Very Computer

by **Nigel Horn** » Fri, 06 Jul 2001 21:04:45

ddi_getlongprop(9F) allocates the memory to hold the property returned to
the called. In sun4u/io/pci/pci_pci.c the DDI_CTLOPS_NREGS and
DDI_CTLOPS_REGSIZE operations do not free the memory allocated in this
operation; see about line 361.

*(int *)result = 0;
if (ddi_getlongprop(DDI_DEV_T_NONE, rdip,
DDI_PROP_DONTPASS | DDI_PROP_CANSLEEP, "reg",
(caddr_t)&drv_regp, &reglen) != DDI_SUCCESS)
return (DDI_FAILURE);

The fix is to add the following line before the
return (DDI_SUCCESS);
statement:

kmem_free(drv_regp, reglen);
return(DDI_SUCCESS);

--
Nigel Horne. Arranger, Composer, Conductor, Typesetter.
Owner of the brass band group of the Internet. ICQ#20252325

by **Casper H.S. Dik - Network Security Engine** » Fri, 06 Jul 2001 23:56:01

[[ PLEASE DON'T SEND ME EMAIL COPIES OF POSTINGS ]]

>ddi_getlongprop(9F) allocates the memory to hold the property returned to
>the called. In sun4u/io/pci/pci_pci.c the DDI_CTLOPS_NREGS and
>DDI_CTLOPS_REGSIZE operations do not free the memory allocated in this
>operation; see about line 361.

Looks like:

4302431 pci_pci: memory leak in ppb_ctlops()

(fixed in S9)

Quote:> *(int *)result = 0;
> if (ddi_getlongprop(DDI_DEV_T_NONE, rdip,
> DDI_PROP_DONTPASS | DDI_PROP_CANSLEEP, "reg",
> (caddr_t)&drv_regp, &reglen) != DDI_SUCCESS)
> return (DDI_FAILURE);
>The fix is to add the following line before the
> return (DDI_SUCCESS);
>statement:
> kmem_free(drv_regp, reglen);
> return(DDI_SUCCESS);

That line is now present in the Solaris source tree.

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

Very Computer

Memory leak in Solaris 2.8 kernel PCI-PCI bridge

Memory leak in Solaris 2.8 kernel PCI-PCI bridge

Memory leak in Solaris 2.8 kernel PCI-PCI bridge