Re: [PATCH v7 3/8] libfdt: Add support for disabling dtb checks

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]



On Thu, Feb 20, 2020 at 02:45:52PM -0700, Simon Glass wrote:
> Support ASSUME_VALID_DTB to disable some sanity checks
> 
> If we assume that the DTB itself is valid then we can skip some checks and
> save code space. Add various conditions to handle this.
> 
> Signed-off-by: Simon Glass <sjg@xxxxxxxxxxxx>

A couple of things to fix in followup:
> diff --git a/libfdt/fdt_ro.c b/libfdt/fdt_ro.c
> index a5c2797..4c26fbe 100644
> --- a/libfdt/fdt_ro.c
> +++ b/libfdt/fdt_ro.c
> @@ -388,7 +388,8 @@ static const struct fdt_property *fdt_get_property_namelen_(const void *fdt,
>  	     (offset = fdt_next_property_offset(fdt, offset))) {
>  		const struct fdt_property *prop;
>  
> -		if (!(prop = fdt_get_property_by_offset_(fdt, offset, lenp))) {
> +		prop = fdt_get_property_by_offset_(fdt, offset, lenp);
> +		if (!can_assume(VALID_DTB) && !prop) {

Again, eliding checks that lead to FDT_ERR_INTERNAL could be split
out.  Basically if they ever trip, we already have a libfdt bug.

We could maybe gate them under #ifndef NDEBUG, since they are
effectively assert()s?

>  			offset = -FDT_ERR_INTERNAL;
>  			break;
>  		}
> diff --git a/libfdt/fdt_rw.c b/libfdt/fdt_rw.c
> index 8795947..707c00a 100644
> --- a/libfdt/fdt_rw.c
> +++ b/libfdt/fdt_rw.c
> @@ -24,6 +24,8 @@ static int fdt_blocks_misordered_(const void *fdt,
>  
>  static int fdt_rw_probe_(void *fdt)
>  {
> +	if (can_assume(VALID_DTB))
> +		return 0;
>  	FDT_RO_PROBE(fdt);
>  
>  	if (fdt_version(fdt) < 17)
> diff --git a/libfdt/fdt_sw.c b/libfdt/fdt_sw.c
> index 76bea22..352193c 100644
> --- a/libfdt/fdt_sw.c
> +++ b/libfdt/fdt_sw.c
> @@ -12,10 +12,13 @@
>  
>  static int fdt_sw_probe_(void *fdt)
>  {
> -	if (fdt_magic(fdt) == FDT_MAGIC)
> -		return -FDT_ERR_BADSTATE;
> -	else if (fdt_magic(fdt) != FDT_SW_MAGIC)
> -		return -FDT_ERR_BADMAGIC;
> +	if (!can_assume(VALID_DTB)) {
> +		if (fdt_magic(fdt) == FDT_MAGIC)
> +			return -FDT_ERR_BADSTATE;
> +		else if (fdt_magic(fdt) != FDT_SW_MAGIC)
> +			return -FDT_ERR_BADMAGIC;
> +	}
> +
>  	return 0;
>  }
>  
> @@ -38,7 +41,7 @@ static int fdt_sw_probe_memrsv_(void *fdt)
>  	if (err)
>  		return err;
>  
> -	if (fdt_off_dt_strings(fdt) != 0)
> +	if (!can_assume(VALID_DTB) && fdt_off_dt_strings(fdt) != 0)
>  		return -FDT_ERR_BADSTATE;

Hrm.  I think checks leading to FDT_ERR_BADSTATE should be moved under
VALID_INPUT rather than VALID_DTB.  They're generally about using the
fdt_sw functions in the wrong order, so they're not really about the
initial dtb input we're working with.  In fact with the sw functions,
there *isn't* an initial dt, since they're about building one from
scratch.

They're about "input" in the sense that they're about how the library
is invoked during an indivudual "session".

>  	return 0;
>  }
> @@ -151,7 +154,8 @@ int fdt_resize(void *fdt, void *buf, int bufsize)
>  	headsize = fdt_off_dt_struct(fdt) + fdt_size_dt_struct(fdt);
>  	tailsize = fdt_size_dt_strings(fdt);
>  
> -	if ((headsize + tailsize) > fdt_totalsize(fdt))
> +	if (!can_assume(VALID_DTB) &&
> +	    headsize + tailsize > fdt_totalsize(fdt))
>  		return -FDT_ERR_INTERNAL;
>  
>  	if ((headsize + tailsize) > bufsize)

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

Attachment: signature.asc
Description: PGP signature


[Index of Archives]     [Device Tree]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux