When gmin_get_config_var() calls efi.get_variable() and the EFI variable is larger than the expected buffer size, two behaviors combine to create a stack buffer overflow:
1. gmin_get_config_var() does not return the proper error code when efi.get_variable() fails. It returns the stale 'ret' value from earlier operations instead of indicating the EFI failure. 2. When efi.get_variable() returns EFI_BUFFER_TOO_SMALL, it updates *out_len to the required buffer size but writes no data to the output buffer. However, due to bug #1, gmin_get_var_int() believes the call succeeded. The caller gmin_get_var_int() then performs: - Allocates val[CFG_VAR_NAME_MAX + 1] (65 bytes) on stack - Calls gmin_get_config_var(dev, is_gmin, var, val, &len) with len=64 - If EFI variable is >64 bytes, efi.get_variable() sets len=required_size - Due to bug #1, thinks call succeeded with len=required_size - Executes val[len] = 0, writing past end of 65-byte stack buffer This creates a stack buffer overflow when EFI variables are larger than 64 bytes. Since EFI variables can be controlled by firmware or system configuration, this could potentially be exploited for code execution. Fix the bug by returning proper error codes from gmin_get_config_var() based on EFI status instead of stale 'ret' value. The gmin_get_var_int() function is called during device initialization for camera sensor configuration on Intel Bay Trail and Cherry Trail platforms using the atomisp camera stack. Reported-by: zepta <[email protected]> Closes: https://lore.kernel.org/all/capbs6koqym7fmdpwouxtexsoe44x4h3f8fw+y_qwq6e+odm...@mail.gmail.com Fixes: 38d4f74bc148 ("media: atomisp_gmin_platform: stop abusing efivar API") Signed-off-by: Kees Cook <[email protected]> --- Cc: Greg Kroah-Hartman <[email protected]> Cc: Ard Biesheuvel <[email protected]> Cc: Andy Shevchenko <[email protected]> Cc: Hans de Goede <[email protected]> Cc: Mauro Carvalho Chehab <[email protected]> Cc: Sakari Ailus <[email protected]> Cc: <[email protected]> Cc: <[email protected]> Note that as an exercise this fix and the commit log body (I wrote the tags) was entirely written by an LLM (and reviewed by me), though I really had to help it focus on where to be looking. Here were my prompts: Is there a buffer overflow problem associated with gmin_get_config_var()'s use of efi.get_variable()? What does efi.get_variable() do to out_len if it fails? Does the function return an error when efi.get_variable fails? What would the caller do when it sees a success but when efi.get_variable changed out_len? Propose fixes and write a commit message with all of the details you just gave me about reachability, impact, etc. Since the EFI_SUCCESS test ends with "return 0" you don't need an explicit "else" block for the error path code. --- .../staging/media/atomisp/pci/atomisp_gmin_platform.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/staging/media/atomisp/pci/atomisp_gmin_platform.c b/drivers/staging/media/atomisp/pci/atomisp_gmin_platform.c index 5f59519ac8e2..964cc3bcc0ac 100644 --- a/drivers/staging/media/atomisp/pci/atomisp_gmin_platform.c +++ b/drivers/staging/media/atomisp/pci/atomisp_gmin_platform.c @@ -1272,14 +1272,15 @@ static int gmin_get_config_var(struct device *maindev, if (efi_rt_services_supported(EFI_RT_SUPPORTED_GET_VARIABLE)) status = efi.get_variable(var16, &GMIN_CFG_VAR_EFI_GUID, NULL, (unsigned long *)out_len, out); - if (status == EFI_SUCCESS) + if (status == EFI_SUCCESS) { dev_info(maindev, "found EFI entry for '%s'\n", var8); - else if (is_gmin) + return 0; + } + if (is_gmin) dev_info(maindev, "Failed to find EFI gmin variable %s\n", var8); else dev_info(maindev, "Failed to find EFI variable %s\n", var8); - - return ret; + return -ENOENT; } int gmin_get_var_int(struct device *dev, bool is_gmin, const char *var, int def) -- 2.34.1
