From: Ard Biesheuvel <ardb@kernel.org>
To: linux-efi@vger.kernel.org, Ingo Molnar <mingo@kernel.org>,
Thomas Gleixner <tglx@linutronix.de>
Cc: Ard Biesheuvel <ardb@kernel.org>,
linux-kernel@vger.kernel.org,
Arvind Sankar <nivedita@alum.mit.edu>,
Atish Patra <atish.patra@wdc.com>,
Palmer Dabbelt <palmerdabbelt@google.com>,
Zou Wei <zou_wei@huawei.com>
Subject: [PATCH 22/33] efi/libstub/arm64: Simplify randomized loading of kernel image
Date: Fri, 24 Apr 2020 15:05:20 +0200 [thread overview]
Message-ID: <20200424130531.30518-23-ardb@kernel.org> (raw)
In-Reply-To: <20200424130531.30518-1-ardb@kernel.org>
The KASLR code path in the arm64 version of the EFI stub incorporates
some overly complicated logic to randomly allocate a region of the right
alignment: there is no need to randomize the placement of the kernel
modulo 2 MiB separately from the placement of the 2 MiB aligned allocation
itself - we can simply follow the same logic used by the non-randomized
placement, which is to allocate at the correct alignment, and only take
TEXT_OFFSET into account if it is not a round multiple of the alignment.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
drivers/firmware/efi/libstub/arm64-stub.c | 32 ++++-------------------
1 file changed, 5 insertions(+), 27 deletions(-)
diff --git a/drivers/firmware/efi/libstub/arm64-stub.c b/drivers/firmware/efi/libstub/arm64-stub.c
index cfd535c13242..6fc3bd9a56db 100644
--- a/drivers/firmware/efi/libstub/arm64-stub.c
+++ b/drivers/firmware/efi/libstub/arm64-stub.c
@@ -52,7 +52,7 @@ efi_status_t handle_kernel_image(unsigned long *image_addr,
{
efi_status_t status;
unsigned long kernel_size, kernel_memsize = 0;
- u64 phys_seed = 0;
+ u32 phys_seed = 0;
if (IS_ENABLED(CONFIG_RANDOMIZE_BASE)) {
if (!nokaslr()) {
@@ -74,36 +74,15 @@ efi_status_t handle_kernel_image(unsigned long *image_addr,
kernel_size = _edata - _text;
kernel_memsize = kernel_size + (_end - _edata);
+ *reserve_size = kernel_memsize + TEXT_OFFSET % min_kimg_align;
if (IS_ENABLED(CONFIG_RANDOMIZE_BASE) && phys_seed != 0) {
- /*
- * Produce a displacement in the interval [0, MIN_KIMG_ALIGN)
- * that doesn't violate this kernel's de-facto alignment
- * constraints.
- */
- u32 mask = (MIN_KIMG_ALIGN - 1) & ~(EFI_KIMG_ALIGN - 1);
- u32 offset = (phys_seed >> 32) & mask;
-
- /*
- * With CONFIG_RANDOMIZE_TEXT_OFFSET=y, TEXT_OFFSET may not
- * be a multiple of EFI_KIMG_ALIGN, and we must ensure that
- * we preserve the misalignment of 'offset' relative to
- * EFI_KIMG_ALIGN so that statically allocated objects whose
- * alignment exceeds PAGE_SIZE appear correctly aligned in
- * memory.
- */
- offset |= TEXT_OFFSET % EFI_KIMG_ALIGN;
-
/*
* If KASLR is enabled, and we have some randomness available,
* locate the kernel at a randomized offset in physical memory.
*/
- *reserve_size = kernel_memsize + offset;
- status = efi_random_alloc(*reserve_size,
- MIN_KIMG_ALIGN, reserve_addr,
- (u32)phys_seed);
-
- *image_addr = *reserve_addr + offset;
+ status = efi_random_alloc(*reserve_size, min_kimg_align,
+ reserve_addr, phys_seed);
} else {
status = EFI_OUT_OF_RESOURCES;
}
@@ -119,7 +98,6 @@ efi_status_t handle_kernel_image(unsigned long *image_addr,
return EFI_SUCCESS;
}
- *reserve_size = kernel_memsize + TEXT_OFFSET % min_kimg_align;
status = efi_low_alloc(*reserve_size,
min_kimg_align, reserve_addr);
@@ -128,9 +106,9 @@ efi_status_t handle_kernel_image(unsigned long *image_addr,
*reserve_size = 0;
return status;
}
- *image_addr = *reserve_addr + TEXT_OFFSET % min_kimg_align;
}
+ *image_addr = *reserve_addr + TEXT_OFFSET % min_kimg_align;
memcpy((void *)*image_addr, _text, kernel_size);
return EFI_SUCCESS;
--
2.17.1
next prev parent reply other threads:[~2020-04-24 13:07 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-24 13:04 [GIT PULL 00/33] EFI updates for v5.8 Ard Biesheuvel
2020-04-24 13:04 ` [PATCH 01/33] efi/libstub: Move arm-stub to a common file Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 02/33] efi/libstub: Make initrd file loader configurable Ard Biesheuvel
2020-04-24 13:15 ` Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 03/33] efi/libstub: Unify EFI call wrappers for non-x86 Ard Biesheuvel
2020-05-03 15:09 ` Guenter Roeck
2020-05-03 16:09 ` Arvind Sankar
2020-05-03 16:11 ` Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 04/33] efi/libstub/arm: Make install_memreserve_table static Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 05/33] efi/gop: Remove redundant current_fb_base Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 06/33] efi/gop: Move check for framebuffer before con_out Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 07/33] efi/gop: Get mode information outside the loop Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 08/33] efi/gop: Factor out locating the gop into a function Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 09/33] efi/gop: Slightly re-arrange logic of find_gop Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 10/33] efi/gop: Move variable declarations into loop block Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 11/33] efi/gop: Use helper macros for populating lfb_base Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 12/33] efi/gop: Use helper macros for find_bits Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 13/33] efi/gop: Remove unreachable code from setup_pixel_info Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 14/33] efi/gop: Add prototypes for query_mode and set_mode Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 15/33] efi/gop: Allow specifying mode number on command line Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 16/33] efi/gop: Allow specifying mode by <xres>x<yres> Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 17/33] efi/gop: Allow specifying depth as well as resolution Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 18/33] efi/gop: Allow automatically choosing the best mode Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 19/33] efi/libstub/random: Align allocate size to EFI_ALLOC_ALIGN Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 20/33] efi/libstub/random: Increase random alloc granularity Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 21/33] efi/libstub/arm64: Replace 'preferred' offset with alignment check Ard Biesheuvel
2020-04-24 13:05 ` Ard Biesheuvel [this message]
2020-04-24 13:05 ` [PATCH 23/33] efi/libstub: Add API function to allocate aligned memory Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 24/33] efi/libstub/arm64: Switch to ordinary page allocator for kernel image Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 25/33] efi/libstub: Move efi_relocate_kernel() into separate source file Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 26/33] efi/arm: Remove __efistub_global annotation Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 27/33] efi/x86: Remove __efistub_global and add relocation check Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 28/33] efi: Kill __efistub_global Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 29/33] efi/libstub: Drop __pure getter for efi_system_table Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 30/33] efi/libstub: Drop __pure getters for EFI stub options Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 31/33] efi/libstub/x86: Avoid getter function for efi_is64 Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 32/33] efi: Clean up config table description arrays Ard Biesheuvel
2020-04-24 13:05 ` [PATCH 33/33] efi: Move arch_tables check to caller Ard Biesheuvel
2020-04-25 8:27 ` [GIT PULL 00/33] EFI updates for v5.8 Ingo Molnar
2020-04-25 9:57 ` Ard Biesheuvel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200424130531.30518-23-ardb@kernel.org \
--to=ardb@kernel.org \
--cc=atish.patra@wdc.com \
--cc=linux-efi@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=nivedita@alum.mit.edu \
--cc=palmerdabbelt@google.com \
--cc=tglx@linutronix.de \
--cc=zou_wei@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).