total, *t), s2, sizeof(s2)); - printf("%s %d/%d bytes at offset %" PRId64 "\n", - op, total, count, offset); - printf("%s, %d ops; %s (%s/sec and %.4f ops/sec)\n", - s1, cnt, ts, s2, tdiv((double)cnt, *t)); - } else {/* bytes,ops,time,bytes/sec,ops/sec */ - printf("%d,%d,%s,%.3f,%.3f\n", - total, cnt, ts, - tdiv((double)total, *t), - tdiv((double)cnt, *t)); - } -} - -/* - * Parse multiple length statements for vectored I/O, and construct an I/O - * vector matching it. - */ -static void * -create_iovec(BlockDriverState *bs, QEMUIOVector *qiov, char **argv, int nr_iov, - int pattern) -{ - size_t *sizes = g_new0(size_t, nr_iov); - size_t count = 0; - void *buf = NULL; - void *p; - int i; - - for (i = 0; i < nr_iov; i++) { - char *arg = argv[i]; - int64_t len; - - len = cvtnum(arg); - if (len < 0) { - printf("non-numeric length argument -- %s\n", arg); - goto fail; - } - - /* should be SIZE_T_MAX, but that doesn't exist */ - if (len > INT_MAX) { - printf("too large length argument -- %s\n", arg); - goto fail; - } - - if (len & 0x1ff) { - printf("length argument %" PRId64 - " is not sector aligned\n", len); - goto fail; - } - - sizes[i] = len; - count += len; - } - - qemu_iovec_init(qiov, nr_iov); - - buf = p = qemu_io_alloc(bs, count, pattern); - - for (i = 0; i < nr_iov; i++) { - qemu_iovec_add(qiov, p, sizes[i]); - p += sizes[i]; - } - -fail: - g_free(sizes); - return buf; -} - -static int do_read(BlockDriverState *bs, char *buf, int64_t offset, int count, - int *total) -{ - int ret; - - ret = bdrv_read(bs, offset >> 9, (uint8_t *)buf, count >> 9); - if (ret < 0) { - return ret; - } - *total = count; - return 1; -} - -static int do_write(BlockDriverState *bs, char *buf, int64_t offset, int count, - int *total) -{ - int ret; - - ret = bdrv_write(bs, offset >> 9, (uint8_t *)buf, count >> 9); - if (ret < 0) { - return ret; - } - *total = count; - return 1; -} - -static int do_pread(BlockDriverState *bs, char *buf, int64_t offset, int count, - int *total) -{ - *total = bdrv_pread(bs, offset, (uint8_t *)buf, count); - if (*total < 0) { - return *total; - } - return 1; -} - -static int do_pwrite(BlockDriverState *bs, char *buf, int64_t offset, int count, - int *total) -{ - *total = bdrv_pwrite(bs, offset, (uint8_t *)buf, count); - if (*total < 0) { - return *total; - } - return 1; -} - -typedef struct { - BlockDriverState *bs; - int64_t offset; - int count; - int *total; - int ret; - bool done; -} CoWriteZeroes; - -static void coroutine_fn co_write_zeroes_entry(void *opaque) -{ - CoWriteZeroes *data = opaque; - - data->ret = bdrv_co_write_zeroes(data->bs, data->offset / BDRV_SECTOR_SIZE, - data->count / BDRV_SECTOR_SIZE); - data->done = true; - if (data->ret < 0) { - *data->total = data->ret; - return; - } - - *data->total = data->count; -} - -static int do_co_write_zeroes(BlockDriverState *bs, int64_t offset, int count, - int *total) -{ - Coroutine *co; - CoWriteZeroes data = { - .bs = bs, - .offset = offset, - .count = count, - .total = total, - .done = false, - }; - - co = qemu_coroutine_create(co_write_zeroes_entry); - qemu_coroutine_enter(co, &data); - while (!data.done) { - qemu_aio_wait(); - } - if (data.ret < 0) { - return data.ret; - } else { - return 1; - } -} - -static int do_write_compressed(BlockDriverState *bs, char *buf, int64_t offset, - int count, int *total) -{ - int ret; - - ret = bdrv_write_compressed(bs, offset >> 9, (uint8_t *)buf, count >> 9); - if (ret < 0) { - return ret; - } - *total = count; - return 1; -} - -static int do_load_vmstate(BlockDriverState *bs, char *buf, int64_t offset, - int count, int *total) -{ - *total = bdrv_load_vmstate(bs, (uint8_t *)buf, offset, count); - if (*total < 0) { - return *total; - } - return 1; -} - -static int do_save_vmstate(BlockDriverState *bs, char *buf, int64_t offset, - int count, int *total) -{ - *total = bdrv_save_vmstate(bs, (uint8_t *)buf, offset, count); - if (*total < 0) { - return *total; - } - return 1; -} - -#define NOT_DONE 0x7fffffff -static void aio_rw_done(void *opaque, int ret) -{ - *(int *)opaque = ret; -} - -static int do_aio_readv(BlockDriverState *bs, QEMUIOVector *qiov, - int64_t offset, int *total) -{ - int async_ret = NOT_DONE; - - bdrv_aio_readv(bs, offset >> 9, qiov, qiov->size >> 9, - aio_rw_done, &async_ret); - while (async_ret == NOT_DONE) { - main_loop_wait(false); - } - - *total = qiov->size; - return async_ret < 0 ? async_ret : 1; -} - -static int do_aio_writev(BlockDriverState *bs, QEMUIOVector *qiov, - int64_t offset, int *total) -{ - int async_ret = NOT_DONE; - - bdrv_aio_writev(bs, offset >> 9, qiov, qiov->size >> 9, - aio_rw_done, &async_ret); - while (async_ret == NOT_DONE) { - main_loop_wait(false); - } - - *total = qiov->size; - return async_ret < 0 ? async_ret : 1; -} - -struct multiwrite_async_ret { - int num_done; - int error; -}; - -static void multiwrite_cb(void *opaque, int ret) -{ - struct multiwrite_async_ret *async_ret = opaque; - - async_ret->num_done++; - if (ret < 0) { - async_ret->error = ret; - } -} - -static int do_aio_multiwrite(BlockDriverState *bs, BlockRequest* reqs, - int num_reqs, int *total) -{ - int i, ret; - struct multiwrite_async_ret async_ret = { - .num_done = 0, - .error = 0, - }; - - *total = 0; - for (i = 0; i < num_reqs; i++) { - reqs[i].cb = multiwrite_cb; - reqs[i].opaque = &async_ret; - *total += reqs[i].qiov->size; - } - - ret = bdrv_aio_multiwrite(bs, reqs, num_reqs); - if (ret < 0) { - return ret; - } - - while (async_ret.num_done < num_reqs) { - main_loop_wait(false); - } - - return async_ret.error < 0 ? async_ret.error : 1; -} - -static void read_help(void) -{ - printf( -"\n" -" reads a range of bytes from the given offset\n" -"\n" -" Example:\n" -" 'read -v 512 1k' - dumps 1 kilobyte read from 512 bytes into the file\n" -"\n" -" Reads a segment of the currently open file, optionally dumping it to the\n" -" standard output stream (with -v option) for subsequent inspection.\n" -" -b, -- read from the VM state rather than the virtual disk\n" -" -C, -- report statistics in a machine parsable format\n" -" -l, -- length for pattern verification (only with -P)\n" -" -p, -- use bdrv_pread to read the file\n" -" -P, -- use a pattern to verify read data\n" -" -q, -- quiet mode, do not show I/O statistics\n" -" -s, -- start offset for pattern verification (only with -P)\n" -" -v, -- dump buffer to standard output\n" -"\n"); -} - -static int read_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t read_cmd = { - .name = "read", - .altname = "r", - .cfunc = read_f, - .argmin = 2, - .argmax = -1, - .args = "[-abCpqv] [-P pattern [-s off] [-l len]] off len", - .oneline = "reads a number of bytes at a specified offset", - .help = read_help, -}; - -static int read_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, pflag = 0, qflag = 0, vflag = 0; - int Pflag = 0, sflag = 0, lflag = 0, bflag = 0; - int c, cnt; - char *buf; - int64_t offset; - int count; - /* Some compilers get confused and warn if this is not initialized. */ - int total = 0; - int pattern = 0, pattern_offset = 0, pattern_count = 0; - - while ((c = getopt(argc, argv, "bCl:pP:qs:v")) != EOF) { - switch (c) { - case 'b': - bflag = 1; - break; - case 'C': - Cflag = 1; - break; - case 'l': - lflag = 1; - pattern_count = cvtnum(optarg); - if (pattern_count < 0) { - printf("non-numeric length argument -- %s\n", optarg); - return 0; - } - break; - case 'p': - pflag = 1; - break; - case 'P': - Pflag = 1; - pattern = parse_pattern(optarg); - if (pattern < 0) { - return 0; - } - break; - case 'q': - qflag = 1; - break; - case 's': - sflag = 1; - pattern_offset = cvtnum(optarg); - if (pattern_offset < 0) { - printf("non-numeric length argument -- %s\n", optarg); - return 0; - } - break; - case 'v': - vflag = 1; - break; - default: - return command_usage(&read_cmd); - } - } - - if (optind != argc - 2) { - return command_usage(&read_cmd); - } - - if (bflag && pflag) { - printf("-b and -p cannot be specified at the same time\n"); - return 0; - } - - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - optind++; - count = cvtnum(argv[optind]); - if (count < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - if (!Pflag && (lflag || sflag)) { - return command_usage(&read_cmd); - } - - if (!lflag) { - pattern_count = count - pattern_offset; - } - - if ((pattern_count < 0) || (pattern_count + pattern_offset > count)) { - printf("pattern verification range exceeds end of read data\n"); - return 0; - } - - if (!pflag) { - if (offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - offset); - return 0; - } - if (count & 0x1ff) { - printf("count %d is not sector aligned\n", - count); - return 0; - } - } - - buf = qemu_io_alloc(bs, count, 0xab); - - gettimeofday(&t1, NULL); - if (pflag) { - cnt = do_pread(bs, buf, offset, count, &total); - } else if (bflag) { - cnt = do_load_vmstate(bs, buf, offset, count, &total); - } else { - cnt = do_read(bs, buf, offset, count, &total); - } - gettimeofday(&t2, NULL); - - if (cnt < 0) { - printf("read failed: %s\n", strerror(-cnt)); - goto out; - } - - if (Pflag) { - void *cmp_buf = g_malloc(pattern_count); - memset(cmp_buf, pattern, pattern_count); - if (memcmp(buf + pattern_offset, cmp_buf, pattern_count)) { - printf("Pattern verification failed at offset %" - PRId64 ", %d bytes\n", - offset + pattern_offset, pattern_count); - } - g_free(cmp_buf); - } - - if (qflag) { - goto out; - } - - if (vflag) { - dump_buffer(buf, offset, count); - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, t1); - print_report("read", &t2, offset, count, total, cnt, Cflag); - -out: - qemu_io_free(buf); - - return 0; -} - -static void readv_help(void) -{ - printf( -"\n" -" reads a range of bytes from the given offset into multiple buffers\n" -"\n" -" Example:\n" -" 'readv -v 512 1k 1k ' - dumps 2 kilobytes read from 512 bytes into the file\n" -"\n" -" Reads a segment of the currently open file, optionally dumping it to the\n" -" standard output stream (with -v option) for subsequent inspection.\n" -" Uses multiple iovec buffers if more than one byte range is specified.\n" -" -C, -- report statistics in a machine parsable format\n" -" -P, -- use a pattern to verify read data\n" -" -v, -- dump buffer to standard output\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int readv_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t readv_cmd = { - .name = "readv", - .cfunc = readv_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cqv] [-P pattern ] off len [len..]", - .oneline = "reads a number of bytes at a specified offset", - .help = readv_help, -}; - -static int readv_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, qflag = 0, vflag = 0; - int c, cnt; - char *buf; - int64_t offset; - /* Some compilers get confused and warn if this is not initialized. */ - int total = 0; - int nr_iov; - QEMUIOVector qiov; - int pattern = 0; - int Pflag = 0; - - while ((c = getopt(argc, argv, "CP:qv")) != EOF) { - switch (c) { - case 'C': - Cflag = 1; - break; - case 'P': - Pflag = 1; - pattern = parse_pattern(optarg); - if (pattern < 0) { - return 0; - } - break; - case 'q': - qflag = 1; - break; - case 'v': - vflag = 1; - break; - default: - return command_usage(&readv_cmd); - } - } - - if (optind > argc - 2) { - return command_usage(&readv_cmd); - } - - - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - optind++; - - if (offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - offset); - return 0; - } - - nr_iov = argc - optind; - buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, 0xab); - if (buf == NULL) { - return 0; - } - - gettimeofday(&t1, NULL); - cnt = do_aio_readv(bs, &qiov, offset, &total); - gettimeofday(&t2, NULL); - - if (cnt < 0) { - printf("readv failed: %s\n", strerror(-cnt)); - goto out; - } - - if (Pflag) { - void *cmp_buf = g_malloc(qiov.size); - memset(cmp_buf, pattern, qiov.size); - if (memcmp(buf, cmp_buf, qiov.size)) { - printf("Pattern verification failed at offset %" - PRId64 ", %zd bytes\n", offset, qiov.size); - } - g_free(cmp_buf); - } - - if (qflag) { - goto out; - } - - if (vflag) { - dump_buffer(buf, offset, qiov.size); - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, t1); - print_report("read", &t2, offset, qiov.size, total, cnt, Cflag); - -out: - qemu_iovec_destroy(&qiov); - qemu_io_free(buf); - return 0; -} - -static void write_help(void) -{ - printf( -"\n" -" writes a range of bytes from the given offset\n" -"\n" -" Example:\n" -" 'write 512 1k' - writes 1 kilobyte at 512 bytes into the open file\n" -"\n" -" Writes into a segment of the currently open file, using a buffer\n" -" filled with a set pattern (0xcdcdcdcd).\n" -" -b, -- write to the VM state rather than the virtual disk\n" -" -c, -- write compressed data with bdrv_write_compressed\n" -" -p, -- use bdrv_pwrite to write the file\n" -" -P, -- use different pattern to fill file\n" -" -C, -- report statistics in a machine parsable format\n" -" -q, -- quiet mode, do not show I/O statistics\n" -" -z, -- write zeroes using bdrv_co_write_zeroes\n" -"\n"); -} - -static int write_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t write_cmd = { - .name = "write", - .altname = "w", - .cfunc = write_f, - .argmin = 2, - .argmax = -1, - .args = "[-bcCpqz] [-P pattern ] off len", - .oneline = "writes a number of bytes at a specified offset", - .help = write_help, -}; - -static int write_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, pflag = 0, qflag = 0, bflag = 0, Pflag = 0, zflag = 0; - int cflag = 0; - int c, cnt; - char *buf = NULL; - int64_t offset; - int count; - /* Some compilers get confused and warn if this is not initialized. */ - int total = 0; - int pattern = 0xcd; - - while ((c = getopt(argc, argv, "bcCpP:qz")) != EOF) { - switch (c) { - case 'b': - bflag = 1; - break; - case 'c': - cflag = 1; - break; - case 'C': - Cflag = 1; - break; - case 'p': - pflag = 1; - break; - case 'P': - Pflag = 1; - pattern = parse_pattern(optarg); - if (pattern < 0) { - return 0; - } - break; - case 'q': - qflag = 1; - break; - case 'z': - zflag = 1; - break; - default: - return command_usage(&write_cmd); - } - } - - if (optind != argc - 2) { - return command_usage(&write_cmd); - } - - if (bflag + pflag + zflag > 1) { - printf("-b, -p, or -z cannot be specified at the same time\n"); - return 0; - } - - if (zflag && Pflag) { - printf("-z and -P cannot be specified at the same time\n"); - return 0; - } - - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - optind++; - count = cvtnum(argv[optind]); - if (count < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - if (!pflag) { - if (offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - offset); - return 0; - } - - if (count & 0x1ff) { - printf("count %d is not sector aligned\n", - count); - return 0; - } - } - - if (!zflag) { - buf = qemu_io_alloc(bs, count, pattern); - } - - gettimeofday(&t1, NULL); - if (pflag) { - cnt = do_pwrite(bs, buf, offset, count, &total); - } else if (bflag) { - cnt = do_save_vmstate(bs, buf, offset, count, &total); - } else if (zflag) { - cnt = do_co_write_zeroes(bs, offset, count, &total); - } else if (cflag) { - cnt = do_write_compressed(bs, buf, offset, count, &total); - } else { - cnt = do_write(bs, buf, offset, count, &total); - } - gettimeofday(&t2, NULL); - - if (cnt < 0) { - printf("write failed: %s\n", strerror(-cnt)); - goto out; - } - - if (qflag) { - goto out; - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, t1); - print_report("wrote", &t2, offset, count, total, cnt, Cflag); - -out: - if (!zflag) { - qemu_io_free(buf); - } - - return 0; -} - -static void -writev_help(void) -{ - printf( -"\n" -" writes a range of bytes from the given offset source from multiple buffers\n" -"\n" -" Example:\n" -" 'write 512 1k 1k' - writes 2 kilobytes at 512 bytes into the open file\n" -"\n" -" Writes into a segment of the currently open file, using a buffer\n" -" filled with a set pattern (0xcdcdcdcd).\n" -" -P, -- use different pattern to fill file\n" -" -C, -- report statistics in a machine parsable format\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int writev_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t writev_cmd = { - .name = "writev", - .cfunc = writev_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cq] [-P pattern ] off len [len..]", - .oneline = "writes a number of bytes at a specified offset", - .help = writev_help, -}; - -static int writev_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, qflag = 0; - int c, cnt; - char *buf; - int64_t offset; - /* Some compilers get confused and warn if this is not initialized. */ - int total = 0; - int nr_iov; - int pattern = 0xcd; - QEMUIOVector qiov; - - while ((c = getopt(argc, argv, "CqP:")) != EOF) { - switch (c) { - case 'C': - Cflag = 1; - break; - case 'q': - qflag = 1; - break; - case 'P': - pattern = parse_pattern(optarg); - if (pattern < 0) { - return 0; - } - break; - default: - return command_usage(&writev_cmd); - } - } - - if (optind > argc - 2) { - return command_usage(&writev_cmd); - } - - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - optind++; - - if (offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - offset); - return 0; - } - - nr_iov = argc - optind; - buf = create_iovec(bs, &qiov, &argv[optind], nr_iov, pattern); - if (buf == NULL) { - return 0; - } - - gettimeofday(&t1, NULL); - cnt = do_aio_writev(bs, &qiov, offset, &total); - gettimeofday(&t2, NULL); - - if (cnt < 0) { - printf("writev failed: %s\n", strerror(-cnt)); - goto out; - } - - if (qflag) { - goto out; - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, t1); - print_report("wrote", &t2, offset, qiov.size, total, cnt, Cflag); -out: - qemu_iovec_destroy(&qiov); - qemu_io_free(buf); - return 0; -} - -static void multiwrite_help(void) -{ - printf( -"\n" -" writes a range of bytes from the given offset source from multiple buffers,\n" -" in a batch of requests that may be merged by qemu\n" -"\n" -" Example:\n" -" 'multiwrite 512 1k 1k ; 4k 1k'\n" -" writes 2 kB at 512 bytes and 1 kB at 4 kB into the open file\n" -"\n" -" Writes into a segment of the currently open file, using a buffer\n" -" filled with a set pattern (0xcdcdcdcd). The pattern byte is increased\n" -" by one for each request contained in the multiwrite command.\n" -" -P, -- use different pattern to fill file\n" -" -C, -- report statistics in a machine parsable format\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int multiwrite_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t multiwrite_cmd = { - .name = "multiwrite", - .cfunc = multiwrite_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cq] [-P pattern ] off len [len..] [; off len [len..]..]", - .oneline = "issues multiple write requests at once", - .help = multiwrite_help, -}; - -static int multiwrite_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, qflag = 0; - int c, cnt; - char **buf; - int64_t offset, first_offset = 0; - /* Some compilers get confused and warn if this is not initialized. */ - int total = 0; - int nr_iov; - int nr_reqs; - int pattern = 0xcd; - QEMUIOVector *qiovs; - int i; - BlockRequest *reqs; - - while ((c = getopt(argc, argv, "CqP:")) != EOF) { - switch (c) { - case 'C': - Cflag = 1; - break; - case 'q': - qflag = 1; - break; - case 'P': - pattern = parse_pattern(optarg); - if (pattern < 0) { - return 0; - } - break; - default: - return command_usage(&writev_cmd); - } - } - - if (optind > argc - 2) { - return command_usage(&writev_cmd); - } - - nr_reqs = 1; - for (i = optind; i < argc; i++) { - if (!strcmp(argv[i], ";")) { - nr_reqs++; - } - } - - reqs = g_malloc0(nr_reqs * sizeof(*reqs)); - buf = g_malloc0(nr_reqs * sizeof(*buf)); - qiovs = g_malloc(nr_reqs * sizeof(*qiovs)); - - for (i = 0; i < nr_reqs && optind < argc; i++) { - int j; - - /* Read the offset of the request */ - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric offset argument -- %s\n", argv[optind]); - goto out; - } - optind++; - - if (offset & 0x1ff) { - printf("offset %lld is not sector aligned\n", - (long long)offset); - goto out; - } - - if (i == 0) { - first_offset = offset; - } - - /* Read lengths for qiov entries */ - for (j = optind; j < argc; j++) { - if (!strcmp(argv[j], ";")) { - break; - } - } - - nr_iov = j - optind; - - /* Build request */ - buf[i] = create_iovec(bs, &qiovs[i], &argv[optind], nr_iov, pattern); - if (buf[i] == NULL) { - goto out; - } - - reqs[i].qiov = &qiovs[i]; - reqs[i].sector = offset >> 9; - reqs[i].nb_sectors = reqs[i].qiov->size >> 9; - - optind = j + 1; - - pattern++; - } - - /* If there were empty requests at the end, ignore them */ - nr_reqs = i; - - gettimeofday(&t1, NULL); - cnt = do_aio_multiwrite(bs, reqs, nr_reqs, &total); - gettimeofday(&t2, NULL); - - if (cnt < 0) { - printf("aio_multiwrite failed: %s\n", strerror(-cnt)); - goto out; - } - - if (qflag) { - goto out; - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, t1); - print_report("wrote", &t2, first_offset, total, total, cnt, Cflag); -out: - for (i = 0; i < nr_reqs; i++) { - qemu_io_free(buf[i]); - if (reqs[i].qiov != NULL) { - qemu_iovec_destroy(&qiovs[i]); - } - } - g_free(buf); - g_free(reqs); - g_free(qiovs); - return 0; -} - -struct aio_ctx { - QEMUIOVector qiov; - int64_t offset; - char *buf; - int qflag; - int vflag; - int Cflag; - int Pflag; - int pattern; - struct timeval t1; -}; - -static void aio_write_done(void *opaque, int ret) -{ - struct aio_ctx *ctx = opaque; - struct timeval t2; - - gettimeofday(&t2, NULL); - - - if (ret < 0) { - printf("aio_write failed: %s\n", strerror(-ret)); - goto out; - } - - if (ctx->qflag) { - goto out; - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, ctx->t1); - print_report("wrote", &t2, ctx->offset, ctx->qiov.size, - ctx->qiov.size, 1, ctx->Cflag); -out: - qemu_io_free(ctx->buf); - qemu_iovec_destroy(&ctx->qiov); - g_free(ctx); -} - -static void aio_read_done(void *opaque, int ret) -{ - struct aio_ctx *ctx = opaque; - struct timeval t2; - - gettimeofday(&t2, NULL); - - if (ret < 0) { - printf("readv failed: %s\n", strerror(-ret)); - goto out; - } - - if (ctx->Pflag) { - void *cmp_buf = g_malloc(ctx->qiov.size); - - memset(cmp_buf, ctx->pattern, ctx->qiov.size); - if (memcmp(ctx->buf, cmp_buf, ctx->qiov.size)) { - printf("Pattern verification failed at offset %" - PRId64 ", %zd bytes\n", ctx->offset, ctx->qiov.size); - } - g_free(cmp_buf); - } - - if (ctx->qflag) { - goto out; - } - - if (ctx->vflag) { - dump_buffer(ctx->buf, ctx->offset, ctx->qiov.size); - } - - /* Finally, report back -- -C gives a parsable format */ - t2 = tsub(t2, ctx->t1); - print_report("read", &t2, ctx->offset, ctx->qiov.size, - ctx->qiov.size, 1, ctx->Cflag); -out: - qemu_io_free(ctx->buf); - qemu_iovec_destroy(&ctx->qiov); - g_free(ctx); -} - -static void aio_read_help(void) -{ - printf( -"\n" -" asynchronously reads a range of bytes from the given offset\n" -"\n" -" Example:\n" -" 'aio_read -v 512 1k 1k ' - dumps 2 kilobytes read from 512 bytes into the file\n" -"\n" -" Reads a segment of the currently open file, optionally dumping it to the\n" -" standard output stream (with -v option) for subsequent inspection.\n" -" The read is performed asynchronously and the aio_flush command must be\n" -" used to ensure all outstanding aio requests have been completed.\n" -" -C, -- report statistics in a machine parsable format\n" -" -P, -- use a pattern to verify read data\n" -" -v, -- dump buffer to standard output\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int aio_read_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t aio_read_cmd = { - .name = "aio_read", - .cfunc = aio_read_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cqv] [-P pattern ] off len [len..]", - .oneline = "asynchronously reads a number of bytes", - .help = aio_read_help, -}; - -static int aio_read_f(BlockDriverState *bs, int argc, char **argv) -{ - int nr_iov, c; - struct aio_ctx *ctx = g_new0(struct aio_ctx, 1); - - while ((c = getopt(argc, argv, "CP:qv")) != EOF) { - switch (c) { - case 'C': - ctx->Cflag = 1; - break; - case 'P': - ctx->Pflag = 1; - ctx->pattern = parse_pattern(optarg); - if (ctx->pattern < 0) { - g_free(ctx); - return 0; - } - break; - case 'q': - ctx->qflag = 1; - break; - case 'v': - ctx->vflag = 1; - break; - default: - g_free(ctx); - return command_usage(&aio_read_cmd); - } - } - - if (optind > argc - 2) { - g_free(ctx); - return command_usage(&aio_read_cmd); - } - - ctx->offset = cvtnum(argv[optind]); - if (ctx->offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - g_free(ctx); - return 0; - } - optind++; - - if (ctx->offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - ctx->offset); - g_free(ctx); - return 0; - } - - nr_iov = argc - optind; - ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, 0xab); - if (ctx->buf == NULL) { - g_free(ctx); - return 0; - } - - gettimeofday(&ctx->t1, NULL); - bdrv_aio_readv(bs, ctx->offset >> 9, &ctx->qiov, - ctx->qiov.size >> 9, aio_read_done, ctx); - return 0; -} - -static void aio_write_help(void) -{ - printf( -"\n" -" asynchronously writes a range of bytes from the given offset source\n" -" from multiple buffers\n" -"\n" -" Example:\n" -" 'aio_write 512 1k 1k' - writes 2 kilobytes at 512 bytes into the open file\n" -"\n" -" Writes into a segment of the currently open file, using a buffer\n" -" filled with a set pattern (0xcdcdcdcd).\n" -" The write is performed asynchronously and the aio_flush command must be\n" -" used to ensure all outstanding aio requests have been completed.\n" -" -P, -- use different pattern to fill file\n" -" -C, -- report statistics in a machine parsable format\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int aio_write_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t aio_write_cmd = { - .name = "aio_write", - .cfunc = aio_write_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cq] [-P pattern ] off len [len..]", - .oneline = "asynchronously writes a number of bytes", - .help = aio_write_help, -}; - -static int aio_write_f(BlockDriverState *bs, int argc, char **argv) -{ - int nr_iov, c; - int pattern = 0xcd; - struct aio_ctx *ctx = g_new0(struct aio_ctx, 1); - - while ((c = getopt(argc, argv, "CqP:")) != EOF) { - switch (c) { - case 'C': - ctx->Cflag = 1; - break; - case 'q': - ctx->qflag = 1; - break; - case 'P': - pattern = parse_pattern(optarg); - if (pattern < 0) { - g_free(ctx); - return 0; - } - break; - default: - g_free(ctx); - return command_usage(&aio_write_cmd); - } - } - - if (optind > argc - 2) { - g_free(ctx); - return command_usage(&aio_write_cmd); - } - - ctx->offset = cvtnum(argv[optind]); - if (ctx->offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - g_free(ctx); - return 0; - } - optind++; - - if (ctx->offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - ctx->offset); - g_free(ctx); - return 0; - } - - nr_iov = argc - optind; - ctx->buf = create_iovec(bs, &ctx->qiov, &argv[optind], nr_iov, pattern); - if (ctx->buf == NULL) { - g_free(ctx); - return 0; - } - - gettimeofday(&ctx->t1, NULL); - bdrv_aio_writev(bs, ctx->offset >> 9, &ctx->qiov, - ctx->qiov.size >> 9, aio_write_done, ctx); - return 0; -} - -static int aio_flush_f(BlockDriverState *bs, int argc, char **argv) -{ - bdrv_drain_all(); - return 0; -} - -static const cmdinfo_t aio_flush_cmd = { - .name = "aio_flush", - .cfunc = aio_flush_f, - .oneline = "completes all outstanding aio requests" -}; - -static int flush_f(BlockDriverState *bs, int argc, char **argv) -{ - bdrv_flush(bs); - return 0; -} - -static const cmdinfo_t flush_cmd = { - .name = "flush", - .altname = "f", - .cfunc = flush_f, - .oneline = "flush all in-core file state to disk", -}; - -static int truncate_f(BlockDriverState *bs, int argc, char **argv) -{ - int64_t offset; - int ret; - - offset = cvtnum(argv[1]); - if (offset < 0) { - printf("non-numeric truncate argument -- %s\n", argv[1]); - return 0; - } - - ret = bdrv_truncate(bs, offset); - if (ret < 0) { - printf("truncate: %s\n", strerror(-ret)); - return 0; - } - - return 0; -} - -static const cmdinfo_t truncate_cmd = { - .name = "truncate", - .altname = "t", - .cfunc = truncate_f, - .argmin = 1, - .argmax = 1, - .args = "off", - .oneline = "truncates the current file at the given offset", -}; - -static int length_f(BlockDriverState *bs, int argc, char **argv) -{ - int64_t size; - char s1[64]; - - size = bdrv_getlength(bs); - if (size < 0) { - printf("getlength: %s\n", strerror(-size)); - return 0; - } - - cvtstr(size, s1, sizeof(s1)); - printf("%s\n", s1); - return 0; -} - - -static const cmdinfo_t length_cmd = { - .name = "length", - .altname = "l", - .cfunc = length_f, - .oneline = "gets the length of the current file", -}; - - -static int info_f(BlockDriverState *bs, int argc, char **argv) -{ - BlockDriverInfo bdi; - char s1[64], s2[64]; - int ret; - - if (bs->drv && bs->drv->format_name) { - printf("format name: %s\n", bs->drv->format_name); - } - if (bs->drv && bs->drv->protocol_name) { - printf("format name: %s\n", bs->drv->protocol_name); - } - - ret = bdrv_get_info(bs, &bdi); - if (ret) { - return 0; - } - - cvtstr(bdi.cluster_size, s1, sizeof(s1)); - cvtstr(bdi.vm_state_offset, s2, sizeof(s2)); - - printf("cluster size: %s\n", s1); - printf("vm state offset: %s\n", s2); - - return 0; -} - - - -static const cmdinfo_t info_cmd = { - .name = "info", - .altname = "i", - .cfunc = info_f, - .oneline = "prints information about the current file", -}; - -static void discard_help(void) -{ - printf( -"\n" -" discards a range of bytes from the given offset\n" -"\n" -" Example:\n" -" 'discard 512 1k' - discards 1 kilobyte from 512 bytes into the file\n" -"\n" -" Discards a segment of the currently open file.\n" -" -C, -- report statistics in a machine parsable format\n" -" -q, -- quiet mode, do not show I/O statistics\n" -"\n"); -} - -static int discard_f(BlockDriverState *bs, int argc, char **argv); - -static const cmdinfo_t discard_cmd = { - .name = "discard", - .altname = "d", - .cfunc = discard_f, - .argmin = 2, - .argmax = -1, - .args = "[-Cq] off len", - .oneline = "discards a number of bytes at a specified offset", - .help = discard_help, -}; - -static int discard_f(BlockDriverState *bs, int argc, char **argv) -{ - struct timeval t1, t2; - int Cflag = 0, qflag = 0; - int c, ret; - int64_t offset; - int count; - - while ((c = getopt(argc, argv, "Cq")) != EOF) { - switch (c) { - case 'C': - Cflag = 1; - break; - case 'q': - qflag = 1; - break; - default: - return command_usage(&discard_cmd); - } - } - - if (optind != argc - 2) { - return command_usage(&discard_cmd); - } - - offset = cvtnum(argv[optind]); - if (offset < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - optind++; - count = cvtnum(argv[optind]); - if (count < 0) { - printf("non-numeric length argument -- %s\n", argv[optind]); - return 0; - } - - gettimeofday(&t1, NULL); - ret = bdrv_discard(bs, offset >> BDRV_SECTOR_BITS, - count >> BDRV_SECTOR_BITS); - gettimeofday(&t2, NULL); - - if (ret < 0) { - printf("discard failed: %s\n", strerror(-ret)); - goto out; - } - - /* Finally, report back -- -C gives a parsable format */ - if (!qflag) { - t2 = tsub(t2, t1); - print_report("discard", &t2, offset, count, count, 1, Cflag); - } - -out: - return 0; -} - -static int alloc_f(BlockDriverState *bs, int argc, char **argv) -{ - int64_t offset, sector_num; - int nb_sectors, remaining; - char s1[64]; - int num, sum_alloc; - int ret; - - offset = cvtnum(argv[1]); - if (offset < 0) { - printf("non-numeric offset argument -- %s\n", argv[1]); - return 0; - } else if (offset & 0x1ff) { - printf("offset %" PRId64 " is not sector aligned\n", - offset); - return 0; - } - - if (argc == 3) { - nb_sectors = cvtnum(argv[2]); - if (nb_sectors < 0) { - printf("non-numeric length argument -- %s\n", argv[2]); - return 0; - } - } else { - nb_sectors = 1; - } - - remaining = nb_sectors; - sum_alloc = 0; - sector_num = offset >> 9; - while (remaining) { - ret = bdrv_is_allocated(bs, sector_num, remaining, &num); - sector_num += num; - remaining -= num; - if (ret) { - sum_alloc += num; - } - if (num == 0) { - nb_sectors -= remaining; - remaining = 0; - } - } - - cvtstr(offset, s1, sizeof(s1)); - - printf("%d/%d sectors allocated at offset %s\n", - sum_alloc, nb_sectors, s1); - return 0; -} - -static const cmdinfo_t alloc_cmd = { - .name = "alloc", - .altname = "a", - .argmin = 1, - .argmax = 2, - .cfunc = alloc_f, - .args = "off [sectors]", - .oneline = "checks if a sector is present in the file", -}; - - -static int map_is_allocated(BlockDriverState *bs, int64_t sector_num, - int64_t nb_sectors, int64_t *pnum) -{ - int num, num_checked; - int ret, firstret; - - num_checked = MIN(nb_sectors, INT_MAX); - ret = bdrv_is_allocated(bs, sector_num, num_checked, &num); - if (ret < 0) { - return ret; - } - - firstret = ret; - *pnum = num; - - while (nb_sectors > 0 && ret == firstret) { - sector_num += num; - nb_sectors -= num; - - num_checked = MIN(nb_sectors, INT_MAX); - ret = bdrv_is_allocated(bs, sector_num, num_checked, &num); - if (ret == firstret) { - *pnum += num; - } else { - break; - } - } - - return firstret; -} - -static int map_f(BlockDriverState *bs, int argc, char **argv) -{ - int64_t offset; - int64_t nb_sectors; - char s1[64]; - int64_t num; - int ret; - const char *retstr; - - offset = 0; - nb_sectors = bs->total_sectors; - - do { - ret = map_is_allocated(bs, offset, nb_sectors, &num); - if (ret < 0) { - error_report("Failed to get allocation status: %s", strerror(-ret)); - return 0; - } - - retstr = ret ? " allocated" : "not allocated"; - cvtstr(offset << 9ULL, s1, sizeof(s1)); - printf("[% 24" PRId64 "] % 8" PRId64 "/% 8" PRId64 " sectors %s " - "at offset %s (%d)\n", - offset << 9ULL, num, nb_sectors, retstr, s1, ret); - - offset += num; - nb_sectors -= num; - } while (offset < bs->total_sectors); - - return 0; -} - -static const cmdinfo_t map_cmd = { - .name = "map", - .argmin = 0, - .argmax = 0, - .cfunc = map_f, - .args = "", - .oneline = "prints the allocated areas of a file", -}; - -static int break_f(BlockDriverState *bs, int argc, char **argv) -{ - int ret; - - ret = bdrv_debug_breakpoint(bs, argv[1], argv[2]); - if (ret < 0) { - printf("Could not set breakpoint: %s\n", strerror(-ret)); - } - - return 0; -} - -static const cmdinfo_t break_cmd = { - .name = "break", - .argmin = 2, - .argmax = 2, - .cfunc = break_f, - .args = "event tag", - .oneline = "sets a breakpoint on event and tags the stopped " - "request as tag", -}; - -static int resume_f(BlockDriverState *bs, int argc, char **argv) -{ - int ret; - - ret = bdrv_debug_resume(bs, argv[1]); - if (ret < 0) { - printf("Could not resume request: %s\n", strerror(-ret)); - } - - return 0; -} - -static const cmdinfo_t resume_cmd = { - .name = "resume", - .argmin = 1, - .argmax = 1, - .cfunc = resume_f, - .args = "tag", - .oneline = "resumes the request tagged as tag", -}; - -static int wait_break_f(BlockDriverState *bs, int argc, char **argv) -{ - while (!bdrv_debug_is_suspended(bs, argv[1])) { - qemu_aio_wait(); - } - - return 0; -} - -static const cmdinfo_t wait_break_cmd = { - .name = "wait_break", - .argmin = 1, - .argmax = 1, - .cfunc = wait_break_f, - .args = "tag", - .oneline = "waits for the suspension of a request", -}; - -static int abort_f(BlockDriverState *bs, int argc, char **argv) -{ - abort(); -} - -static const cmdinfo_t abort_cmd = { - .name = "abort", - .cfunc = abort_f, - .flags = CMD_NOFILE_OK, - .oneline = "simulate a program crash using abort(3)", -}; +extern int qemuio_misalign; static int close_f(BlockDriverState *bs, int argc, char **argv) { @@ -1916,18 +137,6 @@ static int open_f(BlockDriverState *bs, int argc, char **argv) return openfile(argv[optind], flags, growable); } -static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) -{ - if (ct->flags & CMD_FLAG_GLOBAL) { - return 1; - } - if (!(ct->flags & CMD_NOFILE_OK) && !bs) { - fprintf(stderr, "no file open, try 'help open'\n"); - return 0; - } - return 1; -} - static void usage(const char *name) { printf( @@ -1998,7 +207,7 @@ int main(int argc, char **argv) readonly = 1; break; case 'm': - misalign = 1; + qemuio_misalign = 1; break; case 'g': growable = 1; @@ -2039,30 +248,8 @@ int main(int argc, char **argv) /* initialize commands */ quit_init(); - help_init(); add_command(&open_cmd); add_command(&close_cmd); - add_command(&read_cmd); - add_command(&readv_cmd); - add_command(&write_cmd); - add_command(&writev_cmd); - add_command(&multiwrite_cmd); - add_command(&aio_read_cmd); - add_command(&aio_write_cmd); - add_command(&aio_flush_cmd); - add_command(&flush_cmd); - add_command(&truncate_cmd); - add_command(&length_cmd); - add_command(&info_cmd); - add_command(&discard_cmd); - add_command(&alloc_cmd); - add_command(&map_cmd); - add_command(&break_cmd); - add_command(&resume_cmd); - add_command(&wait_break_cmd); - add_command(&abort_cmd); - - add_check_command(init_check_command); /* open the device */ if (!readonly) { -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c 2013-05-28 15:27 ` [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c Kevin Wolf @ 2013-05-29 20:29 ` Eric Blake 2013-06-04 10:11 ` Kevin Wolf 0 siblings, 1 reply; 45+ messages in thread From: Eric Blake @ 2013-05-29 20:29 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 1304 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > This is the implementation of all qemu-io commands that make sense to be > called from the qemu monitor, i.e. everything except open, close and > quit. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > Makefile | 2 +- > qemu-io-cmds.c | 1835 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > qemu-io.c | 1817 +------------------------------------------------------ > 3 files changed, 1838 insertions(+), 1816 deletions(-) > create mode 100644 qemu-io-cmds.c I checked the bulk of this patch via: $ diff -u <(sed -n '/^-/ s///p' file) <(sed -n '/^\+/ s///p' file) The bulk of the patch is blind code motion plus a bit of touchup; I did however spot one case where you were too eager in your search-and-replace: * Memory allocation helpers. * - * Make sure memory is aligned by default, or purposefully misaligned if + * Make sure memory is aligned by default, or purposefully qemuio_misaligned if * that is specified on the command line. > + > +int qemuio_misalign; Should this variable be typed 'bool'? Other than that, 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c 2013-05-29 20:29 ` Eric Blake @ 2013-06-04 10:11 ` Kevin Wolf 0 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-06-04 10:11 UTC (permalink / raw) To: Eric Blake; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino Am 29.05.2013 um 22:29 hat Eric Blake geschrieben: > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > This is the implementation of all qemu-io commands that make sense to be > > called from the qemu monitor, i.e. everything except open, close and > > quit. > > > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > > --- > > Makefile | 2 +- > > qemu-io-cmds.c | 1835 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > qemu-io.c | 1817 +------------------------------------------------------ > > 3 files changed, 1838 insertions(+), 1816 deletions(-) > > create mode 100644 qemu-io-cmds.c > > I checked the bulk of this patch via: > > $ diff -u <(sed -n '/^-/ s///p' file) <(sed -n '/^\+/ s///p' file) > > The bulk of the patch is blind code motion plus a bit of touchup; I did > however spot one case where you were too eager in your search-and-replace: > > * Memory allocation helpers. > * > - * Make sure memory is aligned by default, or purposefully misaligned if > + * Make sure memory is aligned by default, or purposefully > qemuio_misaligned if > * that is specified on the command line. Whoops, thanks for spotting this. :-) > > + > > +int qemuio_misalign; > > Should this variable be typed 'bool'? It should, but this is merely code motion, so I won't fix it here. Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (5 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-29 21:11 ` Eric Blake 2013-06-05 12:33 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function Kevin Wolf ` (9 subsequent siblings) 16 siblings, 2 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino It's duplicated code. Move it to qemu-io-cmds.c because it's not dependent on any static data of the qemu-io tool. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 43 +++++-------------------------------------- cmd.h | 3 ++- qemu-io-cmds.c | 24 ++++++++++++++++++++++++ 3 files changed, 31 insertions(+), 39 deletions(-) diff --git a/cmd.c b/cmd.c index d501aab..7ae978f 100644 --- a/cmd.c +++ b/cmd.c @@ -138,28 +138,11 @@ static char *get_prompt(void); void command_loop(void) { - int c, i, done = 0, fetchable = 0, prompted = 0; + int i, done = 0, fetchable = 0, prompted = 0; char *input; - char **v; - const cmdinfo_t *ct; for (i = 0; !done && i < ncmdline; i++) { - input = strdup(cmdline[i]); - if (!input) { - fprintf(stderr, _("cannot strdup command '%s': %s\n"), - cmdline[i], strerror(errno)); - exit(1); - } - v = breakline(input, &c); - if (c) { - ct = find_command(v[0]); - if (ct) { - done = command(ct, c, v); - } else { - fprintf(stderr, _("command \"%s\" not found\n"), v[0]); - } - } - doneline(input, v); + done = qemuio_command(cmdline[i]); } if (cmdline) { g_free(cmdline); @@ -179,20 +162,13 @@ void command_loop(void) if (!fetchable) { continue; } + input = fetchline(); if (input == NULL) { break; } - v = breakline(input, &c); - if (c) { - ct = find_command(v[0]); - if (ct) { - done = command(ct, c, v); - } else { - fprintf(stderr, _("command \"%s\" not found\n"), v[0]); - } - } - doneline(input, v); + done = qemuio_command(input); + free(input); prompted = 0; fetchable = 0; @@ -328,15 +304,6 @@ char **breakline(char *input, int *count) return rval; } -void -doneline( - char *input, - char **vec) -{ - free(input); - free(vec); -} - #define EXABYTES(x) ((long long)(x) << 60) #define PETABYTES(x) ((long long)(x) << 50) #define TERABYTES(x) ((long long)(x) << 40) diff --git a/cmd.h b/cmd.h index ccf6336..d676408 100644 --- a/cmd.h +++ b/cmd.h @@ -59,7 +59,6 @@ int command(const cmdinfo_t *ci, int argc, char **argv); /* from input.h */ char **breakline(char *input, int *count); -void doneline(char *input, char **vec); char *fetchline(void); void cvtstr(double value, char *str, size_t sz); @@ -77,4 +76,6 @@ void timestr(struct timeval *tv, char *str, size_t sz, int flags); extern char *progname; +bool qemuio_command(const char *cmd); + #endif /* __COMMAND_H__ */ diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index 57b751e..bdfe5f2 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -1807,6 +1807,30 @@ static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) return 1; } +bool qemuio_command(const char *cmd) +{ + char *input; + const cmdinfo_t *ct; + char **v; + int c; + bool done = false; + + input = g_strdup(cmd); + v = breakline(input, &c); + if (c) { + ct = find_command(v[0]); + if (ct) { + done = command(ct, c, v); + } else { + fprintf(stderr, "command \"%s\" not found\n", v[0]); + } + } + g_free(input); + g_free(v); + + return done; +} + static void __attribute((constructor)) init_qemuio_commands(void) { /* initialize commands */ -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command 2013-05-28 15:27 ` [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command Kevin Wolf @ 2013-05-29 21:11 ` Eric Blake 2013-06-05 12:33 ` Stefan Hajnoczi 1 sibling, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-29 21:11 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 585 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > It's duplicated code. Move it to qemu-io-cmds.c because it's not > dependent on any static data of the qemu-io tool. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 43 +++++-------------------------------------- > cmd.h | 3 ++- > qemu-io-cmds.c | 24 ++++++++++++++++++++++++ > 3 files changed, 31 insertions(+), 39 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command 2013-05-28 15:27 ` [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command Kevin Wolf 2013-05-29 21:11 ` Eric Blake @ 2013-06-05 12:33 ` Stefan Hajnoczi 1 sibling, 0 replies; 45+ messages in thread From: Stefan Hajnoczi @ 2013-06-05 12:33 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino On Tue, May 28, 2013 at 05:27:27PM +0200, Kevin Wolf wrote: > diff --git a/cmd.h b/cmd.h > index ccf6336..d676408 100644 > --- a/cmd.h > +++ b/cmd.h > @@ -59,7 +59,6 @@ int command(const cmdinfo_t *ci, int argc, char **argv); > > /* from input.h */ > char **breakline(char *input, int *count); > -void doneline(char *input, char **vec); > char *fetchline(void); > > void cvtstr(double value, char *str, size_t sz); > @@ -77,4 +76,6 @@ void timestr(struct timeval *tv, char *str, size_t sz, int flags); > > extern char *progname; > > +bool qemuio_command(const char *cmd); If we keep cmd.h generic then qemuio_command() should go into another header file. ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (6 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-29 21:25 ` Eric Blake 2013-06-05 12:37 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function Kevin Wolf ` (8 subsequent siblings) 16 siblings, 2 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino No reason to treat it different from other commands. Move it to qemu-io-cmds.c, adapt the coding style and register it like any other command. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 79 ---------------------------------------------------------- cmd.h | 1 - qemu-io-cmds.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++- 3 files changed, 66 insertions(+), 81 deletions(-) diff --git a/cmd.c b/cmd.c index 7ae978f..2941ad3 100644 --- a/cmd.c +++ b/cmd.c @@ -439,82 +439,3 @@ quit_init(void) add_command(&quit_cmd); } - -/* from libxcmd/help.c */ - -static cmdinfo_t help_cmd; -static void help_onecmd(const char *cmd, const cmdinfo_t *ct); -static void help_oneline(const char *cmd, const cmdinfo_t *ct); - -static void -help_all(void) -{ - const cmdinfo_t *ct; - - for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) - help_oneline(ct->name, ct); - printf(_("\nUse 'help commandname' for extended help.\n")); -} - -static int -help_f( - BlockDriverState *bs, - int argc, - char **argv) -{ - const cmdinfo_t *ct; - - if (argc == 1) { - help_all(); - return 0; - } - ct = find_command(argv[1]); - if (ct == NULL) { - printf(_("command %s not found\n"), argv[1]); - return 0; - } - help_onecmd(argv[1], ct); - return 0; -} - -static void -help_onecmd( - const char *cmd, - const cmdinfo_t *ct) -{ - help_oneline(cmd, ct); - if (ct->help) - ct->help(); -} - -static void -help_oneline( - const char *cmd, - const cmdinfo_t *ct) -{ - if (cmd) - printf("%s ", cmd); - else { - printf("%s ", ct->name); - if (ct->altname) - printf("(or %s) ", ct->altname); - } - if (ct->args) - printf("%s ", ct->args); - printf("-- %s\n", ct->oneline); -} - -void -help_init(void) -{ - help_cmd.name = _("help"); - help_cmd.altname = _("?"); - help_cmd.cfunc = help_f; - help_cmd.argmin = 0; - help_cmd.argmax = 1; - help_cmd.flags = CMD_FLAG_GLOBAL; - help_cmd.args = _("[command]"); - help_cmd.oneline = _("help for one or all commands"); - - add_command(&help_cmd); -} diff --git a/cmd.h b/cmd.h index d676408..89e7c6e 100644 --- a/cmd.h +++ b/cmd.h @@ -42,7 +42,6 @@ typedef struct cmdinfo { extern cmdinfo_t *cmdtab; extern int ncmds; -void help_init(void); void quit_init(void); typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index bdfe5f2..e2d65d7 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -1795,6 +1795,71 @@ static const cmdinfo_t abort_cmd = { .oneline = "simulate a program crash using abort(3)", }; +static void help_oneline(const char *cmd, const cmdinfo_t *ct) +{ + if (cmd) { + printf("%s ", cmd); + } else { + printf("%s ", ct->name); + if (ct->altname) { + printf("(or %s) ", ct->altname); + } + } + + if (ct->args) { + printf("%s ", ct->args); + } + printf("-- %s\n", ct->oneline); +} + +static void help_onecmd(const char *cmd, const cmdinfo_t *ct) +{ + help_oneline(cmd, ct); + if (ct->help) { + ct->help(); + } +} + +static void help_all(void) +{ + const cmdinfo_t *ct; + + for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) { + help_oneline(ct->name, ct); + } + printf("\nUse 'help commandname' for extended help.\n"); +} + +static int help_f(BlockDriverState *bs, int argc, char **argv) +{ + const cmdinfo_t *ct; + + if (argc == 1) { + help_all(); + return 0; + } + + ct = find_command(argv[1]); + if (ct == NULL) { + printf("command %s not found\n", argv[1]); + return 0; + } + + help_onecmd(argv[1], ct); + return 0; +} + +static const cmdinfo_t help_cmd = { + .name = "help", + .altname = "?", + .cfunc = help_f, + .argmin = 0, + .argmax = 1, + .flags = CMD_FLAG_GLOBAL, + .args = "[command]", + .oneline = "help for one or all commands", +}; + static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) { if (ct->flags & CMD_FLAG_GLOBAL) { @@ -1834,7 +1899,7 @@ bool qemuio_command(const char *cmd) static void __attribute((constructor)) init_qemuio_commands(void) { /* initialize commands */ - help_init(); + add_command(&help_cmd); add_command(&read_cmd); add_command(&readv_cmd); add_command(&write_cmd); -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function 2013-05-28 15:27 ` [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function Kevin Wolf @ 2013-05-29 21:25 ` Eric Blake 2013-06-04 10:20 ` Kevin Wolf 2013-06-05 12:37 ` Stefan Hajnoczi 1 sibling, 1 reply; 45+ messages in thread From: Eric Blake @ 2013-05-29 21:25 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 1574 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > No reason to treat it different from other commands. Move it to > qemu-io-cmds.c, adapt the coding style and register it like any other > command. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 79 ---------------------------------------------------------- > cmd.h | 1 - > qemu-io-cmds.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++- > 3 files changed, 66 insertions(+), 81 deletions(-) > -void > -help_init(void) > -{ > - help_cmd.name = _("help"); > - help_cmd.altname = _("?"); > - help_cmd.cfunc = help_f; > - help_cmd.argmin = 0; > - help_cmd.argmax = 1; > - help_cmd.flags = CMD_FLAG_GLOBAL; > - help_cmd.args = _("[command]"); > - help_cmd.oneline = _("help for one or all commands"); The old code marked strings for translation with _(), ... > +static const cmdinfo_t help_cmd = { > + .name = "help", > + .altname = "?", > + .cfunc = help_f, > + .argmin = 0, > + .argmax = 1, > + .flags = CMD_FLAG_GLOBAL, > + .args = "[command]", > + .oneline = "help for one or all commands", ...whereas the new code uses fixed literals. Then again, po/messages.po doesn't contain any mention of cmd.c, so it's not like we were actually translating the strings, even if they were marked for translation. So no real change. 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function 2013-05-29 21:25 ` Eric Blake @ 2013-06-04 10:20 ` Kevin Wolf 0 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-06-04 10:20 UTC (permalink / raw) To: Eric Blake; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino Am 29.05.2013 um 23:25 hat Eric Blake geschrieben: > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > No reason to treat it different from other commands. Move it to > > qemu-io-cmds.c, adapt the coding style and register it like any other > > command. > > > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > > --- > > cmd.c | 79 ---------------------------------------------------------- > > cmd.h | 1 - > > qemu-io-cmds.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++- > > 3 files changed, 66 insertions(+), 81 deletions(-) > > > -void > > -help_init(void) > > -{ > > - help_cmd.name = _("help"); > > - help_cmd.altname = _("?"); > > - help_cmd.cfunc = help_f; > > - help_cmd.argmin = 0; > > - help_cmd.argmax = 1; > > - help_cmd.flags = CMD_FLAG_GLOBAL; > > - help_cmd.args = _("[command]"); > > - help_cmd.oneline = _("help for one or all commands"); > > The old code marked strings for translation with _(), ... > > > +static const cmdinfo_t help_cmd = { > > + .name = "help", > > + .altname = "?", > > + .cfunc = help_f, > > + .argmin = 0, > > + .argmax = 1, > > + .flags = CMD_FLAG_GLOBAL, > > + .args = "[command]", > > + .oneline = "help for one or all commands", > > ...whereas the new code uses fixed literals. Then again, po/messages.po > doesn't contain any mention of cmd.c, so it's not like we were actually > translating the strings, even if they were marked for translation. So > no real change. In fact, the real reason is that cmd.c has this line: #define _(x) x /* not gettext support yet */ Apart from that, translating command names has never been a great idea anyway, so at least that part wouldn't a big loss even if gettext was actually used... Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function 2013-05-28 15:27 ` [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function Kevin Wolf 2013-05-29 21:25 ` Eric Blake @ 2013-06-05 12:37 ` Stefan Hajnoczi 1 sibling, 0 replies; 45+ messages in thread From: Stefan Hajnoczi @ 2013-06-05 12:37 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino On Tue, May 28, 2013 at 05:27:28PM +0200, Kevin Wolf wrote: > No reason to treat it different from other commands. Move it to > qemu-io-cmds.c, adapt the coding style and register it like any other > command. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 79 ---------------------------------------------------------- > cmd.h | 1 - > qemu-io-cmds.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++- > 3 files changed, 66 insertions(+), 81 deletions(-) > > diff --git a/cmd.c b/cmd.c > index 7ae978f..2941ad3 100644 > --- a/cmd.c > +++ b/cmd.c 'help' is a generic command, it's not specific to qemu-io. IMO cmd.c is the right place for it since other command-line tools might also need 'help'. ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (7 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-29 22:11 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c Kevin Wolf ` (7 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino This one only makes sense in the context of the qemu-io tool, so move it to qemu-io.c. Adapt coding style and register it like other commands. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 29 ----------------------------- cmd.h | 2 -- qemu-io.c | 17 ++++++++++++++++- 3 files changed, 16 insertions(+), 32 deletions(-) diff --git a/cmd.c b/cmd.c index 2941ad3..8496e74 100644 --- a/cmd.c +++ b/cmd.c @@ -410,32 +410,3 @@ timestr( snprintf(ts, size, "0.%04u sec", (unsigned int) (usec * 10000)); } } - - -/* from libxcmd/quit.c */ - -static cmdinfo_t quit_cmd; - -/* ARGSUSED */ -static int -quit_f( - BlockDriverState *bs, - int argc, - char **argv) -{ - return 1; -} - -void -quit_init(void) -{ - quit_cmd.name = _("quit"); - quit_cmd.altname = _("q"); - quit_cmd.cfunc = quit_f; - quit_cmd.argmin = -1; - quit_cmd.argmax = -1; - quit_cmd.flags = CMD_FLAG_GLOBAL; - quit_cmd.oneline = _("exit the program"); - - add_command(&quit_cmd); -} diff --git a/cmd.h b/cmd.h index 89e7c6e..5b6f61b 100644 --- a/cmd.h +++ b/cmd.h @@ -42,8 +42,6 @@ typedef struct cmdinfo { extern cmdinfo_t *cmdtab; extern int ncmds; -void quit_init(void); - typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); void add_command(const cmdinfo_t *ci); diff --git a/qemu-io.c b/qemu-io.c index cfee840..83788b4 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -137,6 +137,21 @@ static int open_f(BlockDriverState *bs, int argc, char **argv) return openfile(argv[optind], flags, growable); } +static int quit_f(BlockDriverState *bs, int argc, char **argv) +{ + return 1; +} + +static const cmdinfo_t quit_cmd = { + .name = "quit", + .altname = "q", + .cfunc = quit_f, + .argmin = -1, + .argmax = -1, + .flags = CMD_FLAG_GLOBAL, + .oneline = "exit the program", +}; + static void usage(const char *name) { printf( @@ -247,7 +262,7 @@ int main(int argc, char **argv) bdrv_init(); /* initialize commands */ - quit_init(); + add_command(&quit_cmd); add_command(&open_cmd); add_command(&close_cmd); -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function 2013-05-28 15:27 ` [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function Kevin Wolf @ 2013-05-29 22:11 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-29 22:11 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 576 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > This one only makes sense in the context of the qemu-io tool, so move it > to qemu-io.c. Adapt coding style and register it like other commands. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 29 ----------------------------- > cmd.h | 2 -- > qemu-io.c | 17 ++++++++++++++++- > 3 files changed, 16 insertions(+), 32 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (8 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-29 22:12 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands Kevin Wolf ` (6 subsequent siblings) 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 21 --------------------- include/qemu-common.h | 1 + util/cutils.c | 21 +++++++++++++++++++++ 3 files changed, 22 insertions(+), 21 deletions(-) diff --git a/cmd.c b/cmd.c index 8496e74..f6bf2c5 100644 --- a/cmd.c +++ b/cmd.c @@ -255,27 +255,6 @@ fetchline(void) } #endif -static char *qemu_strsep(char **input, const char *delim) -{ - char *result = *input; - if (result != NULL) { - char *p; - - for (p = result; *p != '\0'; p++) { - if (strchr(delim, *p)) { - break; - } - } - if (*p == '\0') { - *input = NULL; - } else { - *p = '\0'; - *input = p + 1; - } - } - return result; -} - char **breakline(char *input, int *count) { int c = 0; diff --git a/include/qemu-common.h b/include/qemu-common.h index d95ea1e..ed8b6e2 100644 --- a/include/qemu-common.h +++ b/include/qemu-common.h @@ -174,6 +174,7 @@ char *pstrcat(char *buf, int buf_size, const char *s); int strstart(const char *str, const char *val, const char **ptr); int stristart(const char *str, const char *val, const char **ptr); int qemu_strnlen(const char *s, int max_len); +char *qemu_strsep(char **input, const char *delim); time_t mktimegm(struct tm *tm); int qemu_fls(int i); int qemu_fdatasync(int fd); diff --git a/util/cutils.c b/util/cutils.c index 8f28896..0116fcd 100644 --- a/util/cutils.c +++ b/util/cutils.c @@ -107,6 +107,27 @@ int qemu_strnlen(const char *s, int max_len) return i; } +char *qemu_strsep(char **input, const char *delim) +{ + char *result = *input; + if (result != NULL) { + char *p; + + for (p = result; *p != '\0'; p++) { + if (strchr(delim, *p)) { + break; + } + } + if (*p == '\0') { + *input = NULL; + } else { + *p = '\0'; + *input = p + 1; + } + } + return result; +} + time_t mktimegm(struct tm *tm) { time_t t; -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c 2013-05-28 15:27 ` [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c Kevin Wolf @ 2013-05-29 22:12 ` Eric Blake 0 siblings, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-29 22:12 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 450 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 21 --------------------- > include/qemu-common.h | 1 + > util/cutils.c | 21 +++++++++++++++++++++ > 3 files changed, 22 insertions(+), 21 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (9 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-30 2:27 ` Eric Blake 2013-06-05 12:42 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 12/16] qemu-io: Move command_loop() and friends Kevin Wolf ` (5 subsequent siblings) 16 siblings, 2 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 113 --------------------------------- cmd.h | 11 +--- qemu-io-cmds.c | 192 ++++++++++++++++++++++++++++++++++++++++++--------------- qemu-io.c | 10 +-- 4 files changed, 148 insertions(+), 178 deletions(-) diff --git a/cmd.c b/cmd.c index f6bf2c5..6616d61 100644 --- a/cmd.c +++ b/cmd.c @@ -31,94 +31,9 @@ /* from libxcmd/command.c */ -cmdinfo_t *cmdtab; -int ncmds; - -static checkfunc_t check_func; static int ncmdline; static char **cmdline; -static int -compare(const void *a, const void *b) -{ - return strcmp(((const cmdinfo_t *)a)->name, - ((const cmdinfo_t *)b)->name); -} - -void add_command(const cmdinfo_t *ci) -{ - cmdtab = g_realloc((void *)cmdtab, ++ncmds * sizeof(*cmdtab)); - cmdtab[ncmds - 1] = *ci; - qsort(cmdtab, ncmds, sizeof(*cmdtab), compare); -} - -static int -check_command( - const cmdinfo_t *ci) -{ - if (check_func) - return check_func(qemuio_bs, ci); - return 1; -} - -void -add_check_command( - checkfunc_t cf) -{ - check_func = cf; -} - -int -command_usage( - const cmdinfo_t *ci) -{ - printf("%s %s -- %s\n", ci->name, ci->args, ci->oneline); - return 0; -} - -int -command( - const cmdinfo_t *ct, - int argc, - char **argv) -{ - char *cmd = argv[0]; - - if (!check_command(ct)) - return 0; - - if (argc-1 < ct->argmin || (ct->argmax != -1 && argc-1 > ct->argmax)) { - if (ct->argmax == -1) - fprintf(stderr, - _("bad argument count %d to %s, expected at least %d arguments\n"), - argc-1, cmd, ct->argmin); - else if (ct->argmin == ct->argmax) - fprintf(stderr, - _("bad argument count %d to %s, expected %d arguments\n"), - argc-1, cmd, ct->argmin); - else - fprintf(stderr, - _("bad argument count %d to %s, expected between %d and %d arguments\n"), - argc-1, cmd, ct->argmin, ct->argmax); - return 0; - } - optind = 0; - return ct->cfunc(qemuio_bs, argc, argv); -} - -const cmdinfo_t * -find_command( - const char *cmd) -{ - cmdinfo_t *ct; - - for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) { - if (strcmp(ct->name, cmd) == 0 || - (ct->altname && strcmp(ct->altname, cmd) == 0)) - return (const cmdinfo_t *)ct; - } - return NULL; -} void add_user_command(char *optarg) { @@ -255,34 +170,6 @@ fetchline(void) } #endif -char **breakline(char *input, int *count) -{ - int c = 0; - char *p; - char **rval = calloc(sizeof(char *), 1); - char **tmp; - - while (rval && (p = qemu_strsep(&input, " ")) != NULL) { - if (!*p) { - continue; - } - c++; - tmp = realloc(rval, sizeof(*rval) * (c + 1)); - if (!tmp) { - free(rval); - rval = NULL; - c = 0; - break; - } else { - rval = tmp; - } - rval[c - 1] = p; - rval[c] = NULL; - } - *count = c; - return rval; -} - #define EXABYTES(x) ((long long)(x) << 60) #define PETABYTES(x) ((long long)(x) << 50) #define TERABYTES(x) ((long long)(x) << 40) diff --git a/cmd.h b/cmd.h index 5b6f61b..0d01a33 100644 --- a/cmd.h +++ b/cmd.h @@ -39,23 +39,16 @@ typedef struct cmdinfo { helpfunc_t help; } cmdinfo_t; -extern cmdinfo_t *cmdtab; -extern int ncmds; - typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); -void add_command(const cmdinfo_t *ci); +void qemuio_add_command(const cmdinfo_t *ci); void add_user_command(char *optarg); void add_check_command(checkfunc_t cf); -const cmdinfo_t *find_command(const char *cmd); - void command_loop(void); -int command_usage(const cmdinfo_t *ci); -int command(const cmdinfo_t *ci, int argc, char **argv); +int qemuio_command_usage(const cmdinfo_t *ci); /* from input.h */ -char **breakline(char *input, int *count); char *fetchline(void); void cvtstr(double value, char *str, size_t sz); diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index e2d65d7..94edd39 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -16,6 +16,110 @@ int qemuio_misalign; +static cmdinfo_t *cmdtab; +static int ncmds; + +static int compare_cmdname(const void *a, const void *b) +{ + return strcmp(((const cmdinfo_t *)a)->name, + ((const cmdinfo_t *)b)->name); +} + +void qemuio_add_command(const cmdinfo_t *ci) +{ + cmdtab = g_realloc(cmdtab, ++ncmds * sizeof(*cmdtab)); + cmdtab[ncmds - 1] = *ci; + qsort(cmdtab, ncmds, sizeof(*cmdtab), compare_cmdname); +} + +int qemuio_command_usage(const cmdinfo_t *ci) +{ + printf("%s %s -- %s\n", ci->name, ci->args, ci->oneline); + return 0; +} + +static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) +{ + if (ct->flags & CMD_FLAG_GLOBAL) { + return 1; + } + if (!(ct->flags & CMD_NOFILE_OK) && !bs) { + fprintf(stderr, "no file open, try 'help open'\n"); + return 0; + } + return 1; +} + +static int command(const cmdinfo_t *ct, int argc, char **argv) +{ + char *cmd = argv[0]; + + if (!init_check_command(qemuio_bs, ct)) { + return 0; + } + + if (argc - 1 < ct->argmin || (ct->argmax != -1 && argc - 1 > ct->argmax)) { + if (ct->argmax == -1) { + fprintf(stderr, + "bad argument count %d to %s, expected at least %d arguments\n", + argc-1, cmd, ct->argmin); + } else if (ct->argmin == ct->argmax) { + fprintf(stderr, + "bad argument count %d to %s, expected %d arguments\n", + argc-1, cmd, ct->argmin); + } else { + fprintf(stderr, + "bad argument count %d to %s, expected between %d and %d arguments\n", + argc-1, cmd, ct->argmin, ct->argmax); + } + return 0; + } + optind = 0; + return ct->cfunc(qemuio_bs, argc, argv); +} + +static const cmdinfo_t *find_command(const char *cmd) +{ + cmdinfo_t *ct; + + for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) { + if (strcmp(ct->name, cmd) == 0 || + (ct->altname && strcmp(ct->altname, cmd) == 0)) + { + return (const cmdinfo_t *)ct; + } + } + return NULL; +} + +static char **breakline(char *input, int *count) +{ + int c = 0; + char *p; + char **rval = g_malloc0(sizeof(char *)); + char **tmp; + + while (rval && (p = qemu_strsep(&input, " ")) != NULL) { + if (!*p) { + continue; + } + c++; + tmp = g_realloc(rval, sizeof(*rval) * (c + 1)); + if (!tmp) { + g_free(rval); + rval = NULL; + c = 0; + break; + } else { + rval = tmp; + } + rval[c - 1] = p; + rval[c] = NULL; + } + *count = c; + return rval; +} + static int64_t cvtnum(const char *s) { char *end; @@ -467,12 +571,12 @@ static int read_f(BlockDriverState *bs, int argc, char **argv) vflag = 1; break; default: - return command_usage(&read_cmd); + return qemuio_command_usage(&read_cmd); } } if (optind != argc - 2) { - return command_usage(&read_cmd); + return qemuio_command_usage(&read_cmd); } if (bflag && pflag) { @@ -494,7 +598,7 @@ static int read_f(BlockDriverState *bs, int argc, char **argv) } if (!Pflag && (lflag || sflag)) { - return command_usage(&read_cmd); + return qemuio_command_usage(&read_cmd); } if (!lflag) { @@ -629,12 +733,12 @@ static int readv_f(BlockDriverState *bs, int argc, char **argv) vflag = 1; break; default: - return command_usage(&readv_cmd); + return qemuio_command_usage(&readv_cmd); } } if (optind > argc - 2) { - return command_usage(&readv_cmd); + return qemuio_command_usage(&readv_cmd); } @@ -769,12 +873,12 @@ static int write_f(BlockDriverState *bs, int argc, char **argv) zflag = 1; break; default: - return command_usage(&write_cmd); + return qemuio_command_usage(&write_cmd); } } if (optind != argc - 2) { - return command_usage(&write_cmd); + return qemuio_command_usage(&write_cmd); } if (bflag + pflag + zflag > 1) { @@ -911,12 +1015,12 @@ static int writev_f(BlockDriverState *bs, int argc, char **argv) } break; default: - return command_usage(&writev_cmd); + return qemuio_command_usage(&writev_cmd); } } if (optind > argc - 2) { - return command_usage(&writev_cmd); + return qemuio_command_usage(&writev_cmd); } offset = cvtnum(argv[optind]); @@ -1023,12 +1127,12 @@ static int multiwrite_f(BlockDriverState *bs, int argc, char **argv) } break; default: - return command_usage(&writev_cmd); + return qemuio_command_usage(&writev_cmd); } } if (optind > argc - 2) { - return command_usage(&writev_cmd); + return qemuio_command_usage(&writev_cmd); } nr_reqs = 1; @@ -1257,13 +1361,13 @@ static int aio_read_f(BlockDriverState *bs, int argc, char **argv) break; default: g_free(ctx); - return command_usage(&aio_read_cmd); + return qemuio_command_usage(&aio_read_cmd); } } if (optind > argc - 2) { g_free(ctx); - return command_usage(&aio_read_cmd); + return qemuio_command_usage(&aio_read_cmd); } ctx->offset = cvtnum(argv[optind]); @@ -1349,13 +1453,13 @@ static int aio_write_f(BlockDriverState *bs, int argc, char **argv) break; default: g_free(ctx); - return command_usage(&aio_write_cmd); + return qemuio_command_usage(&aio_write_cmd); } } if (optind > argc - 2) { g_free(ctx); - return command_usage(&aio_write_cmd); + return qemuio_command_usage(&aio_write_cmd); } ctx->offset = cvtnum(argv[optind]); @@ -1547,12 +1651,12 @@ static int discard_f(BlockDriverState *bs, int argc, char **argv) qflag = 1; break; default: - return command_usage(&discard_cmd); + return qemuio_command_usage(&discard_cmd); } } if (optind != argc - 2) { - return command_usage(&discard_cmd); + return qemuio_command_usage(&discard_cmd); } offset = cvtnum(argv[optind]); @@ -1860,18 +1964,6 @@ static const cmdinfo_t help_cmd = { .oneline = "help for one or all commands", }; -static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) -{ - if (ct->flags & CMD_FLAG_GLOBAL) { - return 1; - } - if (!(ct->flags & CMD_NOFILE_OK) && !bs) { - fprintf(stderr, "no file open, try 'help open'\n"); - return 0; - } - return 1; -} - bool qemuio_command(const char *cmd) { char *input; @@ -1899,26 +1991,24 @@ bool qemuio_command(const char *cmd) static void __attribute((constructor)) init_qemuio_commands(void) { /* initialize commands */ - add_command(&help_cmd); - add_command(&read_cmd); - add_command(&readv_cmd); - add_command(&write_cmd); - add_command(&writev_cmd); - add_command(&multiwrite_cmd); - add_command(&aio_read_cmd); - add_command(&aio_write_cmd); - add_command(&aio_flush_cmd); - add_command(&flush_cmd); - add_command(&truncate_cmd); - add_command(&length_cmd); - add_command(&info_cmd); - add_command(&discard_cmd); - add_command(&alloc_cmd); - add_command(&map_cmd); - add_command(&break_cmd); - add_command(&resume_cmd); - add_command(&wait_break_cmd); - add_command(&abort_cmd); - - add_check_command(init_check_command); + qemuio_add_command(&help_cmd); + qemuio_add_command(&read_cmd); + qemuio_add_command(&readv_cmd); + qemuio_add_command(&write_cmd); + qemuio_add_command(&writev_cmd); + qemuio_add_command(&multiwrite_cmd); + qemuio_add_command(&aio_read_cmd); + qemuio_add_command(&aio_write_cmd); + qemuio_add_command(&aio_flush_cmd); + qemuio_add_command(&flush_cmd); + qemuio_add_command(&truncate_cmd); + qemuio_add_command(&length_cmd); + qemuio_add_command(&info_cmd); + qemuio_add_command(&discard_cmd); + qemuio_add_command(&alloc_cmd); + qemuio_add_command(&map_cmd); + qemuio_add_command(&break_cmd); + qemuio_add_command(&resume_cmd); + qemuio_add_command(&wait_break_cmd); + qemuio_add_command(&abort_cmd); } diff --git a/qemu-io.c b/qemu-io.c index 83788b4..ba31c5b 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -122,7 +122,7 @@ static int open_f(BlockDriverState *bs, int argc, char **argv) growable = 1; break; default: - return command_usage(&open_cmd); + return qemuio_command_usage(&open_cmd); } } @@ -131,7 +131,7 @@ static int open_f(BlockDriverState *bs, int argc, char **argv) } if (optind != argc - 1) { - return command_usage(&open_cmd); + return qemuio_command_usage(&open_cmd); } return openfile(argv[optind], flags, growable); @@ -262,9 +262,9 @@ int main(int argc, char **argv) bdrv_init(); /* initialize commands */ - add_command(&quit_cmd); - add_command(&open_cmd); - add_command(&close_cmd); + qemuio_add_command(&quit_cmd); + qemuio_add_command(&open_cmd); + qemuio_add_command(&close_cmd); /* open the device */ if (!readonly) { -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands 2013-05-28 15:27 ` [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands Kevin Wolf @ 2013-05-30 2:27 ` Eric Blake 2013-06-05 12:42 ` Stefan Hajnoczi 1 sibling, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-30 2:27 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 514 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 113 --------------------------------- > cmd.h | 11 +--- > qemu-io-cmds.c | 192 ++++++++++++++++++++++++++++++++++++++++++--------------- > qemu-io.c | 10 +-- > 4 files changed, 148 insertions(+), 178 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands 2013-05-28 15:27 ` [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands Kevin Wolf 2013-05-30 2:27 ` Eric Blake @ 2013-06-05 12:42 ` Stefan Hajnoczi 1 sibling, 0 replies; 45+ messages in thread From: Stefan Hajnoczi @ 2013-06-05 12:42 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino On Tue, May 28, 2013 at 05:27:31PM +0200, Kevin Wolf wrote: > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > cmd.c | 113 --------------------------------- > cmd.h | 11 +--- > qemu-io-cmds.c | 192 ++++++++++++++++++++++++++++++++++++++++++--------------- > qemu-io.c | 10 +-- > 4 files changed, 148 insertions(+), 178 deletions(-) I haven't figured this out yet. These patches move a bunch of generic code and make it qemu-io-specific. That seems the wrong direction to go. ^ permalink raw reply [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 12/16] qemu-io: Move command_loop() and friends 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (10 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 13/16] qemu-io: Move remaining helpers from cmd.c Kevin Wolf ` (4 subsequent siblings) 16 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.c | 139 -------------------------------------------------------------- cmd.h | 9 ---- qemu-io.c | 139 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 139 insertions(+), 148 deletions(-) diff --git a/cmd.c b/cmd.c index 6616d61..26d38a8 100644 --- a/cmd.c +++ b/cmd.c @@ -31,145 +31,6 @@ /* from libxcmd/command.c */ -static int ncmdline; -static char **cmdline; - - -void add_user_command(char *optarg) -{ - cmdline = g_realloc(cmdline, ++ncmdline * sizeof(char *)); - cmdline[ncmdline-1] = optarg; -} - -static void prep_fetchline(void *opaque) -{ - int *fetchable = opaque; - - qemu_set_fd_handler(STDIN_FILENO, NULL, NULL, NULL); - *fetchable= 1; -} - -static char *get_prompt(void); - -void command_loop(void) -{ - int i, done = 0, fetchable = 0, prompted = 0; - char *input; - - for (i = 0; !done && i < ncmdline; i++) { - done = qemuio_command(cmdline[i]); - } - if (cmdline) { - g_free(cmdline); - return; - } - - while (!done) { - if (!prompted) { - printf("%s", get_prompt()); - fflush(stdout); - qemu_set_fd_handler(STDIN_FILENO, prep_fetchline, NULL, &fetchable); - prompted = 1; - } - - main_loop_wait(false); - - if (!fetchable) { - continue; - } - - input = fetchline(); - if (input == NULL) { - break; - } - done = qemuio_command(input); - free(input); - - prompted = 0; - fetchable = 0; - } - qemu_set_fd_handler(STDIN_FILENO, NULL, NULL, NULL); -} - -/* from libxcmd/input.c */ - -#if defined(ENABLE_READLINE) -# include <readline/history.h> -# include <readline/readline.h> -#elif defined(ENABLE_EDITLINE) -# include <histedit.h> -#endif - -static char * -get_prompt(void) -{ - static char prompt[FILENAME_MAX + 2 /*"> "*/ + 1 /*"\0"*/ ]; - - if (!prompt[0]) - snprintf(prompt, sizeof(prompt), "%s> ", progname); - return prompt; -} - -#if defined(ENABLE_READLINE) -char * -fetchline(void) -{ - char *line; - - line = readline(get_prompt()); - if (line && *line) - add_history(line); - return line; -} -#elif defined(ENABLE_EDITLINE) -static char *el_get_prompt(EditLine *e) { return get_prompt(); } -char * -fetchline(void) -{ - static EditLine *el; - static History *hist; - HistEvent hevent; - char *line; - int count; - - if (!el) { - hist = history_init(); - history(hist, &hevent, H_SETSIZE, 100); - el = el_init(progname, stdin, stdout, stderr); - el_source(el, NULL); - el_set(el, EL_SIGNAL, 1); - el_set(el, EL_PROMPT, el_get_prompt); - el_set(el, EL_HIST, history, (const char *)hist); - } - line = strdup(el_gets(el, &count)); - if (line) { - if (count > 0) - line[count-1] = '\0'; - if (*line) - history(hist, &hevent, H_ENTER, line); - } - return line; -} -#else -# define MAXREADLINESZ 1024 -char * -fetchline(void) -{ - char *p, *line = malloc(MAXREADLINESZ); - - if (!line) - return NULL; - if (!fgets(line, MAXREADLINESZ, stdin)) { - free(line); - return NULL; - } - p = line + strlen(line); - if (p != line && p[-1] == '\n') - p[-1] = '\0'; - return line; -} -#endif - #define EXABYTES(x) ((long long)(x) << 60) #define PETABYTES(x) ((long long)(x) << 50) #define TERABYTES(x) ((long long)(x) << 40) diff --git a/cmd.h b/cmd.h index 0d01a33..da0c7cf 100644 --- a/cmd.h +++ b/cmd.h @@ -39,18 +39,11 @@ typedef struct cmdinfo { helpfunc_t help; } cmdinfo_t; -typedef int (*checkfunc_t)(BlockDriverState *bs, const cmdinfo_t *ci); - void qemuio_add_command(const cmdinfo_t *ci); -void add_user_command(char *optarg); -void add_check_command(checkfunc_t cf); -void command_loop(void); int qemuio_command_usage(const cmdinfo_t *ci); /* from input.h */ -char *fetchline(void); - void cvtstr(double value, char *str, size_t sz); struct timeval tsub(struct timeval t1, struct timeval t2); @@ -64,8 +57,6 @@ enum { void timestr(struct timeval *tv, char *str, size_t sz, int flags); -extern char *progname; - bool qemuio_command(const char *cmd); #endif /* __COMMAND_H__ */ diff --git a/qemu-io.c b/qemu-io.c index ba31c5b..3297f33 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -29,6 +29,10 @@ char *progname; BlockDriverState *qemuio_bs; extern int qemuio_misalign; +/* qemu-io commands passed using -c */ +static int ncmdline; +static char **cmdline; + static int close_f(BlockDriverState *bs, int argc, char **argv) { bdrv_delete(bs); @@ -174,6 +178,141 @@ static void usage(const char *name) } +#if defined(ENABLE_READLINE) +# include <readline/history.h> +# include <readline/readline.h> +#elif defined(ENABLE_EDITLINE) +# include <histedit.h> +#endif + +static char *get_prompt(void) +{ + static char prompt[FILENAME_MAX + 2 /*"> "*/ + 1 /*"\0"*/ ]; + + if (!prompt[0]) { + snprintf(prompt, sizeof(prompt), "%s> ", progname); + } + + return prompt; +} + +#if defined(ENABLE_READLINE) +static char *fetchline(void) +{ + char *line = readline(get_prompt()); + if (line && *line) { + add_history(line); + } + return line; +} +#elif defined(ENABLE_EDITLINE) +static char *el_get_prompt(EditLine *e) +{ + return get_prompt(); +} + +static char *fetchline(void) +{ + static EditLine *el; + static History *hist; + HistEvent hevent; + char *line; + int count; + + if (!el) { + hist = history_init(); + history(hist, &hevent, H_SETSIZE, 100); + el = el_init(progname, stdin, stdout, stderr); + el_source(el, NULL); + el_set(el, EL_SIGNAL, 1); + el_set(el, EL_PROMPT, el_get_prompt); + el_set(el, EL_HIST, history, (const char *)hist); + } + line = strdup(el_gets(el, &count)); + if (line) { + if (count > 0) { + line[count-1] = '\0'; + } + if (*line) { + history(hist, &hevent, H_ENTER, line); + } + } + return line; +} +#else +# define MAXREADLINESZ 1024 +static char *fetchline(void) +{ + char *p, *line = g_malloc(MAXREADLINESZ); + + if (!fgets(line, MAXREADLINESZ, stdin)) { + g_free(line); + return NULL; + } + + p = line + strlen(line); + if (p != line && p[-1] == '\n') { + p[-1] = '\0'; + } + + return line; +} +#endif + +static void prep_fetchline(void *opaque) +{ + int *fetchable = opaque; + + qemu_set_fd_handler(STDIN_FILENO, NULL, NULL, NULL); + *fetchable= 1; +} + +static void command_loop(void) +{ + int i, done = 0, fetchable = 0, prompted = 0; + char *input; + + for (i = 0; !done && i < ncmdline; i++) { + done = qemuio_command(cmdline[i]); + } + if (cmdline) { + g_free(cmdline); + return; + } + + while (!done) { + if (!prompted) { + printf("%s", get_prompt()); + fflush(stdout); + qemu_set_fd_handler(STDIN_FILENO, prep_fetchline, NULL, &fetchable); + prompted = 1; + } + + main_loop_wait(false); + + if (!fetchable) { + continue; + } + + input = fetchline(); + if (input == NULL) { + break; + } + done = qemuio_command(input); + g_free(input); + + prompted = 0; + fetchable = 0; + } + qemu_set_fd_handler(STDIN_FILENO, NULL, NULL, NULL); +} + +static void add_user_command(char *optarg) +{ + cmdline = g_realloc(cmdline, ++ncmdline * sizeof(char *)); + cmdline[ncmdline-1] = optarg; +} + int main(int argc, char **argv) { int readonly = 0; -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 13/16] qemu-io: Move remaining helpers from cmd.c 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (11 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 12/16] qemu-io: Move command_loop() and friends Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 14/16] qemu-io: Interface cleanup Kevin Wolf ` (3 subsequent siblings) 16 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- Makefile | 2 +- cmd.c | 139 --------------------------------------------------------- cmd.h | 14 ------ qemu-io-cmds.c | 104 ++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 105 insertions(+), 154 deletions(-) delete mode 100644 cmd.c diff --git a/Makefile b/Makefile index df2212b..581150f 100644 --- a/Makefile +++ b/Makefile @@ -186,7 +186,7 @@ qemu-img.o: qemu-img-cmds.h qemu-img$(EXESUF): qemu-img.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-nbd$(EXESUF): qemu-nbd.o $(block-obj-y) libqemuutil.a libqemustub.a -qemu-io$(EXESUF): qemu-io.o qemu-io-cmds.o cmd.o $(block-obj-y) libqemuutil.a libqemustub.a +qemu-io$(EXESUF): qemu-io.o qemu-io-cmds.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-bridge-helper$(EXESUF): qemu-bridge-helper.o diff --git a/cmd.c b/cmd.c deleted file mode 100644 index 26d38a8..0000000 --- a/cmd.c +++ /dev/null @@ -1,139 +0,0 @@ -/* - * Copyright (c) 2003-2005 Silicon Graphics, Inc. - * All Rights Reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License as - * published by the Free Software Foundation. - * - * This program is distributed in the hope that it would be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, see <http://www.gnu.org/licenses/>. - */ - -#include <stdio.h> -#include <stdlib.h> -#include <string.h> -#include <ctype.h> -#include <errno.h> -#include <sys/time.h> -#include <getopt.h> - -#include "cmd.h" -#include "block/aio.h" -#include "qemu/main-loop.h" - -#define _(x) x /* not gettext support yet */ - -/* from libxcmd/command.c */ - -#define EXABYTES(x) ((long long)(x) << 60) -#define PETABYTES(x) ((long long)(x) << 50) -#define TERABYTES(x) ((long long)(x) << 40) -#define GIGABYTES(x) ((long long)(x) << 30) -#define MEGABYTES(x) ((long long)(x) << 20) -#define KILOBYTES(x) ((long long)(x) << 10) - -#define TO_EXABYTES(x) ((x) / EXABYTES(1)) -#define TO_PETABYTES(x) ((x) / PETABYTES(1)) -#define TO_TERABYTES(x) ((x) / TERABYTES(1)) -#define TO_GIGABYTES(x) ((x) / GIGABYTES(1)) -#define TO_MEGABYTES(x) ((x) / MEGABYTES(1)) -#define TO_KILOBYTES(x) ((x) / KILOBYTES(1)) - -void -cvtstr( - double value, - char *str, - size_t size) -{ - char *trim; - const char *suffix; - - if (value >= EXABYTES(1)) { - suffix = " EiB"; - snprintf(str, size - 4, "%.3f", TO_EXABYTES(value)); - } else if (value >= PETABYTES(1)) { - suffix = " PiB"; - snprintf(str, size - 4, "%.3f", TO_PETABYTES(value)); - } else if (value >= TERABYTES(1)) { - suffix = " TiB"; - snprintf(str, size - 4, "%.3f", TO_TERABYTES(value)); - } else if (value >= GIGABYTES(1)) { - suffix = " GiB"; - snprintf(str, size - 4, "%.3f", TO_GIGABYTES(value)); - } else if (value >= MEGABYTES(1)) { - suffix = " MiB"; - snprintf(str, size - 4, "%.3f", TO_MEGABYTES(value)); - } else if (value >= KILOBYTES(1)) { - suffix = " KiB"; - snprintf(str, size - 4, "%.3f", TO_KILOBYTES(value)); - } else { - suffix = " bytes"; - snprintf(str, size - 6, "%f", value); - } - - trim = strstr(str, ".000"); - if (trim) { - strcpy(trim, suffix); - } else { - strcat(str, suffix); - } -} - -struct timeval -tsub(struct timeval t1, struct timeval t2) -{ - t1.tv_usec -= t2.tv_usec; - if (t1.tv_usec < 0) { - t1.tv_usec += 1000000; - t1.tv_sec--; - } - t1.tv_sec -= t2.tv_sec; - return t1; -} - -double -tdiv(double value, struct timeval tv) -{ - return value / ((double)tv.tv_sec + ((double)tv.tv_usec / 1000000.0)); -} - -#define HOURS(sec) ((sec) / (60 * 60)) -#define MINUTES(sec) (((sec) % (60 * 60)) / 60) -#define SECONDS(sec) ((sec) % 60) - -void -timestr( - struct timeval *tv, - char *ts, - size_t size, - int format) -{ - double usec = (double)tv->tv_usec / 1000000.0; - - if (format & TERSE_FIXED_TIME) { - if (!HOURS(tv->tv_sec)) { - snprintf(ts, size, "%u:%02u.%02u", - (unsigned int) MINUTES(tv->tv_sec), - (unsigned int) SECONDS(tv->tv_sec), - (unsigned int) (usec * 100)); - return; - } - format |= VERBOSE_FIXED_TIME; /* fallback if hours needed */ - } - - if ((format & VERBOSE_FIXED_TIME) || tv->tv_sec) { - snprintf(ts, size, "%u:%02u:%02u.%02u", - (unsigned int) HOURS(tv->tv_sec), - (unsigned int) MINUTES(tv->tv_sec), - (unsigned int) SECONDS(tv->tv_sec), - (unsigned int) (usec * 100)); - } else { - snprintf(ts, size, "0.%04u sec", (unsigned int) (usec * 10000)); - } -} diff --git a/cmd.h b/cmd.h index da0c7cf..9907795 100644 --- a/cmd.h +++ b/cmd.h @@ -43,20 +43,6 @@ void qemuio_add_command(const cmdinfo_t *ci); int qemuio_command_usage(const cmdinfo_t *ci); -/* from input.h */ -void cvtstr(double value, char *str, size_t sz); - -struct timeval tsub(struct timeval t1, struct timeval t2); -double tdiv(double value, struct timeval tv); - -enum { - DEFAULT_TIME = 0x0, - TERSE_FIXED_TIME = 0x1, - VERBOSE_FIXED_TIME = 0x2 -}; - -void timestr(struct timeval *tv, char *str, size_t sz, int flags); - bool qemuio_command(const char *cmd); #endif /* __COMMAND_H__ */ diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index 94edd39..82f9be6 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -126,6 +126,110 @@ static int64_t cvtnum(const char *s) return strtosz_suffix(s, &end, STRTOSZ_DEFSUFFIX_B); } +#define EXABYTES(x) ((long long)(x) << 60) +#define PETABYTES(x) ((long long)(x) << 50) +#define TERABYTES(x) ((long long)(x) << 40) +#define GIGABYTES(x) ((long long)(x) << 30) +#define MEGABYTES(x) ((long long)(x) << 20) +#define KILOBYTES(x) ((long long)(x) << 10) + +#define TO_EXABYTES(x) ((x) / EXABYTES(1)) +#define TO_PETABYTES(x) ((x) / PETABYTES(1)) +#define TO_TERABYTES(x) ((x) / TERABYTES(1)) +#define TO_GIGABYTES(x) ((x) / GIGABYTES(1)) +#define TO_MEGABYTES(x) ((x) / MEGABYTES(1)) +#define TO_KILOBYTES(x) ((x) / KILOBYTES(1)) + +static void cvtstr(double value, char *str, size_t size) +{ + char *trim; + const char *suffix; + + if (value >= EXABYTES(1)) { + suffix = " EiB"; + snprintf(str, size - 4, "%.3f", TO_EXABYTES(value)); + } else if (value >= PETABYTES(1)) { + suffix = " PiB"; + snprintf(str, size - 4, "%.3f", TO_PETABYTES(value)); + } else if (value >= TERABYTES(1)) { + suffix = " TiB"; + snprintf(str, size - 4, "%.3f", TO_TERABYTES(value)); + } else if (value >= GIGABYTES(1)) { + suffix = " GiB"; + snprintf(str, size - 4, "%.3f", TO_GIGABYTES(value)); + } else if (value >= MEGABYTES(1)) { + suffix = " MiB"; + snprintf(str, size - 4, "%.3f", TO_MEGABYTES(value)); + } else if (value >= KILOBYTES(1)) { + suffix = " KiB"; + snprintf(str, size - 4, "%.3f", TO_KILOBYTES(value)); + } else { + suffix = " bytes"; + snprintf(str, size - 6, "%f", value); + } + + trim = strstr(str, ".000"); + if (trim) { + strcpy(trim, suffix); + } else { + strcat(str, suffix); + } +} + + + +static struct timeval tsub(struct timeval t1, struct timeval t2) +{ + t1.tv_usec -= t2.tv_usec; + if (t1.tv_usec < 0) { + t1.tv_usec += 1000000; + t1.tv_sec--; + } + t1.tv_sec -= t2.tv_sec; + return t1; +} + +static double tdiv(double value, struct timeval tv) +{ + return value / ((double)tv.tv_sec + ((double)tv.tv_usec / 1000000.0)); +} + +#define HOURS(sec) ((sec) / (60 * 60)) +#define MINUTES(sec) (((sec) % (60 * 60)) / 60) +#define SECONDS(sec) ((sec) % 60) + +enum { + DEFAULT_TIME = 0x0, + TERSE_FIXED_TIME = 0x1, + VERBOSE_FIXED_TIME = 0x2, +}; + +static void timestr(struct timeval *tv, char *ts, size_t size, int format) +{ + double usec = (double)tv->tv_usec / 1000000.0; + + if (format & TERSE_FIXED_TIME) { + if (!HOURS(tv->tv_sec)) { + snprintf(ts, size, "%u:%02u.%02u", + (unsigned int) MINUTES(tv->tv_sec), + (unsigned int) SECONDS(tv->tv_sec), + (unsigned int) (usec * 100)); + return; + } + format |= VERBOSE_FIXED_TIME; /* fallback if hours needed */ + } + + if ((format & VERBOSE_FIXED_TIME) || tv->tv_sec) { + snprintf(ts, size, "%u:%02u:%02u.%02u", + (unsigned int) HOURS(tv->tv_sec), + (unsigned int) MINUTES(tv->tv_sec), + (unsigned int) SECONDS(tv->tv_sec), + (unsigned int) (usec * 100)); + } else { + snprintf(ts, size, "0.%04u sec", (unsigned int) (usec * 10000)); + } +} + /* * Parse the pattern argument to various sub-commands. * -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 14/16] qemu-io: Interface cleanup 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (12 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 13/16] qemu-io: Move remaining helpers from cmd.c Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 15/16] qemu-io: Use the qemu version for -V Kevin Wolf ` (2 subsequent siblings) 16 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- cmd.h | 48 ------------------------------------------------ include/qemu-io.h | 46 ++++++++++++++++++++++++++++++++++++++++++++++ qemu-io-cmds.c | 14 +++++++------- qemu-io.c | 7 +++---- 4 files changed, 56 insertions(+), 59 deletions(-) delete mode 100644 cmd.h create mode 100644 include/qemu-io.h diff --git a/cmd.h b/cmd.h deleted file mode 100644 index 9907795..0000000 --- a/cmd.h +++ /dev/null @@ -1,48 +0,0 @@ -/* - * Copyright (c) 2000-2005 Silicon Graphics, Inc. - * All Rights Reserved. - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License as - * published by the Free Software Foundation. - * - * This program is distributed in the hope that it would be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, see <http://www.gnu.org/licenses/>. - */ -#ifndef __COMMAND_H__ -#define __COMMAND_H__ - -#include "qemu-common.h" - -#define CMD_FLAG_GLOBAL ((int)0x80000000) /* don't iterate "args" */ - -extern BlockDriverState *qemuio_bs; - -typedef int (*cfunc_t)(BlockDriverState *bs, int argc, char **argv); -typedef void (*helpfunc_t)(void); - -typedef struct cmdinfo { - const char *name; - const char *altname; - cfunc_t cfunc; - int argmin; - int argmax; - int canpush; - int flags; - const char *args; - const char *oneline; - helpfunc_t help; -} cmdinfo_t; - -void qemuio_add_command(const cmdinfo_t *ci); - -int qemuio_command_usage(const cmdinfo_t *ci); - -bool qemuio_command(const char *cmd); - -#endif /* __COMMAND_H__ */ diff --git a/include/qemu-io.h b/include/qemu-io.h new file mode 100644 index 0000000..a418b46 --- /dev/null +++ b/include/qemu-io.h @@ -0,0 +1,46 @@ +/* + * Copyright (c) 2000-2005 Silicon Graphics, Inc. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef QEMU_IO_H +#define QEMU_IO_H + +#include "qemu-common.h" + +#define CMD_FLAG_GLOBAL ((int)0x80000000) /* don't iterate "args" */ + +typedef int (*cfunc_t)(BlockDriverState *bs, int argc, char **argv); +typedef void (*helpfunc_t)(void); + +typedef struct cmdinfo { + const char* name; + const char* altname; + cfunc_t cfunc; + int argmin; + int argmax; + int canpush; + int flags; + const char *args; + const char *oneline; + helpfunc_t help; +} cmdinfo_t; + +bool qemuio_command(BlockDriverState *bs, const char *cmd); + +void qemuio_add_command(const cmdinfo_t *ci); +int qemuio_command_usage(const cmdinfo_t *ci); + +#endif /* QEMU_IO_H */ diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index 82f9be6..87fddfa 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -8,9 +8,8 @@ * See the COPYING file in the top-level directory. */ -#include "qemu-common.h" +#include "qemu-io.h" #include "block/block_int.h" -#include "cmd.h" #define CMD_NOFILE_OK 0x01 @@ -50,11 +49,12 @@ static int init_check_command(BlockDriverState *bs, const cmdinfo_t *ct) return 1; } -static int command(const cmdinfo_t *ct, int argc, char **argv) +static int command(BlockDriverState *bs, const cmdinfo_t *ct, int argc, + char **argv) { char *cmd = argv[0]; - if (!init_check_command(qemuio_bs, ct)) { + if (!init_check_command(bs, ct)) { return 0; } @@ -75,7 +75,7 @@ static int command(const cmdinfo_t *ct, int argc, char **argv) return 0; } optind = 0; - return ct->cfunc(qemuio_bs, argc, argv); + return ct->cfunc(bs, argc, argv); } static const cmdinfo_t *find_command(const char *cmd) @@ -2068,7 +2068,7 @@ static const cmdinfo_t help_cmd = { .oneline = "help for one or all commands", }; -bool qemuio_command(const char *cmd) +bool qemuio_command(BlockDriverState *bs, const char *cmd) { char *input; const cmdinfo_t *ct; @@ -2081,7 +2081,7 @@ bool qemuio_command(const char *cmd) if (c) { ct = find_command(v[0]); if (ct) { - done = command(ct, c, v); + done = command(bs, ct, c, v); } else { fprintf(stderr, "command \"%s\" not found\n", v[0]); } diff --git a/qemu-io.c b/qemu-io.c index 3297f33..228567a 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -14,10 +14,9 @@ #include <getopt.h> #include <libgen.h> -#include "qemu-common.h" +#include "qemu-io.h" #include "qemu/main-loop.h" #include "block/block_int.h" -#include "cmd.h" #include "trace/control.h" #define VERSION "0.0.1" @@ -273,7 +272,7 @@ static void command_loop(void) char *input; for (i = 0; !done && i < ncmdline; i++) { - done = qemuio_command(cmdline[i]); + done = qemuio_command(qemuio_bs, cmdline[i]); } if (cmdline) { g_free(cmdline); @@ -298,7 +297,7 @@ static void command_loop(void) if (input == NULL) { break; } - done = qemuio_command(input); + done = qemuio_command(qemuio_bs, input); g_free(input); prompted = 0; -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 15/16] qemu-io: Use the qemu version for -V 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (13 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 14/16] qemu-io: Interface cleanup Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor Kevin Wolf 2013-06-05 13:02 ` [Qemu-devel] [PATCH 00/16] " Stefan Hajnoczi 16 siblings, 0 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino Always printing 0.0.1 and never updating the version number wasn't very useful. qemu-io is released with qemu, so using the same version number makes most sense. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-io.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/qemu-io.c b/qemu-io.c index 228567a..6d16192 100644 --- a/qemu-io.c +++ b/qemu-io.c @@ -19,8 +19,6 @@ #include "block/block_int.h" #include "trace/control.h" -#define VERSION "0.0.1" - #define CMD_NOFILE_OK 0x01 char *progname; @@ -380,7 +378,7 @@ int main(int argc, char **argv) } break; case 'V': - printf("%s version %s\n", progname, VERSION); + printf("%s version %s\n", progname, QEMU_VERSION); exit(0); case 'h': usage(progname); -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (14 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 15/16] qemu-io: Use the qemu version for -V Kevin Wolf @ 2013-05-28 15:27 ` Kevin Wolf 2013-05-28 16:07 ` Eric Blake 2013-06-05 13:02 ` [Qemu-devel] [PATCH 00/16] " Stefan Hajnoczi 16 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-05-28 15:27 UTC (permalink / raw) To: qemu-devel; +Cc: kwolf, pbonzini, stefanha, lcapitulino The QMP version is flagged with a __org.qemu.debug- prefix in order to reinforce the statement that qemu-io is for testing and debugging only, with no API guarantees. The HMP version is simply called 'qemu-io' for convenience. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- Makefile | 2 +- Makefile.objs | 1 + blockdev.c | 15 +++++++++++++++ hmp-commands.hx | 16 ++++++++++++++++ hmp.c | 10 ++++++++++ hmp.h | 1 + qapi-schema.json | 16 ++++++++++++++++ qmp-commands.hx | 28 ++++++++++++++++++++++++++++ 8 files changed, 88 insertions(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 581150f..2b85af8 100644 --- a/Makefile +++ b/Makefile @@ -186,7 +186,7 @@ qemu-img.o: qemu-img-cmds.h qemu-img$(EXESUF): qemu-img.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-nbd$(EXESUF): qemu-nbd.o $(block-obj-y) libqemuutil.a libqemustub.a -qemu-io$(EXESUF): qemu-io.o qemu-io-cmds.o $(block-obj-y) libqemuutil.a libqemustub.a +qemu-io$(EXESUF): qemu-io.o $(block-obj-y) libqemuutil.a libqemustub.a qemu-bridge-helper$(EXESUF): qemu-bridge-helper.o diff --git a/Makefile.objs b/Makefile.objs index 286ce06..5b288ba 100644 --- a/Makefile.objs +++ b/Makefile.objs @@ -13,6 +13,7 @@ block-obj-$(CONFIG_POSIX) += aio-posix.o block-obj-$(CONFIG_WIN32) += aio-win32.o block-obj-y += block/ block-obj-y += qapi-types.o qapi-visit.o +block-obj-y += qemu-io-cmds.o block-obj-y += qemu-coroutine.o qemu-coroutine-lock.o qemu-coroutine-io.o block-obj-y += qemu-coroutine-sleep.o diff --git a/blockdev.c b/blockdev.c index 789ad9f..9f9f920 100644 --- a/blockdev.c +++ b/blockdev.c @@ -43,6 +43,7 @@ #include "qmp-commands.h" #include "trace.h" #include "sysemu/arch_init.h" +#include "qemu-io.h" static QTAILQ_HEAD(drivelist, DriveInfo) drives = QTAILQ_HEAD_INITIALIZER(drives); extern QemuOptsList qemu_common_drive_opts; @@ -1604,6 +1605,20 @@ BlockJobInfoList *qmp_query_block_jobs(Error **errp) return dummy.next; } +void qmp___org_qemu_debug_qemu_io_command(const char *device, const char *cmd, + Error **errp) +{ + BlockDriverState *bs; + + bs = bdrv_find(device); + if (!bs) { + error_set(errp, QERR_DEVICE_NOT_FOUND, device); + return; + } + + qemuio_command(bs, cmd); +} + QemuOptsList qemu_common_drive_opts = { .name = "drive", .head = QTAILQ_HEAD_INITIALIZER(qemu_common_drive_opts.head), diff --git a/hmp-commands.hx b/hmp-commands.hx index 9cea415..9d4c04f 100644 --- a/hmp-commands.hx +++ b/hmp-commands.hx @@ -1551,6 +1551,22 @@ Removes the chardev @var{id}. ETEXI { + .name = "qemu-io", + .args_type = "device:B,command:s", + .params = "[device] \"[command]\"", + .help = "run a qemu-io command on a block device", + .mhandler.cmd = hmp_qemu_io, + }, + +STEXI +@item chardev_remove id +@findex chardev_remove + +Removes the chardev @var{id}. + +ETEXI + + { .name = "info", .args_type = "item:s?", .params = "[subcommand]", diff --git a/hmp.c b/hmp.c index 4fb76ec..9cea8d8 100644 --- a/hmp.c +++ b/hmp.c @@ -1425,3 +1425,13 @@ void hmp_chardev_remove(Monitor *mon, const QDict *qdict) qmp_chardev_remove(qdict_get_str(qdict, "id"), &local_err); hmp_handle_error(mon, &local_err); } + +void hmp_qemu_io(Monitor *mon, const QDict *qdict) +{ + Error *local_err = NULL; + const char* device = qdict_get_str(qdict, "device"); + const char* command = qdict_get_str(qdict, "command"); + + qmp___org_qemu_debug_qemu_io_command(device, command, &local_err); + hmp_handle_error(mon, &local_err); +} diff --git a/hmp.h b/hmp.h index 95fe76e..56d2e92 100644 --- a/hmp.h +++ b/hmp.h @@ -85,5 +85,6 @@ void hmp_nbd_server_add(Monitor *mon, const QDict *qdict); void hmp_nbd_server_stop(Monitor *mon, const QDict *qdict); void hmp_chardev_add(Monitor *mon, const QDict *qdict); void hmp_chardev_remove(Monitor *mon, const QDict *qdict); +void hmp_qemu_io(Monitor *mon, const QDict *qdict); #endif diff --git a/qapi-schema.json b/qapi-schema.json index ef1f657..cc4a083 100644 --- a/qapi-schema.json +++ b/qapi-schema.json @@ -3618,3 +3618,19 @@ '*cpuid-input-ecx': 'int', 'cpuid-register': 'X86CPURegister32', 'features': 'int' } } + +## +# @__org.qemu.debug-qemu-io-command +# +# Execute a qemu-io command +# +# @device: The block device on which the command should be executed +# +# @command: The command to execute +# +# Returns: Nothing on success +# +# Since: 1.6 (testing and debugging use only, no API stability) +## +{ 'command': '__org.qemu.debug-qemu-io-command', + 'data': {'device': 'str', 'command': 'str'} } diff --git a/qmp-commands.hx b/qmp-commands.hx index ffd130e..8682bea 100644 --- a/qmp-commands.hx +++ b/qmp-commands.hx @@ -2932,3 +2932,31 @@ Example: <- { "return": {} } EQMP + + { + .name = "__org.qemu.debug-qemu-io-command", + .args_type = "device:s,command:s", + .mhandler.cmd_new = qmp_marshal_input___org_qemu_debug_qemu_io_command, + }, + + +SQMP +__org.qemu.debug-qemu-io-command +-------------------------------- + +Execute a qemu-io command. Debugging and testing use only. This API is not +stable. + +Arguments: + +- "device": The block device on which the command should be executed + (json-string) +- "command": The command to execute (json-string) + +Example: + +-> { "execute": "__org.qemu.debug-qemu-io-command", + "arguments": { "device" : "ide0-hd0", "command" : "flush" } } +<- { "return": {} } + +EQMP -- 1.8.1.4 ^ permalink raw reply related [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-28 15:27 ` [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor Kevin Wolf @ 2013-05-28 16:07 ` Eric Blake 2013-05-29 8:13 ` Kevin Wolf 0 siblings, 1 reply; 45+ messages in thread From: Eric Blake @ 2013-05-28 16:07 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 2841 bytes --] On 05/28/2013 09:27 AM, Kevin Wolf wrote: > The QMP version is flagged with a __org.qemu.debug- prefix in order to > reinforce the statement that qemu-io is for testing and debugging only, > with no API guarantees. Correct use of naming conventions. Hmm, I wonder if the recent addition of an 'abort' action to 'transaction' should be renamed __org.qemu.debug-abort, to make it obvious that it is another case of a QMP command useful mainly for testing, and not real-life use. > > The HMP version is simply called 'qemu-io' for convenience. > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > --- > +++ b/hmp-commands.hx > @@ -1551,6 +1551,22 @@ Removes the chardev @var{id}. > ETEXI > > { > + .name = "qemu-io", > + .args_type = "device:B,command:s", > + .params = "[device] \"[command]\"", > + .help = "run a qemu-io command on a block device", > + .mhandler.cmd = hmp_qemu_io, > + }, > + > +STEXI > +@item chardev_remove id > +@findex chardev_remove > + > +Removes the chardev @var{id}. Is this the right documentation? > +++ b/qapi-schema.json > @@ -3618,3 +3618,19 @@ > '*cpuid-input-ecx': 'int', > 'cpuid-register': 'X86CPURegister32', > 'features': 'int' } } > + > +## > +# @__org.qemu.debug-qemu-io-command > +# > +# Execute a qemu-io command > +# > +# @device: The block device on which the command should be executed > +# > +# @command: The command to execute > +# > +# Returns: Nothing on success > +# > +# Since: 1.6 (testing and debugging use only, no API stability) > +## > +{ 'command': '__org.qemu.debug-qemu-io-command', > + 'data': {'device': 'str', 'command': 'str'} } I would have asked that 'command' be an enum, except that we promise no API stability so there's no need to lock us into an enum that could be introspected :) This command looks fine as-is. > diff --git a/qmp-commands.hx b/qmp-commands.hx > index ffd130e..8682bea 100644 > --- a/qmp-commands.hx > +++ b/qmp-commands.hx > @@ -2932,3 +2932,31 @@ Example: > <- { "return": {} } > > EQMP > + > + { > + .name = "__org.qemu.debug-qemu-io-command", > + .args_type = "device:s,command:s", > + .mhandler.cmd_new = qmp_marshal_input___org_qemu_debug_qemu_io_command, > + }, > + > + > +SQMP > +__org.qemu.debug-qemu-io-command > +-------------------------------- Do we really need to have SQMP documentation of the command, or is listing of its handler sufficient? In other words, if this is debug-only, can we just leave it undocumented instead of listing an example only to declare the example non-portable? -- 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-28 16:07 ` Eric Blake @ 2013-05-29 8:13 ` Kevin Wolf 2013-05-29 12:05 ` Eric Blake 2013-05-29 17:51 ` Luiz Capitulino 0 siblings, 2 replies; 45+ messages in thread From: Kevin Wolf @ 2013-05-29 8:13 UTC (permalink / raw) To: Eric Blake; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > reinforce the statement that qemu-io is for testing and debugging only, > > with no API guarantees. > > Correct use of naming conventions. > > Hmm, I wonder if the recent addition of an 'abort' action to > 'transaction' should be renamed __org.qemu.debug-abort, to make it > obvious that it is another case of a QMP command useful mainly for > testing, and not real-life use. Makes sense to me. But first I'd like to get Luiz's ack for this, because I think I'm the first one to use an __org.qemu prefix, and I'm the first one trying to introduce a QMP command without API stability. > > > > The HMP version is simply called 'qemu-io' for convenience. > > > > Signed-off-by: Kevin Wolf <kwolf@redhat.com> > > --- > > > +++ b/hmp-commands.hx > > @@ -1551,6 +1551,22 @@ Removes the chardev @var{id}. > > ETEXI > > > > { > > + .name = "qemu-io", > > + .args_type = "device:B,command:s", > > + .params = "[device] \"[command]\"", > > + .help = "run a qemu-io command on a block device", > > + .mhandler.cmd = hmp_qemu_io, > > + }, > > + > > +STEXI > > +@item chardev_remove id > > +@findex chardev_remove > > + > > +Removes the chardev @var{id}. > > Is this the right documentation? Obviously not. I find these files where you declare monitor commands horribly confusing, and this is the result... > > +++ b/qapi-schema.json > > @@ -3618,3 +3618,19 @@ > > '*cpuid-input-ecx': 'int', > > 'cpuid-register': 'X86CPURegister32', > > 'features': 'int' } } > > + > > +## > > +# @__org.qemu.debug-qemu-io-command > > +# > > +# Execute a qemu-io command > > +# > > +# @device: The block device on which the command should be executed > > +# > > +# @command: The command to execute > > +# > > +# Returns: Nothing on success > > +# > > +# Since: 1.6 (testing and debugging use only, no API stability) > > +## > > +{ 'command': '__org.qemu.debug-qemu-io-command', > > + 'data': {'device': 'str', 'command': 'str'} } > > I would have asked that 'command' be an enum, except that we promise no > API stability so there's no need to lock us into an enum that could be > introspected :) This command looks fine as-is. Wouldn't even work because the command includes parameters and things. This is really what you would type on the qemu-io command line. So a typical example might look like: { "execute": "__org.qemu.debug-qemu-io-command", "arguments": { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } There's nothing structured about it, but for testing and debugging this should be good enough. > > diff --git a/qmp-commands.hx b/qmp-commands.hx > > index ffd130e..8682bea 100644 > > --- a/qmp-commands.hx > > +++ b/qmp-commands.hx > > @@ -2932,3 +2932,31 @@ Example: > > <- { "return": {} } > > > > EQMP > > + > > + { > > + .name = "__org.qemu.debug-qemu-io-command", > > + .args_type = "device:s,command:s", > > + .mhandler.cmd_new = qmp_marshal_input___org_qemu_debug_qemu_io_command, > > + }, > > + > > + > > +SQMP > > +__org.qemu.debug-qemu-io-command > > +-------------------------------- > > Do we really need to have SQMP documentation of the command, or is > listing of its handler sufficient? In other words, if this is > debug-only, can we just leave it undocumented instead of listing an > example only to declare the example non-portable? I don't mind either way. Luiz? Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-29 8:13 ` Kevin Wolf @ 2013-05-29 12:05 ` Eric Blake 2013-05-29 17:51 ` Luiz Capitulino 1 sibling, 0 replies; 45+ messages in thread From: Eric Blake @ 2013-05-29 12:05 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino [-- Attachment #1: Type: text/plain, Size: 2567 bytes --] On 05/29/2013 02:13 AM, Kevin Wolf wrote: > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: >> On 05/28/2013 09:27 AM, Kevin Wolf wrote: >>> The QMP version is flagged with a __org.qemu.debug- prefix in order to >>> reinforce the statement that qemu-io is for testing and debugging only, >>> with no API guarantees. >> >> Correct use of naming conventions. >> >> Hmm, I wonder if the recent addition of an 'abort' action to >> 'transaction' should be renamed __org.qemu.debug-abort, to make it >> obvious that it is another case of a QMP command useful mainly for >> testing, and not real-life use. > > Makes sense to me. > > But first I'd like to get Luiz's ack for this, because I think I'm the > first one to use an __org.qemu prefix, and I'm the first one trying to > introduce a QMP command without API stability. qom-get/qom-set is somewhat of an unstable API (not that the parameters of the call will change, but that the underlying qom model is not locked down, and therefore you cannot blindly rely on a qom command that worked in one version to continue working in future versions). [Hmm, do we have a good way to introspect the qom model?] But yes, this is indeed a different class of unstable interface, and worth agreeing that we have best practice in mind before committing to it. >>> +{ 'command': '__org.qemu.debug-qemu-io-command', >>> + 'data': {'device': 'str', 'command': 'str'} } >> >> I would have asked that 'command' be an enum, except that we promise no >> API stability so there's no need to lock us into an enum that could be >> introspected :) This command looks fine as-is. > > Wouldn't even work because the command includes parameters and things. > This is really what you would type on the qemu-io command line. So a > typical example might look like: > > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } > > There's nothing structured about it, but for testing and debugging this > should be good enough. I would suggest that you consider using 'command':['str'], so that at least you have structure that allows you to pass separate arguments (and thus pass an argument with spaces in the future, should the need arise): { "execute": "__org.qemu.debug-qemu-io-command", "arguments": { "device": "ide0-hd0", "command": [ "write", "-P", "0x12", "4M", "512k" ] } } -- 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: 621 bytes --] ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-29 8:13 ` Kevin Wolf 2013-05-29 12:05 ` Eric Blake @ 2013-05-29 17:51 ` Luiz Capitulino 2013-06-04 10:08 ` Kevin Wolf 1 sibling, 1 reply; 45+ messages in thread From: Luiz Capitulino @ 2013-05-29 17:51 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha On Wed, 29 May 2013 10:13:42 +0200 Kevin Wolf <kwolf@redhat.com> wrote: > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > > reinforce the statement that qemu-io is for testing and debugging only, > > > with no API guarantees. > > > > Correct use of naming conventions. > > > > Hmm, I wonder if the recent addition of an 'abort' action to > > 'transaction' should be renamed __org.qemu.debug-abort, to make it > > obvious that it is another case of a QMP command useful mainly for > > testing, and not real-life use. > > Makes sense to me. > > But first I'd like to get Luiz's ack for this, because I think I'm the > first one to use an __org.qemu prefix, and I'm the first one trying to > introduce a QMP command without API stability. I think that should be fine. However, it's not a perfect match for QMP as you don't expect mngt to use it anytime soon and that the command's syntax is not QMP friendly: > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } What about adding a HMP-only command (the good old way) and use it through human-monitor-command? IMO, this matches better your current use-case and the API instability of the command. ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-05-29 17:51 ` Luiz Capitulino @ 2013-06-04 10:08 ` Kevin Wolf 2013-06-04 12:40 ` Luiz Capitulino 0 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-06-04 10:08 UTC (permalink / raw) To: Luiz Capitulino; +Cc: pbonzini, qemu-devel, stefanha Am 29.05.2013 um 19:51 hat Luiz Capitulino geschrieben: > On Wed, 29 May 2013 10:13:42 +0200 > Kevin Wolf <kwolf@redhat.com> wrote: > > > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > > > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > > > reinforce the statement that qemu-io is for testing and debugging only, > > > > with no API guarantees. > > > > > > Correct use of naming conventions. > > > > > > Hmm, I wonder if the recent addition of an 'abort' action to > > > 'transaction' should be renamed __org.qemu.debug-abort, to make it > > > obvious that it is another case of a QMP command useful mainly for > > > testing, and not real-life use. > > > > Makes sense to me. > > > > But first I'd like to get Luiz's ack for this, because I think I'm the > > first one to use an __org.qemu prefix, and I'm the first one trying to > > introduce a QMP command without API stability. > > I think that should be fine. However, it's not a perfect match for QMP > as you don't expect mngt to use it anytime soon and that the command's > syntax is not QMP friendly: > > > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } > > What about adding a HMP-only command (the good old way) and use it > through human-monitor-command? > > IMO, this matches better your current use-case and the API instability > of the command. Works for me, but wasn't the plan to make HMP purely a wrapper around QMP? Then adding HMP-only commands would be counterproductive. So I assumed that QMP is a must. I didn't even know that the code still allows you to have HMP-only commands. :-) So you prefer a respin with the QMP part dropped? Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-06-04 10:08 ` Kevin Wolf @ 2013-06-04 12:40 ` Luiz Capitulino 2013-06-04 12:49 ` Kevin Wolf 0 siblings, 1 reply; 45+ messages in thread From: Luiz Capitulino @ 2013-06-04 12:40 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha On Tue, 4 Jun 2013 12:08:23 +0200 Kevin Wolf <kwolf@redhat.com> wrote: > Am 29.05.2013 um 19:51 hat Luiz Capitulino geschrieben: > > On Wed, 29 May 2013 10:13:42 +0200 > > Kevin Wolf <kwolf@redhat.com> wrote: > > > > > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > > > > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > > > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > > > > reinforce the statement that qemu-io is for testing and debugging only, > > > > > with no API guarantees. > > > > > > > > Correct use of naming conventions. > > > > > > > > Hmm, I wonder if the recent addition of an 'abort' action to > > > > 'transaction' should be renamed __org.qemu.debug-abort, to make it > > > > obvious that it is another case of a QMP command useful mainly for > > > > testing, and not real-life use. > > > > > > Makes sense to me. > > > > > > But first I'd like to get Luiz's ack for this, because I think I'm the > > > first one to use an __org.qemu prefix, and I'm the first one trying to > > > introduce a QMP command without API stability. > > > > I think that should be fine. However, it's not a perfect match for QMP > > as you don't expect mngt to use it anytime soon and that the command's > > syntax is not QMP friendly: > > > > > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > > > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } > > > > What about adding a HMP-only command (the good old way) and use it > > through human-monitor-command? > > > > IMO, this matches better your current use-case and the API instability > > of the command. > > Works for me, but wasn't the plan to make HMP purely a wrapper around > QMP? Then adding HMP-only commands would be counterproductive. So I > assumed that QMP is a must. I didn't even know that the code still > allows you to have HMP-only commands. :-) Yes, the long term plan is to have all HMP commands calling QMP counterparts. But the command you're adding doesn't fit QMP's design very well. I suggested adding it as HMP-only for now because it's the simplest thing to do for this very specific case. If more test-only non-QMP-friendly commands appear, then we'll need to think of a more general solution. Now, something has just occurred to me. Why isn't it a good idea having this command as a stable API? Wouldn't it be a good idea to allow out-of-tree test tools like autotest to use it? > So you prefer a respin with the QMP part dropped? Yes. ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-06-04 12:40 ` Luiz Capitulino @ 2013-06-04 12:49 ` Kevin Wolf 2013-06-04 13:09 ` Luiz Capitulino 0 siblings, 1 reply; 45+ messages in thread From: Kevin Wolf @ 2013-06-04 12:49 UTC (permalink / raw) To: Luiz Capitulino; +Cc: pbonzini, qemu-devel, stefanha Am 04.06.2013 um 14:40 hat Luiz Capitulino geschrieben: > On Tue, 4 Jun 2013 12:08:23 +0200 > Kevin Wolf <kwolf@redhat.com> wrote: > > > Am 29.05.2013 um 19:51 hat Luiz Capitulino geschrieben: > > > On Wed, 29 May 2013 10:13:42 +0200 > > > Kevin Wolf <kwolf@redhat.com> wrote: > > > > > > > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > > > > > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > > > > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > > > > > reinforce the statement that qemu-io is for testing and debugging only, > > > > > > with no API guarantees. > > > > > > > > > > Correct use of naming conventions. > > > > > > > > > > Hmm, I wonder if the recent addition of an 'abort' action to > > > > > 'transaction' should be renamed __org.qemu.debug-abort, to make it > > > > > obvious that it is another case of a QMP command useful mainly for > > > > > testing, and not real-life use. > > > > > > > > Makes sense to me. > > > > > > > > But first I'd like to get Luiz's ack for this, because I think I'm the > > > > first one to use an __org.qemu prefix, and I'm the first one trying to > > > > introduce a QMP command without API stability. > > > > > > I think that should be fine. However, it's not a perfect match for QMP > > > as you don't expect mngt to use it anytime soon and that the command's > > > syntax is not QMP friendly: > > > > > > > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > > > > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } > > > > > > What about adding a HMP-only command (the good old way) and use it > > > through human-monitor-command? > > > > > > IMO, this matches better your current use-case and the API instability > > > of the command. > > > > Works for me, but wasn't the plan to make HMP purely a wrapper around > > QMP? Then adding HMP-only commands would be counterproductive. So I > > assumed that QMP is a must. I didn't even know that the code still > > allows you to have HMP-only commands. :-) > > Yes, the long term plan is to have all HMP commands calling QMP > counterparts. But the command you're adding doesn't fit QMP's design > very well. It fits the design about as well as human-monitor-command. Both are passthrough commands, one to HMP, the other one to qemu-io. > I suggested adding it as HMP-only for now because it's the simplest > thing to do for this very specific case. > > If more test-only non-QMP-friendly commands appear, then we'll need > to think of a more general solution. What does "for now" mean, and what will the proper long-term solution look like? > Now, something has just occurred to me. Why isn't it a good idea having > this command as a stable API? Wouldn't it be a good idea to allow > out-of-tree test tools like autotest to use it? Because qemu-io exposes internals of the block layer that we don't want to make a stable API. Autotest can have test cases using this, but they'd have to be against a specific qemu version. > > So you prefer a respin with the QMP part dropped? > > Yes. Okay, I'll do that. Kevin ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor 2013-06-04 12:49 ` Kevin Wolf @ 2013-06-04 13:09 ` Luiz Capitulino 0 siblings, 0 replies; 45+ messages in thread From: Luiz Capitulino @ 2013-06-04 13:09 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha On Tue, 4 Jun 2013 14:49:55 +0200 Kevin Wolf <kwolf@redhat.com> wrote: > Am 04.06.2013 um 14:40 hat Luiz Capitulino geschrieben: > > On Tue, 4 Jun 2013 12:08:23 +0200 > > Kevin Wolf <kwolf@redhat.com> wrote: > > > > > Am 29.05.2013 um 19:51 hat Luiz Capitulino geschrieben: > > > > On Wed, 29 May 2013 10:13:42 +0200 > > > > Kevin Wolf <kwolf@redhat.com> wrote: > > > > > > > > > Am 28.05.2013 um 18:07 hat Eric Blake geschrieben: > > > > > > On 05/28/2013 09:27 AM, Kevin Wolf wrote: > > > > > > > The QMP version is flagged with a __org.qemu.debug- prefix in order to > > > > > > > reinforce the statement that qemu-io is for testing and debugging only, > > > > > > > with no API guarantees. > > > > > > > > > > > > Correct use of naming conventions. > > > > > > > > > > > > Hmm, I wonder if the recent addition of an 'abort' action to > > > > > > 'transaction' should be renamed __org.qemu.debug-abort, to make it > > > > > > obvious that it is another case of a QMP command useful mainly for > > > > > > testing, and not real-life use. > > > > > > > > > > Makes sense to me. > > > > > > > > > > But first I'd like to get Luiz's ack for this, because I think I'm the > > > > > first one to use an __org.qemu prefix, and I'm the first one trying to > > > > > introduce a QMP command without API stability. > > > > > > > > I think that should be fine. However, it's not a perfect match for QMP > > > > as you don't expect mngt to use it anytime soon and that the command's > > > > syntax is not QMP friendly: > > > > > > > > > { "execute": "__org.qemu.debug-qemu-io-command", "arguments": > > > > > { "device": "ide0-hd0", "command": "write -P 0x12 4M 512k" } } > > > > > > > > What about adding a HMP-only command (the good old way) and use it > > > > through human-monitor-command? > > > > > > > > IMO, this matches better your current use-case and the API instability > > > > of the command. > > > > > > Works for me, but wasn't the plan to make HMP purely a wrapper around > > > QMP? Then adding HMP-only commands would be counterproductive. So I > > > assumed that QMP is a must. I didn't even know that the code still > > > allows you to have HMP-only commands. :-) > > > > Yes, the long term plan is to have all HMP commands calling QMP > > counterparts. But the command you're adding doesn't fit QMP's design > > very well. > > It fits the design about as well as human-monitor-command. Both are > passthrough commands, one to HMP, the other one to qemu-io. Yes, but human-monitor-command is a stop gap command meant to be an exception. It's usage should decrease over time. I'd expect the opposite for qemu-io-command. > > I suggested adding it as HMP-only for now because it's the simplest > > thing to do for this very specific case. > > > > If more test-only non-QMP-friendly commands appear, then we'll need > > to think of a more general solution. > > What does "for now" mean, and what will the proper long-term solution > look like? For now means: until we know what the proper solution should be. I Honestly don't know what's the best/proper solution here, that's why I suggested having this in HMP first, so that we don't commit ourselves to anything prematurely. > > Now, something has just occurred to me. Why isn't it a good idea having > > this command as a stable API? Wouldn't it be a good idea to allow > > out-of-tree test tools like autotest to use it? > > Because qemu-io exposes internals of the block layer that we don't want > to make a stable API. What kind of internals? In the example you posted I see just a generic enough write command. > Autotest can have test cases using this, but > they'd have to be against a specific qemu version. > > > > So you prefer a respin with the QMP part dropped? > > > > Yes. > > Okay, I'll do that. > > Kevin > ^ permalink raw reply [flat|nested] 45+ messages in thread
* Re: [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf ` (15 preceding siblings ...) 2013-05-28 15:27 ` [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor Kevin Wolf @ 2013-06-05 13:02 ` Stefan Hajnoczi 16 siblings, 0 replies; 45+ messages in thread From: Stefan Hajnoczi @ 2013-06-05 13:02 UTC (permalink / raw) To: Kevin Wolf; +Cc: pbonzini, qemu-devel, stefanha, lcapitulino On Tue, May 28, 2013 at 05:27:20PM +0200, Kevin Wolf wrote: > This is a prerequisite for some kind of tests. It involves reorganising the > qemu-io code so that the command part can be separated and doesn't pollute the > global namespace any more, so we can link it with qemu. > > Kevin Wolf (16): > qemu-io: Remove unused args_command > cutils: Support 'P' and 'E' suffixes in strtosz() > qemu-io: Make cvtnum() a wrapper around strtosz_suffix() > qemu-io: Handle cvtnum() errors in 'alloc' > qemu-io: Don't use global bs in command implementations > qemu-io: Split off commands to qemu-io-cmds.c > qemu-io: Factor out qemuio_command > qemu-io: Move 'help' function > qemu-io: Move 'quit' function > qemu-io: Move qemu_strsep() to cutils.c > qemu-io: Move functions for registering and running commands > qemu-io: Move command_loop() and friends > qemu-io: Move remaining helpers from cmd.c > qemu-io: Interface cleanup > qemu-io: Use the qemu version for -V > Make qemu-io commands available in the monitor > > Makefile | 2 +- > Makefile.objs | 1 + > blockdev.c | 15 + > cmd.c | 612 ------------- > cmd.h | 79 -- > hmp-commands.hx | 16 + > hmp.c | 10 + > hmp.h | 1 + > include/qemu-common.h | 3 + > include/qemu-io.h | 46 + > monitor.c | 8 +- > qapi-schema.json | 16 + > qemu-img.c | 10 +- > qemu-io-cmds.c | 2118 ++++++++++++++++++++++++++++++++++++++++++++ > qemu-io.c | 1988 ++++------------------------------------- > qmp-commands.hx | 28 + > tests/qemu-iotests/049.out | 8 +- > util/cutils.c | 25 + > 18 files changed, 2466 insertions(+), 2520 deletions(-) > delete mode 100644 cmd.c > delete mode 100644 cmd.h > create mode 100644 include/qemu-io.h > create mode 100644 qemu-io-cmds.c Mostly good but I think cmd.c should be preserved. It's a generic command-line dispatcher and shouldn't be squashed into qemu-io-cmds.c. ^ permalink raw reply [flat|nested] 45+ messages in thread
end of thread, other threads:[~2013-06-05 13:02 UTC | newest] Thread overview: 45+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2013-05-28 15:27 [Qemu-devel] [PATCH 00/16] Make qemu-io commands available in the monitor Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 01/16] qemu-io: Remove unused args_command Kevin Wolf 2013-05-28 16:18 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 02/16] cutils: Support 'P' and 'E' suffixes in strtosz() Kevin Wolf 2013-05-28 16:29 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 03/16] qemu-io: Make cvtnum() a wrapper around strtosz_suffix() Kevin Wolf 2013-05-28 16:42 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 04/16] qemu-io: Handle cvtnum() errors in 'alloc' Kevin Wolf 2013-05-28 16:53 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 05/16] qemu-io: Don't use global bs in command implementations Kevin Wolf 2013-05-28 15:37 ` Kevin Wolf 2013-05-28 17:02 ` Eric Blake 2013-06-05 12:28 ` Stefan Hajnoczi 2013-06-05 12:39 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 06/16] qemu-io: Split off commands to qemu-io-cmds.c Kevin Wolf 2013-05-29 20:29 ` Eric Blake 2013-06-04 10:11 ` Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 07/16] qemu-io: Factor out qemuio_command Kevin Wolf 2013-05-29 21:11 ` Eric Blake 2013-06-05 12:33 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 08/16] qemu-io: Move 'help' function Kevin Wolf 2013-05-29 21:25 ` Eric Blake 2013-06-04 10:20 ` Kevin Wolf 2013-06-05 12:37 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 09/16] qemu-io: Move 'quit' function Kevin Wolf 2013-05-29 22:11 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 10/16] qemu-io: Move qemu_strsep() to cutils.c Kevin Wolf 2013-05-29 22:12 ` Eric Blake 2013-05-28 15:27 ` [Qemu-devel] [PATCH 11/16] qemu-io: Move functions for registering and running commands Kevin Wolf 2013-05-30 2:27 ` Eric Blake 2013-06-05 12:42 ` Stefan Hajnoczi 2013-05-28 15:27 ` [Qemu-devel] [PATCH 12/16] qemu-io: Move command_loop() and friends Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 13/16] qemu-io: Move remaining helpers from cmd.c Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 14/16] qemu-io: Interface cleanup Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 15/16] qemu-io: Use the qemu version for -V Kevin Wolf 2013-05-28 15:27 ` [Qemu-devel] [PATCH 16/16] Make qemu-io commands available in the monitor Kevin Wolf 2013-05-28 16:07 ` Eric Blake 2013-05-29 8:13 ` Kevin Wolf 2013-05-29 12:05 ` Eric Blake 2013-05-29 17:51 ` Luiz Capitulino 2013-06-04 10:08 ` Kevin Wolf 2013-06-04 12:40 ` Luiz Capitulino 2013-06-04 12:49 ` Kevin Wolf 2013-06-04 13:09 ` Luiz Capitulino 2013-06-05 13:02 ` [Qemu-devel] [PATCH 00/16] " Stefan Hajnoczi
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.