Btrfs-progs: fix parse_limit function to return errors when parsing unit
PATCH 1/1 Btrfs-progs: fix parse_limit function to return errors when parsing unit
From: Wang Shilong <[email protected]>
Steps to reproduce:
btrfs qgroup limit m <mnt>/subv
Here, unit(k/K/g/G/m/M/t/T) all will trigger the problem. For the above command, the original code will parse the limit value as 0 and return successfully.It is wrong,fix it.
--- cmds-qgroup.c | 9 ++++++++- 1 files changed, 8 insertions(+), 1 deletions(-)
cmds-qgroup.c +8 -1
@@ -214,6 +214,8 @@ static int parse_limit(const char *p, unsigned long long *s) {char *endptr;unsigned long long size; +int len =strlen (p); +int unit =0 ;if (strcasecmp (p,"none" ) ==0 ) { *s =0 ;
@@ -232,6 +234,7 @@ static int parse_limit(const char *p, unsigned long long *s) size *=1024 ;case 'K' :case 'k' : + unit =1 ; size *=1024 ; ++endptr;break ;
@@ -241,7 +244,11 @@ static int parse_limit(const char *p, unsigned long long *s)return 0 ; } -if (*endptr) +/* +* (len==1 && unit) means +* limit passed is like k,K,m,M,g,G,t,T. +*/ +if (*endptr || (len ==1 && unit))return 0 ; *s = size;
Signature · 1 line
1.7.7.6
Quoted 8 lines · Wang Shilong
From: Wang Shilong <[email protected]>
Steps to reproduce:
btrfs qgroup limit m <mnt>/subvHere, unit(k/K/g/G/m/M/t/T) all will trigger the problem. For the above command, the original code will parse the limit value as 0 and return successfully.It is wrong,fix it.
I don't think we should allow to accept bare units for specifing numbers, so in that case fail with 'unrecognized limit'.
david
Quoted 6 lines · David Sterba
Quoted 5 lines · Wang Shilong
From: Wang Shilong <[email protected]>
Steps to reproduce:
btrfs qgroup limit m <mnt>/subv
Quoted 6 lines
Quoted 3 lines
Here, unit(k/K/g/G/m/M/t/T) all will trigger the problem. For the above command, the original code will parse the limit value as 0 and return successfully.It is wrong,fix it.
I don't think we should allow to accept bare units for specifing numbers, so in that case fail with 'unrecognized limit'.
I agree with that.
Maybe since David didn't quite want to take my patch 01/17, and Zach thought we needed better code for it all, and Goffredo didn't like the exit() from parse_size, and now this problem, we need a single, concerted effort to fix up size parsing.
Who knew it was so hard? :)
-Eric
Quoted 6 lines
david
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hello, Eric, David
2013/2/27 Eric Sandeen <[email protected]>:
Quoted 14 lines · David Sterba
Quoted 6 lines · Wang Shilong
Quoted 5 lines
From: Wang Shilong <[email protected]>
Steps to reproduce:
btrfs qgroup limit m <mnt>/subvQuoted 6 lines
Quoted 3 lines
Here, unit(k/K/g/G/m/M/t/T) all will trigger the problem. For the above command, the original code will parse the limit value as 0 and return successfully.It is wrong,fix it.
I don't think we should allow to accept bare units for specifing numbers, so in that case fail with 'unrecognized limit'.
Quoted 6 lines
I agree with that.
Maybe since David didn't quite want to take my patch 01/17, and Zach thought we needed better code for it all, and Goffredo didn't like the
exit()fromparse_size, and now this problem, we need a single, concerted effort to fix up size parsing.
I Agre exit() should not be used in parse_size... But i don't agree with Zach about better code for it...
Maybe like this:
int parse_size(const char *s, u64 * size)..
Many places that use the function parse_size/parse_limit, they have to be corrected after change....
Thanks,
Wang
Quoted 12 lines
Who knew it was so hard? :)
-Eric
Quoted 6 lines
david
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html