-
Notifications
You must be signed in to change notification settings - Fork 305
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Enable the NX compatibility flag by default. #530
Conversation
Currently by default, when we build shim we do not set the PE NX-compatibility DLL Characteristic flag. This signifies to the firmware that shim (including the components it loads) is not prepared for several related firmware changes: - non-executable stack - non-executable pages from AllocatePages()/AllocatePool()/etc. - non-writable 0 page (not strictly related but some firmware will be transitioning at the same time) - the need to use the UEFI 2.10 Memory Attribute Protocol to set page permissions. This patch changes that default to be enabled by default. Distributors of shim will need to ensure that either their builds disable this bit (using "post-process-pe -N"), or that the bootloaders and kernels you support loading are all compliant with this change. A new make variable, POST_PROCESS_PE_FLAGS, has been added to simplify doing so. Signed-off-by: Peter Jones <[email protected]>
I have tested this patch on my side. Currently I didn't see problem when booting with grub2 and even with fallback/mokmanager. My testing is on top of OVMF 202205. Should we enable any NV function when building OVMF for testing? |
OK! I am too fast to say OK. I have tried to change the PcdImageProtectionPolicy and PcdDxeNxMemoryProtectionPolicy default value to enable the proection, and built OVMF. My change is: --- edk2.orig/MdeModulePkg/MdeModulePkg.dec Then I tested patched shim with grub2. Looks no problem, I can see grub2 menu. But I got a exception dump when running kernel: Loading Linux 5.19.2-1-default ... |
Yeah, x86 kernel is currently known broken, needs the work Evgeniy Baskov has been doing in this (and related) threads: |
Thanks for you point it out! I will also looking at Evgeniy's patches. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks okay to me
I tested his v2 series and found some problems that he's fixed. Here is the latest series (V4). |
Where are we up to with the NX support? I'm looking at a 15.7 build now, and I guess I'll have to turn this on... |
Please consider making a new tag of shim including this. I was not aware it was required and spent time making a 15.7 shim and now am unsure what to do with it: rhboot/shim-review#313 |
What's changed * Various CVE fixes: CVE-2023-40546 mok: fix LogError() invocation CVE-2023-40547 - avoid incorrectly trusting HTTP headers CVE-2023-40548 Fix integer overflow on SBAT section size on 32-bit system CVE-2023-40549 Authenticode: verify that the signature header is in bounds. CVE-2023-40550 pe: Fix an out-of-bound read in verify_buffer_sbat() CVE-2023-40551: pe-relocate: Fix bounds check for MZ binaries * Add make infrastructure to set the NX_COMPAT flag by @vathpela in #530 * Make sbat_var.S parse right with buggy gcc/binutils by @vathpela in #535 * Drop invalid calls to CRYPTO_set_mem_functions by @nicholasbishop in #537 * pe: Align section size up to page size for mem attrs by @nicholasbishop in #539 * test-sbat: Fix exit code by @vathpela in #540 * pe: Add IS_PAGE_ALIGNED macro by @nicholasbishop in #541 * CryptoPkg/BaseCryptLib: Fix buffer overflow issue in realloc wrapper by @nicholasbishop in #546 * Don't loop forever in load_certs() with buggy firmware by @rmetrich in #547 * Block Debian grub binaries with SBAT < 4 by @steve-mcintyre in #550 * Shim unable to locate grubx64 in PXE boot mode when grubx64 is stored in a different file path by @Alberto-Perez-Guevara in #551 * Further improve load_certs() for non-compliant drivers/firmwares by @pbatard in #560 * pe: only process RelocDir->Size of reloc section by @mikebeaton in #562 * Rename 'msecs' to 'usecs' to avoid potential confusion by @aronowski in #563 * Optionally allow to keep shim protocol installed by @bluca in #565 * SBAT-related documents formatting and spelling by @aronowski in #566 * Add SbatLevel_Variable.txt to document the various revocations by @jsetje in #569 * Add a security contact email address in README.md by @vathpela in #572 * Use -Wno-unused-but-set-variable for Cryptlib and OpenSSL by @vathpela in #576 * mok: fix LogError() invocation by @vathpela in #577 * Minor housekeeping by @vathpela in #578 * Test ImageAddress() by @vathpela in #579 * FreePages() is used to return memory allocated by AllocatePages() by @dennis-tseng99 in #580 * Size should minus 1 when calculating 'RelocBaseEnd' by @jsetje in #581 * Verify signature before verifying sbat levels by @jsetje in #583 * Add libFuzzer support for csv.c and sbat.c by @vathpela in #584 * mok: Avoid underflow in maximum variable size calculation by @alpernebbi in #587 * Housekeeping by @vathpela in #605 Signed-off-by: Peter Jones <[email protected]>
What's changed * Various CVE fixes: CVE-2023-40546 mok: fix LogError() invocation CVE-2023-40547 - avoid incorrectly trusting HTTP headers CVE-2023-40548 Fix integer overflow on SBAT section size on 32-bit system CVE-2023-40549 Authenticode: verify that the signature header is in bounds. CVE-2023-40550 pe: Fix an out-of-bound read in verify_buffer_sbat() CVE-2023-40551: pe-relocate: Fix bounds check for MZ binaries * Add make infrastructure to set the NX_COMPAT flag by @vathpela in rhboot#530 * Make sbat_var.S parse right with buggy gcc/binutils by @vathpela in rhboot#535 * Drop invalid calls to CRYPTO_set_mem_functions by @nicholasbishop in rhboot#537 * pe: Align section size up to page size for mem attrs by @nicholasbishop in rhboot#539 * test-sbat: Fix exit code by @vathpela in rhboot#540 * pe: Add IS_PAGE_ALIGNED macro by @nicholasbishop in rhboot#541 * CryptoPkg/BaseCryptLib: Fix buffer overflow issue in realloc wrapper by @nicholasbishop in rhboot#546 * Don't loop forever in load_certs() with buggy firmware by @rmetrich in rhboot#547 * Block Debian grub binaries with SBAT < 4 by @steve-mcintyre in rhboot#550 * Shim unable to locate grubx64 in PXE boot mode when grubx64 is stored in a different file path by @Alberto-Perez-Guevara in rhboot#551 * Further improve load_certs() for non-compliant drivers/firmwares by @pbatard in rhboot#560 * pe: only process RelocDir->Size of reloc section by @mikebeaton in rhboot#562 * Rename 'msecs' to 'usecs' to avoid potential confusion by @aronowski in rhboot#563 * Optionally allow to keep shim protocol installed by @bluca in rhboot#565 * SBAT-related documents formatting and spelling by @aronowski in rhboot#566 * Add SbatLevel_Variable.txt to document the various revocations by @jsetje in rhboot#569 * Add a security contact email address in README.md by @vathpela in rhboot#572 * Use -Wno-unused-but-set-variable for Cryptlib and OpenSSL by @vathpela in rhboot#576 * mok: fix LogError() invocation by @vathpela in rhboot#577 * Minor housekeeping by @vathpela in rhboot#578 * Test ImageAddress() by @vathpela in rhboot#579 * FreePages() is used to return memory allocated by AllocatePages() by @dennis-tseng99 in rhboot#580 * Size should minus 1 when calculating 'RelocBaseEnd' by @jsetje in rhboot#581 * Verify signature before verifying sbat levels by @jsetje in rhboot#583 * Add libFuzzer support for csv.c and sbat.c by @vathpela in rhboot#584 * mok: Avoid underflow in maximum variable size calculation by @alpernebbi in rhboot#587 * Housekeeping by @vathpela in rhboot#605 Signed-off-by: Peter Jones <[email protected]>
Currently by default, when we build shim we do not set the PE NX-compatibility DLL Characteristic flag. This signifies to the firmware that shim (including the components it loads) is not prepared for several related firmware changes:
This patch changes that default to be enabled by default. Distributors of shim will need to ensure that either their builds disable this bit (using "post-process-pe -N"), or that the bootloaders and kernels you support loading are all compliant with this change. A new make variable, POST_PROCESS_PE_FLAGS, has been added to simplify doing so.
Signed-off-by: Peter Jones [email protected]