qemu-io: Rework alloc command
PATCH 1/1 qemu-io: Rework alloc command
The alloc command in qemu-io is mostly useless currently. Instead of doing a single call to bdrv_is_allocated, we must call bdrv_is_allocated in a loop until we have found out for each requested sector if it is allocated or not (bdrv_is_allocated returns a number of sectors that are known to be in the same state as the first one, but it is not required to include all of them)
This changes the output format of the alloc command so that a change to the expected output of qemu-iotests 019 is necessary once this is included.
--- qemu-io.c | 22 ++++++++++++++-------- 1 files changed, 14 insertions(+), 8 deletions(-)
qemu-io.c +14 -8
@@ -1170,11 +1170,10 @@ static int alloc_f(int argc, char **argv) { int64_t offset; -int nb_sectors; +int nb_sectors, remaining;char s1[64 ]; -int num; +int num, sum_alloc;int ret; -const char *retstr; offset =cvtnum (argv[1 ]);if (offset &0x1ff ) {
@@ -1188,16 +1187,23 @@ alloc_f(int argc, char **argv)else nb_sectors =1 ; - ret =bdrv_is_allocated (bs, offset >>9 , nb_sectors, &num); + remaining = nb_sectors; + sum_alloc =0 ; +while (remaining) { + ret =bdrv_is_allocated (bs, offset >>9 , nb_sectors, &num); + remaining -= num; +if (ret) { + sum_alloc += num; + } + }cvtstr (offset, s1,sizeof (s1)); - retstr = ret ?"allocated" :"not allocated" ;if (nb_sectors ==1 ) -printf ("sector %s at offset %s\n" , retstr, s1); +printf ("sector allocated at offset %s\n" , s1);else -printf ("%d/%d sectors %s at offset %s\n" , - num, nb_sectors, retstr, s1); +printf ("%d/%d sectors allocated at offset %s\n" , + sum_alloc, nb_sectors, s1);return 0 ; }
Signature · 1 line
1.6.0.6
Quoted 8 lines · Kevin Wolf
The alloc command in qemu-io is mostly useless currently. Instead of doing a single call to
bdrv_is_allocated, we must callbdrv_is_allocatedin a loop until we have found out for each requested sector if it is allocated or not (bdrv_is_allocatedreturns a number of sectors that are known to be in the same state as the first one, but it is not required to include all of them)This changes the output format of the alloc command so that a change to the expected output of qemu-iotests 019 is necessary once this is included.
Looks good to me. Can you post the corresponding qemu-iotests patch, too?
Quoted 7 lines · Christoph Hellwig
Quoted 6 lines · Kevin Wolf
The alloc command in qemu-io is mostly useless currently. Instead of doing a single call to
bdrv_is_allocated, we must callbdrv_is_allocatedin a loop until we have found out for each requested sector if it is allocated or not (bdrv_is_allocatedreturns a number of sectors that are known to be in the same state as the first one, but it is not required to include all of them)
Quoted 5 lines
Quoted 2 lines
This changes the output format of the alloc command so that a change to the expected output of qemu-iotests 019 is necessary once this is included.
Looks good to me. Can you post the corresponding qemu-iotests patch, too?
I wanted to post it when the qemu-io change is in, but if you want to apply it earlier (makes qcow2 and vmdk fail until the qemu-io patch is committed), here is it.
Kevin
Quoted 1 line
From
41f80a74b4177e5396dab7916f8f7b40a65cef07Mon Sep 17 00:00:00 2001
From: Kevin Wolf <[email protected]>
Date: Tue, 21 Jul 2009 09:16:37 +0200
Subject: [PATCH] qemu-iotests: Fix expected result for 019 after qemu-io change
The output of the alloc command in qemu-io has changed.
--- 019.out | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
019.out +2 -2
@@ -539,8 +539,8 @@ qemu-io> wrote 65536/65536 bytes at offset 4296146944 qemu-io> No errors were found on the image. Checking if backing clusters are allocated when they shouldn't -qemu-io> 128/128 sectors not allocated at offset 1 MiB -qemu-io> qemu-io> 128/128 sectors not allocated at offset 4.001 GiB +qemu-io> 0/128 sectors allocated at offset 1 MiB +qemu-io> qemu-io> 0/128 sectors allocated at offset 4.001 GiB qemu-io> Reading === IO: pattern 0
Signature · 1 line
1.6.0.6
Quoted 3 lines · Kevin Wolf
I wanted to post it when the qemu-io change is in, but if you want to apply it earlier (makes qcow2 and vmdk fail until the qemu-io patch is committed), here is it.
Thanks, I've put it in now that the qemu-io change went in.