> On Jul 14, 2020, at 5:42 PM, Arvind Sankar <nivedita@xxxxxxxxxxxx> wrote: > > On Wed, Jul 08, 2020 at 11:50:19AM -0700, Nick Terrell wrote: >> From: Nick Terrell <terrelln@xxxxxx> >> >> * Add unzstd() and the zstd decompress interface. >> * Add zstd support to decompress_method(). >> >> The decompress_method() and unzstd() functions are used to decompress >> the initramfs and the initrd. The __decompress() function is used in >> the preboot environment to decompress a zstd compressed kernel. >> >> The zstd decompression function allows the input and output buffers to >> overlap because that is used by x86 kernel decompression. >> >> Reviewed-by: Kees Cook <keescook@xxxxxxxxxxxx> >> Tested-by: Sedat Dilek <sedat.dilek@xxxxxxxxx> >> Signed-off-by: Nick Terrell <terrelln@xxxxxx> >> + * >> + * __DISABLE_EXPORTS stops zstd and xxhash from declaring themselves >> + * as modules by disabling the EXPORT_SYMBOL macro. >> + */ > > Hi Nick, this doesn't actually work, because misc.c includes export.h > via > "misc.h" -> <linux/linkage.h> -> <linux/export.h> > and EXPORT_SYMBOL is already defined before __DISABLE_EXPORTS is defined > here. > > Adding -D__EXPORT_SYMBOLS to KBUILD_CFLAGS in arch/x86/boot/compressed/Makefile > fixes that, and then you don't have to define it here. > > Also, __DISABLE_EXPORTS does not currently inhibit MODULE_INFO, so a > .modinfo section gets emitted into misc.o. I *think* this is harmless > and it will get discarded when building compressed/vmlinux, but thought > I'd point it out. Thanks for pointing that out! I’m surprised that the zstd compressed kernels built and ran successfully, I vaguely remember that failing in the past. I will put up a fix shortly. -Nick
