* [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-06-28 7:06 ` Peng Fan (OSS)
0 siblings, 0 replies; 18+ messages in thread
From: Peng Fan (OSS) @ 2021-06-28 7:06 UTC (permalink / raw)
To: ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: shawnguo, s.hauer, kernel, festevam, linux-imx, linux-remoteproc,
linux-arm-kernel, linux-kernel, aisheng.dong, Peng Fan
From: Peng Fan <peng.fan@nxp.com>
It seems luckliy work on i.MX platform, but it is wrong.
Need use memcpy_toio, not memcpy_fromio.
Fixes: 40df0a91b2a52 ("remoteproc: add is_iomem to da_to_va")
Tested-by: Dong Aisheng <aisheng.dong@nxp.com> (i.MX8MQ)
Reported-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
---
drivers/remoteproc/remoteproc_elf_loader.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/remoteproc/remoteproc_elf_loader.c b/drivers/remoteproc/remoteproc_elf_loader.c
index 469c52e62faf..e8078efb3dec 100644
--- a/drivers/remoteproc/remoteproc_elf_loader.c
+++ b/drivers/remoteproc/remoteproc_elf_loader.c
@@ -220,7 +220,7 @@ int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
/* put the segment where the remote processor expects it */
if (filesz) {
if (is_iomem)
- memcpy_fromio(ptr, (void __iomem *)(elf_data + offset), filesz);
+ memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
else
memcpy(ptr, elf_data + offset, filesz);
}
--
2.30.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply related [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-06-28 8:52 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 8:52 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6806 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210625]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: riscv-randconfig-s031-20210628 (attached as .config)
compiler: riscv64-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=riscv
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *addr @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *addr
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *buffer @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *buffer
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 33286 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-06-28 8:52 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 8:52 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6965 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210625]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: riscv-randconfig-s031-20210628 (attached as .config)
compiler: riscv64-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=riscv
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *addr @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *addr
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *buffer @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *buffer
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 33286 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-06-28 13:45 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 13:45 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6590 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210628]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: i386-randconfig-s002-20210628 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce:
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=i386
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem * @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const * @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 45366 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-06-28 13:45 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 13:45 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6747 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210628]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: i386-randconfig-s002-20210628 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce:
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
make W=1 C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=i386
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem * @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const * @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 45366 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-06-28 14:23 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 14:23 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6762 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210628]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: parisc-randconfig-s032-20210628 (attached as .config)
compiler: hppa-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=parisc
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *dst @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *dst
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *src @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *src
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 22225 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-06-28 14:23 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-06-28 14:23 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6921 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210628]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: parisc-randconfig-s032-20210628 (attached as .config)
compiler: hppa-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=parisc
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *dst @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *dst
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *src @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *src
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 22225 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-06-28 16:16 ` Bjorn Andersson
-1 siblings, 0 replies; 18+ messages in thread
From: Bjorn Andersson @ 2021-06-28 16:16 UTC (permalink / raw)
To: Peng Fan (OSS)
Cc: ohad, mathieu.poirier, o.rempel, shawnguo, s.hauer, kernel,
festevam, linux-imx, linux-remoteproc, linux-arm-kernel,
linux-kernel, aisheng.dong, Peng Fan
On Mon 28 Jun 02:06 CDT 2021, Peng Fan (OSS) wrote:
> From: Peng Fan <peng.fan@nxp.com>
>
> It seems luckliy work on i.MX platform, but it is wrong.
> Need use memcpy_toio, not memcpy_fromio.
>
> Fixes: 40df0a91b2a52 ("remoteproc: add is_iomem to da_to_va")
> Tested-by: Dong Aisheng <aisheng.dong@nxp.com> (i.MX8MQ)
> Reported-by: Dong Aisheng <aisheng.dong@nxp.com>
> Signed-off-by: Peng Fan <peng.fan@nxp.com>
> ---
> drivers/remoteproc/remoteproc_elf_loader.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/remoteproc/remoteproc_elf_loader.c b/drivers/remoteproc/remoteproc_elf_loader.c
> index 469c52e62faf..e8078efb3dec 100644
> --- a/drivers/remoteproc/remoteproc_elf_loader.c
> +++ b/drivers/remoteproc/remoteproc_elf_loader.c
> @@ -220,7 +220,7 @@ int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
> /* put the segment where the remote processor expects it */
> if (filesz) {
> if (is_iomem)
> - memcpy_fromio(ptr, (void __iomem *)(elf_data + offset), filesz);
> + memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
Ouch, we should have caught that during the review, thanks for spotting
and fixing this.
That said, it's the ptr that's __iomem, so the cast should be on the
first parameter, not the second.
Regards,
Bjorn
> else
> memcpy(ptr, elf_data + offset, filesz);
> }
> --
> 2.30.0
>
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-06-28 16:16 ` Bjorn Andersson
0 siblings, 0 replies; 18+ messages in thread
From: Bjorn Andersson @ 2021-06-28 16:16 UTC (permalink / raw)
To: Peng Fan (OSS)
Cc: ohad, mathieu.poirier, o.rempel, shawnguo, s.hauer, kernel,
festevam, linux-imx, linux-remoteproc, linux-arm-kernel,
linux-kernel, aisheng.dong, Peng Fan
On Mon 28 Jun 02:06 CDT 2021, Peng Fan (OSS) wrote:
> From: Peng Fan <peng.fan@nxp.com>
>
> It seems luckliy work on i.MX platform, but it is wrong.
> Need use memcpy_toio, not memcpy_fromio.
>
> Fixes: 40df0a91b2a52 ("remoteproc: add is_iomem to da_to_va")
> Tested-by: Dong Aisheng <aisheng.dong@nxp.com> (i.MX8MQ)
> Reported-by: Dong Aisheng <aisheng.dong@nxp.com>
> Signed-off-by: Peng Fan <peng.fan@nxp.com>
> ---
> drivers/remoteproc/remoteproc_elf_loader.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/remoteproc/remoteproc_elf_loader.c b/drivers/remoteproc/remoteproc_elf_loader.c
> index 469c52e62faf..e8078efb3dec 100644
> --- a/drivers/remoteproc/remoteproc_elf_loader.c
> +++ b/drivers/remoteproc/remoteproc_elf_loader.c
> @@ -220,7 +220,7 @@ int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
> /* put the segment where the remote processor expects it */
> if (filesz) {
> if (is_iomem)
> - memcpy_fromio(ptr, (void __iomem *)(elf_data + offset), filesz);
> + memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
Ouch, we should have caught that during the review, thanks for spotting
and fixing this.
That said, it's the ptr that's __iomem, so the cast should be on the
first parameter, not the second.
Regards,
Bjorn
> else
> memcpy(ptr, elf_data + offset, filesz);
> }
> --
> 2.30.0
>
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-07-01 16:15 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-01 16:15 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6798 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: arm-randconfig-s031-20210630 (attached as .config)
compiler: arm-linux-gnueabi-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=arm
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *to @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: expected void volatile [noderef] __iomem *to
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *from @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: expected void const *from
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 34855 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-07-01 16:15 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-01 16:15 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6957 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master remoteproc/for-next rpmsg/for-next]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: arm-randconfig-s031-20210630 (attached as .config)
compiler: arm-linux-gnueabi-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=arm
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *to @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: expected void volatile [noderef] __iomem *to
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *from @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: expected void const *from
drivers/remoteproc/remoteproc_elf_loader.c:219:33: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 34855 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-07-05 7:22 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-05 7:22 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6763 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-07-05 7:22 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-05 7:22 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6922 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-07-05 21:07 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-05 21:07 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6763 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-07-05 21:07 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-05 21:07 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6922 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210701]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
2021-06-28 7:06 ` Peng Fan (OSS)
@ 2021-07-06 12:55 ` kernel test robot
-1 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-06 12:55 UTC (permalink / raw)
To: Peng Fan (OSS), ohad, bjorn.andersson, mathieu.poirier, o.rempel
Cc: kbuild-all, shawnguo, s.hauer, kernel, festevam, linux-imx,
linux-remoteproc
[-- Attachment #1: Type: text/plain, Size: 6763 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210706]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH] remoteproc: elf_loader: fix loading segment when is_iomem true
@ 2021-07-06 12:55 ` kernel test robot
0 siblings, 0 replies; 18+ messages in thread
From: kernel test robot @ 2021-07-06 12:55 UTC (permalink / raw)
To: kbuild-all
[-- Attachment #1: Type: text/plain, Size: 6922 bytes --]
Hi "Peng,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on soc/for-next]
[also build test WARNING on linus/master v5.13 next-20210706]
[cannot apply to xlnx/master]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
base: https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git for-next
config: powerpc64-randconfig-s032-20210705 (attached as .config)
compiler: powerpc-linux-gcc (GCC) 9.3.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# apt-get install sparse
# sparse version: v0.6.3-341-g8af24329-dirty
# https://github.com/0day-ci/linux/commit/2c90d7d15776bf485ad028de63002efbc2facc66
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Peng-Fan-OSS/remoteproc-elf_loader-fix-loading-segment-when-is_iomem-true/20210628-143358
git checkout 2c90d7d15776bf485ad028de63002efbc2facc66
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' ARCH=powerpc64
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>
sparse warnings: (new ones prefixed by >>)
>> drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: sparse: incorrect type in argument 1 (different address spaces) @@ expected void volatile [noderef] __iomem *d @@ got void *[assigned] ptr @@
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: expected void volatile [noderef] __iomem *d
drivers/remoteproc/remoteproc_elf_loader.c:219:45: sparse: got void *[assigned] ptr
>> drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: sparse: incorrect type in argument 2 (different address spaces) @@ expected void const *s @@ got void [noderef] __iomem * @@
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: expected void const *s
drivers/remoteproc/remoteproc_elf_loader.c:219:51: sparse: got void [noderef] __iomem *
vim +219 drivers/remoteproc/remoteproc_elf_loader.c
131
132 /**
133 * rproc_elf_load_segments() - load firmware segments to memory
134 * @rproc: remote processor which will be booted using these fw segments
135 * @fw: the ELF firmware image
136 *
137 * This function loads the firmware segments to memory, where the remote
138 * processor expects them.
139 *
140 * Some remote processors will expect their code and data to be placed
141 * in specific device addresses, and can't have them dynamically assigned.
142 *
143 * We currently support only those kind of remote processors, and expect
144 * the program header's paddr member to contain those addresses. We then go
145 * through the physically contiguous "carveout" memory regions which we
146 * allocated (and mapped) earlier on behalf of the remote processor,
147 * and "translate" device address to kernel addresses, so we can copy the
148 * segments where they are expected.
149 *
150 * Currently we only support remote processors that required carveout
151 * allocations and got them mapped onto their iommus. Some processors
152 * might be different: they might not have iommus, and would prefer to
153 * directly allocate memory for every segment/resource. This is not yet
154 * supported, though.
155 */
156 int rproc_elf_load_segments(struct rproc *rproc, const struct firmware *fw)
157 {
158 struct device *dev = &rproc->dev;
159 const void *ehdr, *phdr;
160 int i, ret = 0;
161 u16 phnum;
162 const u8 *elf_data = fw->data;
163 u8 class = fw_elf_get_class(fw);
164 u32 elf_phdr_get_size = elf_size_of_phdr(class);
165
166 ehdr = elf_data;
167 phnum = elf_hdr_get_e_phnum(class, ehdr);
168 phdr = elf_data + elf_hdr_get_e_phoff(class, ehdr);
169
170 /* go through the available ELF segments */
171 for (i = 0; i < phnum; i++, phdr += elf_phdr_get_size) {
172 u64 da = elf_phdr_get_p_paddr(class, phdr);
173 u64 memsz = elf_phdr_get_p_memsz(class, phdr);
174 u64 filesz = elf_phdr_get_p_filesz(class, phdr);
175 u64 offset = elf_phdr_get_p_offset(class, phdr);
176 u32 type = elf_phdr_get_p_type(class, phdr);
177 void *ptr;
178 bool is_iomem;
179
180 if (type != PT_LOAD)
181 continue;
182
183 dev_dbg(dev, "phdr: type %d da 0x%llx memsz 0x%llx filesz 0x%llx\n",
184 type, da, memsz, filesz);
185
186 if (filesz > memsz) {
187 dev_err(dev, "bad phdr filesz 0x%llx memsz 0x%llx\n",
188 filesz, memsz);
189 ret = -EINVAL;
190 break;
191 }
192
193 if (offset + filesz > fw->size) {
194 dev_err(dev, "truncated fw: need 0x%llx avail 0x%zx\n",
195 offset + filesz, fw->size);
196 ret = -EINVAL;
197 break;
198 }
199
200 if (!rproc_u64_fit_in_size_t(memsz)) {
201 dev_err(dev, "size (%llx) does not fit in size_t type\n",
202 memsz);
203 ret = -EOVERFLOW;
204 break;
205 }
206
207 /* grab the kernel address for this device address */
208 ptr = rproc_da_to_va(rproc, da, memsz, &is_iomem);
209 if (!ptr) {
210 dev_err(dev, "bad phdr da 0x%llx mem 0x%llx\n", da,
211 memsz);
212 ret = -EINVAL;
213 break;
214 }
215
216 /* put the segment where the remote processor expects it */
217 if (filesz) {
218 if (is_iomem)
> 219 memcpy_toio(ptr, (void __iomem *)(elf_data + offset), filesz);
220 else
221 memcpy(ptr, elf_data + offset, filesz);
222 }
223
224 /*
225 * Zero out remaining memory for this segment.
226 *
227 * This isn't strictly required since dma_alloc_coherent already
228 * did this for us. albeit harmless, we may consider removing
229 * this.
230 */
231 if (memsz > filesz) {
232 if (is_iomem)
233 memset_io((void __iomem *)(ptr + filesz), 0, memsz - filesz);
234 else
235 memset(ptr + filesz, 0, memsz - filesz);
236 }
237 }
238
239 return ret;
240 }
241 EXPORT_SYMBOL(rproc_elf_load_segments);
242
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 48889 bytes --]
^ permalink raw reply [flat|nested] 18+ messages in thread