Re: [ros-dev] [ros-diffs] [pschweitzer] 64617: [NTFS] - Add a directory entry into FCB which will be a copy of the $FILE_NAME attribute of the FILE record. Though, it won't contain the name itself (in the FCB instead), so k...

[Thomas Faber]

On 2014-10-08 21:32, [e-mail] wrote:
> MftRecord = ExAllocatePoolWithTag(NonPagedPool,
>                                  Vcb->NtfsInfo.BytesPerFileRecord,
>                                  TAG_NTFS);
>
> if (MftRecord == NULL)
>
> {
>    return NULL;
>
> }
>
> if (!NT_SUCCESS(ReadFileRecord(Vcb, NTFS_FILE_ROOT, MftRecord)))
>
> {
>    return NULL;
>
> }
>
> FileName = GetFileNameFromRecord(MftRecord);
>
> if (!FileName)
>
> {
>    return NULL;
>
> }
>
> Fcb = NtfsCreateFCB(L"\", Vcb);
>
> -//    memset(Fcb->entry.Filename, ' ', 11);
>
> -//    Fcb->Entry.DataLengthL = Vcb->CdInfo.RootSize;
> -//    Fcb->Entry.ExtentLocationL = Vcb->CdInfo.RootStart;
> -//    Fcb->Entry.FileFlags = 0x02; // FILE_ATTRIBUTE_DIRECTORY;
> if (!Fcb)
> {
>    return NULL;
> }

These failure cases look like they're all at least leaking MftRecord :p

[Pierre Schweitzer]

Ah yes! Oopsie :-$

Let's not add more leaks into it!

On 08/10/2014 21:54, Thomas Faber wrote:
> On 2014-10-08 21:32, [e-mail] wrote:
> > MftRecord = ExAllocatePoolWithTag(NonPagedPool,
> >                                  Vcb->NtfsInfo.BytesPerFileRecord,
> >                                  TAG_NTFS);
> >
> > if (MftRecord == NULL)
> >
> > {
> >    return NULL;
> >
> > }
> >
> > if (!NT_SUCCESS(ReadFileRecord(Vcb, NTFS_FILE_ROOT, MftRecord)))
> >
> > {
> >    return NULL;
> >
> > }
> >
> > FileName = GetFileNameFromRecord(MftRecord);
> >
> > if (!FileName)
> >
> > {
> >    return NULL;
> >
> > }
> >
> > Fcb = NtfsCreateFCB(L"\", Vcb);
> >
> > -//    memset(Fcb->entry.Filename, ' ', 11);
> >
> > -//    Fcb->Entry.DataLengthL = Vcb->CdInfo.RootSize;
> > -//    Fcb->Entry.ExtentLocationL = Vcb->CdInfo.RootStart;
> > -//    Fcb->Entry.FileFlags = 0x02; // FILE_ATTRIBUTE_DIRECTORY;
> > if (!Fcb)
> > {
> >    return NULL;
> > }
>
> These failure cases look like they're all at least leaking MftRecord :p
>
> Ros-dev mailing list
> [e-mail]
> http://www.reactos.org/mailman/listinfo/ros-dev
-- 
Pierre Schweitzer <pierre at reactos.org>
System & Network Administrator
Senior Kernel Developer
ReactOS Deutschland e.V.