Messages in this thread | | | From | David Laight <> | Subject | RE: [PATCH v2] x86/boot: save fields explicitly, zero out everything else | Date | Wed, 7 Aug 2019 11:41:06 +0000 |
| |
From: john.hubbard@gmail.com > Sent: 31 July 2019 06:46 > > Recent gcc compilers (gcc 9.1) generate warnings about an > out of bounds memset, if you trying memset across several fields > of a struct. This generated a couple of warnings on x86_64 builds. > > Fix this by explicitly saving the fields in struct boot_params > that are intended to be preserved, and zeroing all the rest. > > Suggested-by: Thomas Gleixner <tglx@linutronix.de> > Suggested-by: H. Peter Anvin <hpa@zytor.com> > Signed-off-by: John Hubbard <jhubbard@nvidia.com> > --- > arch/x86/include/asm/bootparam_utils.h | 62 +++++++++++++++++++------- > 1 file changed, 47 insertions(+), 15 deletions(-) > > diff --git a/arch/x86/include/asm/bootparam_utils.h b/arch/x86/include/asm/bootparam_utils.h > index 101eb944f13c..514aee24b8de 100644 > --- a/arch/x86/include/asm/bootparam_utils.h > +++ b/arch/x86/include/asm/bootparam_utils.h > @@ -18,6 +18,20 @@ > * Note: efi_info is commonly left uninitialized, but that field has a > * private magic, so it is better to leave it unchanged. > */ > + > +#define sizeof_mbr(type, member) ({ sizeof(((type *)0)->member); }) > + > +#define BOOT_PARAM_PRESERVE(struct_member) \ > + { \ > + .start = offsetof(struct boot_params, struct_member), \ > + .len = sizeof_mbr(struct boot_params, struct_member), \ > + } > + > +struct boot_params_to_save { > + unsigned int start; > + unsigned int len; > +}; > + > static void sanitize_boot_params(struct boot_params *boot_params) > { > /* > @@ -35,21 +49,39 @@ static void sanitize_boot_params(struct boot_params *boot_params) > * problems again. > */ > if (boot_params->sentinel) { > - /* fields in boot_params are left uninitialized, clear them */ > - boot_params->acpi_rsdp_addr = 0; > - memset(&boot_params->ext_ramdisk_image, 0, > - (char *)&boot_params->efi_info - > - (char *)&boot_params->ext_ramdisk_image); > - memset(&boot_params->kbd_status, 0, > - (char *)&boot_params->hdr - > - (char *)&boot_params->kbd_status); > - memset(&boot_params->_pad7[0], 0, > - (char *)&boot_params->edd_mbr_sig_buffer[0] - > - (char *)&boot_params->_pad7[0]); > - memset(&boot_params->_pad8[0], 0, > - (char *)&boot_params->eddbuf[0] - > - (char *)&boot_params->_pad8[0]); > - memset(&boot_params->_pad9[0], 0, sizeof(boot_params->_pad9)); ...
How about replacing the above first using: #define zero_struct_fields(ptr, from, to) memset(&ptr->from, 0, (char *)&ptr->to - (char *)&ptr->from) zero_struct_fields(boot_params, ext_ramdisk_image, efi_info); ... Which is absolutely identical to the original code.
The replacing the define with: #define so(s, m) offsetof(typeof(*s), m) #define zero_struct_fields(ptr, from, to) memset((char *)ptr + so(ptr, from), 0, so(ptr, to) - so(ptr, from)) which gcc probably doesn't complain about, but should generate identical code again. There might be an existing define for so().
David
- Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
| |