Re: CVS commit: src/sys/dev/dm
Le 18/08/2014 19:16, Alistair G. Crooks a écrit :
Module Name: src
Committed By: agc
Date: Mon Aug 18 17:16:42 UTC 2014
Modified Files:
src/sys/dev/dm: dm_target_stripe.c
Log Message:
Avoid a memory leak - from maxv
To generate a diff of this commit:
cvs rdiff -u -r1.19 -r1.20 src/sys/dev/dm/dm_target_stripe.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
I have a doubt for this one.
for (strpi = DM_STRIPE_DEV_OFFSET; strpi strpc; strpi += 2) {
printf(Stripe target device name %s -- offset %s\n,
argv[strpi], argv[strpi+1]);
tlc = kmem_alloc(sizeof(*tlc), KM_NOSLEEP);
if ((tlc-pdev = dm_pdev_insert(argv[strpi])) == NULL) {
kmem_free(tsc, sizeof(*tsc));
XXX kmem_free(tlc, sizeof(*tlc));
return ENOENT;
}
tlc-offset = atoi(argv[strpi+1]);
/* Insert striping device to linked list. */
XXX TAILQ_INSERT_TAIL(tsc-stripe_devs, tlc, entries);
}
The tlc's inserted into the list are not freed, are they?
Re: CVS commit: src/sys/dev/dm
On Aug 19, 6:50am, Maxime Villard wrote:
} Le 18/08/2014 19:16, Alistair G. Crooks a écrit :
}
} Module Name:src
} Committed By: agc
} Date: Mon Aug 18 17:16:42 UTC 2014
}
} Modified Files:
} src/sys/dev/dm: dm_target_stripe.c
}
} Log Message:
} Avoid a memory leak - from maxv
}
} To generate a diff of this commit:
} cvs rdiff -u -r1.19 -r1.20 src/sys/dev/dm/dm_target_stripe.c
}
} I have a doubt for this one.
}
} for (strpi = DM_STRIPE_DEV_OFFSET; strpi strpc; strpi += 2) {
} printf(Stripe target device name %s -- offset %s\n,
} argv[strpi], argv[strpi+1]);
}
} tlc = kmem_alloc(sizeof(*tlc), KM_NOSLEEP);
} if ((tlc-pdev = dm_pdev_insert(argv[strpi])) == NULL) {
} kmem_free(tsc, sizeof(*tsc));
} XXX kmem_free(tlc, sizeof(*tlc));
} return ENOENT;
} }
} tlc-offset = atoi(argv[strpi+1]);
}
} /* Insert striping device to linked list. */
} XXX TAILQ_INSERT_TAIL(tsc-stripe_devs, tlc, entries);
} }
}
} The tlc's inserted into the list are not freed, are they?
Notice the return ENOENT after the two calls to kmem_free().
If the first XXX line is executed then the second XXX line won't be.
}-- End of excerpt from Maxime Villard
Re: CVS commit: src/sys/dev/dm
Le 19/08/2014 10:31, John Nemeth a écrit :
On Aug 19, 6:50am, Maxime Villard wrote:
} Le 18/08/2014 19:16, Alistair G. Crooks a écrit :
}
} Module Name: src
} Committed By: agc
} Date: Mon Aug 18 17:16:42 UTC 2014
}
} Modified Files:
}src/sys/dev/dm: dm_target_stripe.c
}
} Log Message:
} Avoid a memory leak - from maxv
}
} To generate a diff of this commit:
} cvs rdiff -u -r1.19 -r1.20 src/sys/dev/dm/dm_target_stripe.c
}
} I have a doubt for this one.
}
} for (strpi = DM_STRIPE_DEV_OFFSET; strpi strpc; strpi += 2) {
} printf(Stripe target device name %s -- offset %s\n,
}argv[strpi], argv[strpi+1]);
}
} tlc = kmem_alloc(sizeof(*tlc), KM_NOSLEEP);
} if ((tlc-pdev = dm_pdev_insert(argv[strpi])) == NULL) {
} kmem_free(tsc, sizeof(*tsc));
} XXX kmem_free(tlc, sizeof(*tlc));
} return ENOENT;
} }
} tlc-offset = atoi(argv[strpi+1]);
}
} /* Insert striping device to linked list. */
} XXX TAILQ_INSERT_TAIL(tsc-stripe_devs, tlc, entries);
} }
}
} The tlc's inserted into the list are not freed, are they?
Notice the return ENOENT after the two calls to kmem_free().
If the first XXX line is executed then the second XXX line won't be.
Yes, but that's a loop. The second XXX line may be reached several times,
before the first one returns ENOENT. And on that case the tlc's previously
allocated are not freed.
Unless I'm not reading clearly?
Re: CVS commit: src/sys/dev/dm
On Tue, May 18, 2010 at 03:10:41PM +, Adam Hamsik wrote: Log Message: Add support for DIOCCACHESYNC ioctl for dm devices. Add new sync function pointer to dm_target_t because that is the only part of dm which know real block device. disk_ioctl_switch parses whole device table and for every entry it calls particular sync routine which propagates DIOCCACHESYNC to real disk. What happens if a logical volume consists of two slice of one physical device? Will the physical device get flushed twice? Kind regards -- Matthias Scheler http://zhadum.org.uk/
Re: CVS commit: src/sys/dev/dm
On May,Tuesday 18 2010, at 5:42 PM, Matthias Scheler wrote: On Tue, May 18, 2010 at 03:10:41PM +, Adam Hamsik wrote: Log Message: Add support for DIOCCACHESYNC ioctl for dm devices. Add new sync function pointer to dm_target_t because that is the only part of dm which know real block device. disk_ioctl_switch parses whole device table and for every entry it calls particular sync routine which propagates DIOCCACHESYNC to real disk. What happens if a logical volume consists of two slice of one physical device? Will the physical device get flushed twice? Yes and there is no easy way how to not do that. Regards Adam.
Re: CVS commit: src/sys/dev/dm
On Mar,Tuesday 23 2010, at 4:09 PM, Jonathan A. Kollasch wrote: Module Name: src Committed By: jakllsch Date: Tue Mar 23 15:09:45 UTC 2010 Modified Files: src/sys/dev/dm: device-mapper.c Log Message: Rework module/builtin code so it works in both cases. (Tested recently in the module case, slightly less recently as builtin.) haad [if it works] go for it now when we have built in case working, would you like to add pseudo-device dm to AMD64, i386 non-modular kernels ? I think that ALL MONOLITHIC *XEN* should be enough. What do you think about this ?(Sorry I have zero free time now, otherwise I would do it myself:() Regards Adam.
