commit bc7b488b1d1c71dc4c5182206911127bc6c410d6 upstream.
While loading the DMC firmware we were double checking the headers made sense, but in no place we checked that we were actually reading memory we were supposed to. This could be wrong in case the firmware file is truncated or malformed.
Before this patch: # ls -l /lib/firmware/i915/icl_dmc_ver1_07.bin -rw-r--r-- 1 root root 25716 Feb 1 12:26 icl_dmc_ver1_07.bin # truncate -s 25700 /lib/firmware/i915/icl_dmc_ver1_07.bin # modprobe i915 # dmesg| grep -i dmc [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm] Finished loading DMC firmware i915/icl_dmc_ver1_07.bin (v1.7)
i.e. it loads random data. Now it fails like below: [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm:csr_load_work_fn [i915]] *ERROR* Truncated DMC firmware, rejecting. i915 0000:00:02.0: Failed to load DMC firmware i915/icl_dmc_ver1_07.bin. Disabling runtime power management. i915 0000:00:02.0: DMC firmware homepage: https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/...
Before reading any part of the firmware file, validate the input first.
Fixes: eb805623d8b1 ("drm/i915/skl: Add support to load SKL CSR firmware.") Signed-off-by: Lucas De Marchi lucas.demarchi@intel.com Reviewed-by: Rodrigo Vivi rodrigo.vivi@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20190605235535.17791-1-lucas.d... (cherry picked from commit bc7b488b1d1c71dc4c5182206911127bc6c410d6) Signed-off-by: Jani Nikula jani.nikula@intel.com [ Lucas: backported to 4.9+ adjusting the context ] Cc: stable@vger.kernel.org # v4.9+ --- drivers/gpu/drm/i915/intel_csr.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 1ea0e1f43397..54d878cb458f 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -280,10 +280,17 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, uint32_t i; uint32_t *dmc_payload; uint32_t required_version; + size_t fsize;
if (!fw) return NULL;
+ fsize = sizeof(struct intel_css_header) + + sizeof(struct intel_package_header) + + sizeof(struct intel_dmc_header); + if (fsize > fw->size) + goto error_truncated; + /* Extract CSS Header information*/ css_header = (struct intel_css_header *)fw->data; if (sizeof(struct intel_css_header) != @@ -349,6 +356,9 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, return NULL; } readcount += dmc_offset; + fsize += dmc_offset; + if (fsize > fw->size) + goto error_truncated;
/* Extract dmc_header information. */ dmc_header = (struct intel_dmc_header *)&fw->data[readcount]; @@ -379,6 +389,10 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv,
/* fw_size is in dwords, so multiplied by 4 to convert into bytes. */ nbytes = dmc_header->fw_size * 4; + fsize += nbytes; + if (fsize > fw->size) + goto error_truncated; + if (nbytes > CSR_MAX_FW_SIZE) { DRM_ERROR("CSR firmware too big (%u) bytes\n", nbytes); return NULL; @@ -392,6 +406,10 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, }
return memcpy(dmc_payload, &fw->data[readcount], nbytes); + +error_truncated: + DRM_ERROR("Truncated DMC firmware, rejecting.\n"); + return NULL; }
static void csr_load_work_fn(struct work_struct *work)
On Tue, Jul 02, 2019 at 12:23:04PM -0700, Lucas De Marchi wrote:
commit bc7b488b1d1c71dc4c5182206911127bc6c410d6 upstream.
While loading the DMC firmware we were double checking the headers made sense, but in no place we checked that we were actually reading memory we were supposed to. This could be wrong in case the firmware file is truncated or malformed.
Before this patch: # ls -l /lib/firmware/i915/icl_dmc_ver1_07.bin -rw-r--r-- 1 root root 25716 Feb 1 12:26 icl_dmc_ver1_07.bin # truncate -s 25700 /lib/firmware/i915/icl_dmc_ver1_07.bin # modprobe i915 # dmesg| grep -i dmc [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm] Finished loading DMC firmware i915/icl_dmc_ver1_07.bin (v1.7)
i.e. it loads random data. Now it fails like below: [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm:csr_load_work_fn [i915]] *ERROR* Truncated DMC firmware, rejecting. i915 0000:00:02.0: Failed to load DMC firmware i915/icl_dmc_ver1_07.bin. Disabling runtime power management. i915 0000:00:02.0: DMC firmware homepage: https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/...
Before reading any part of the firmware file, validate the input first.
Fixes: eb805623d8b1 ("drm/i915/skl: Add support to load SKL CSR firmware.") Signed-off-by: Lucas De Marchi lucas.demarchi@intel.com Reviewed-by: Rodrigo Vivi rodrigo.vivi@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20190605235535.17791-1-lucas.d... (cherry picked from commit bc7b488b1d1c71dc4c5182206911127bc6c410d6) Signed-off-by: Jani Nikula jani.nikula@intel.com [ Lucas: backported to 4.9+ adjusting the context ] Cc: stable@vger.kernel.org # v4.9+
What about a 4.14.y and 4.19.y backport as well? I can't take this without those as we do not want people to upgrade and have a regression.
thanks,
greg k-h
On Wed, Jul 03, 2019 at 02:14:16PM +0200, Greg KH wrote:
On Tue, Jul 02, 2019 at 12:23:04PM -0700, Lucas De Marchi wrote:
commit bc7b488b1d1c71dc4c5182206911127bc6c410d6 upstream.
While loading the DMC firmware we were double checking the headers made sense, but in no place we checked that we were actually reading memory we were supposed to. This could be wrong in case the firmware file is truncated or malformed.
Before this patch: # ls -l /lib/firmware/i915/icl_dmc_ver1_07.bin -rw-r--r-- 1 root root 25716 Feb 1 12:26 icl_dmc_ver1_07.bin # truncate -s 25700 /lib/firmware/i915/icl_dmc_ver1_07.bin # modprobe i915 # dmesg| grep -i dmc [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm] Finished loading DMC firmware i915/icl_dmc_ver1_07.bin (v1.7)
i.e. it loads random data. Now it fails like below: [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm:csr_load_work_fn [i915]] *ERROR* Truncated DMC firmware, rejecting. i915 0000:00:02.0: Failed to load DMC firmware i915/icl_dmc_ver1_07.bin. Disabling runtime power management. i915 0000:00:02.0: DMC firmware homepage: https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/...
Before reading any part of the firmware file, validate the input first.
Fixes: eb805623d8b1 ("drm/i915/skl: Add support to load SKL CSR firmware.") Signed-off-by: Lucas De Marchi lucas.demarchi@intel.com Reviewed-by: Rodrigo Vivi rodrigo.vivi@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20190605235535.17791-1-lucas.d... (cherry picked from commit bc7b488b1d1c71dc4c5182206911127bc6c410d6) Signed-off-by: Jani Nikula jani.nikula@intel.com [ Lucas: backported to 4.9+ adjusting the context ] Cc: stable@vger.kernel.org # v4.9+
What about a 4.14.y and 4.19.y backport as well? I can't take this without those as we do not want people to upgrade and have a regression.
The documentation about stable process explicitely says the meaning of the tag is 'For each "-stable" tree starting with the specified version.'. I tried to make it clear by using the '+' suffix as I saw in other commits in stable tree.
This patch applies cleanly to 4.9, 4.14 and 4.19. This commit is already applied in 5.1 as it didn't need any backport. That was the intention, let me know if that is not the proper way.
The only missing stable is 4.4, but that needs more changes to the patch.
Lucas De Marchi
On Wed, Jul 03, 2019 at 09:24:03AM -0700, Lucas De Marchi wrote:
On Wed, Jul 03, 2019 at 02:14:16PM +0200, Greg KH wrote:
On Tue, Jul 02, 2019 at 12:23:04PM -0700, Lucas De Marchi wrote:
commit bc7b488b1d1c71dc4c5182206911127bc6c410d6 upstream.
While loading the DMC firmware we were double checking the headers made sense, but in no place we checked that we were actually reading memory we were supposed to. This could be wrong in case the firmware file is truncated or malformed.
Before this patch: # ls -l /lib/firmware/i915/icl_dmc_ver1_07.bin -rw-r--r-- 1 root root 25716 Feb 1 12:26 icl_dmc_ver1_07.bin # truncate -s 25700 /lib/firmware/i915/icl_dmc_ver1_07.bin # modprobe i915 # dmesg| grep -i dmc [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm] Finished loading DMC firmware i915/icl_dmc_ver1_07.bin (v1.7)
i.e. it loads random data. Now it fails like below: [drm:intel_csr_ucode_init [i915]] Loading i915/icl_dmc_ver1_07.bin [drm:csr_load_work_fn [i915]] *ERROR* Truncated DMC firmware, rejecting. i915 0000:00:02.0: Failed to load DMC firmware i915/icl_dmc_ver1_07.bin. Disabling runtime power management. i915 0000:00:02.0: DMC firmware homepage: https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/...
Before reading any part of the firmware file, validate the input first.
Fixes: eb805623d8b1 ("drm/i915/skl: Add support to load SKL CSR firmware.") Signed-off-by: Lucas De Marchi lucas.demarchi@intel.com Reviewed-by: Rodrigo Vivi rodrigo.vivi@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20190605235535.17791-1-lucas.d... (cherry picked from commit bc7b488b1d1c71dc4c5182206911127bc6c410d6) Signed-off-by: Jani Nikula jani.nikula@intel.com [ Lucas: backported to 4.9+ adjusting the context ] Cc: stable@vger.kernel.org # v4.9+
What about a 4.14.y and 4.19.y backport as well? I can't take this without those as we do not want people to upgrade and have a regression.
The documentation about stable process explicitely says the meaning of the tag is 'For each "-stable" tree starting with the specified version.'. I tried to make it clear by using the '+' suffix as I saw in other commits in stable tree.
This patch applies cleanly to 4.9, 4.14 and 4.19. This commit is already applied in 5.1 as it didn't need any backport. That was the intention, let me know if that is not the proper way.
The only missing stable is 4.4, but that needs more changes to the patch.
This works, I've queued it up for 4.9-4.19, thank you!
-- Thanks, Sasha
linux-stable-mirror@lists.linaro.org