phy: renesas: phy-rcar-gen2: Add support for r8a77470
Hello Biju Das,
The patch b7187e001a10: "phy: renesas: phy-rcar-gen2: Add support for r8a77470" from Apr 10, 2019, leads to the following static checker warning:
drivers/phy/renesas/phy-rcar-gen2.c:403 rcar_gen2_phy_probe() warn: array off by one? 'data->select_value[channel_num]'
drivers/phy/renesas/phy-rcar-gen2.c
262 static const u32 pci_select_value[][PHYS_PER_CHANNEL] = {
263 [0] = { USBHS_UGCTRL2_USB0SEL_PCI, USBHS_UGCTRL2_USB0SEL_HS_USB },
264 [2] = { USBHS_UGCTRL2_USB2SEL_PCI, USBHS_UGCTRL2_USB2SEL_USB30 },This select array has three elements.
265 };
266
267 static const u32 usb20_select_value[][PHYS_PER_CHANNEL] = {
268 { USBHS_UGCTRL2_USB0SEL_USB20, USBHS_UGCTRL2_USB0SEL_HS_USB20 },
But this one only has two.
269 };
270
271 static const struct rcar_gen2_phy_data rcar_gen2_usb_phy_data = {
272 .gen2_phy_ops = &rcar_gen2_phy_ops,
273 .select_value = pci_select_value,
274 };
275
276 static const struct rcar_gen2_phy_data rz_g1c_usb_phy_data = {
277 .gen2_phy_ops = &rz_g1c_phy_ops,
278 .select_value = usb20_select_value,
279 };
[ snip ]
382 for_each_child_of_node(dev->of_node, np) {
383 struct rcar_gen2_channel *channel = drv->channels + i;
384 u32 channel_num;
385 int error, n;
386
387 channel->of_node = np;
388 channel->drv = drv;
389 channel->selected_phy = -1;
390
391 error = of_property_read_u32(np, "reg", &channel_num);
392 if (error || channel_num > 2) {
^^^^^^^^^^^^^^^The code seems to assume they all have 3 elements
393 dev_err(dev, "Invalid \"reg\" property\n");
394 return error;
395 }
396 channel->select_mask = select_mask[channel_num];
397
398 for (n = 0; n < PHYS_PER_CHANNEL; n++) {
399 struct rcar_gen2_phy *phy = &channel->phys[n];
400
401 phy->channel = channel;
402 phy->number = n;
403 phy->select_value = data->select_value[channel_num][n];
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^Smatch warning.
404
405 phy->phy = devm_phy_create(dev, NULL,
406 data->gen2_phy_ops);
407 if (IS_ERR(phy->phy)) {
408 dev_err(dev, "Failed to create PHY\n");
409 return PTR_ERR(phy->phy);
410 }
411 phy_set_drvdata(phy->phy, phy);
412 }
413
414 i++;
415 }regards,
dan carpenter
Hi Dan Carpenter,
Thanks for the report. Yes, the hardcoded check "channel_num > 2" to be replaced with
some variable check to fix this issue.
How to reproduce this issue? So that I can I provide a fix.
I tried using smatch tool, but I get a different warning.
biju@be1yocto:/data/biju/linux-next$ make -j8 uImage LOADADDR=80008000 CHECK="/data/biju/smatch/smatch -p=kernel" C=1 | tee warns.txt CALL scripts/atomic/check-atomics.sh CALL scripts/checksyscalls.sh CHK include/generated/compile.h CHECK drivers/phy/renesas/phy-rcar-gen2.c drivers/phy/renesas/phy-rcar-gen2.c:354 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:360 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:409 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:420 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' CC drivers/phy/renesas/phy-rcar-gen2.o AR drivers/phy/renesas/built-in.a AR drivers/phy/built-in.a AR drivers/built-in.a GEN .version CHK include/generated/compile.h UPD include/generated/compile.h CHECK init/version.c CC init/version.o AR init/built-in.a LD vmlinux.o MODPOST vmlinux.o KSYM .tmp_kallsyms1.o KSYM .tmp_kallsyms2.o LD vmlinux SORTEX vmlinux SYSMAP System.map OBJCOPY arch/arm/boot/Image Kernel: arch/arm/boot/Image is ready GZIP arch/arm/boot/compressed/piggy_data AS arch/arm/boot/compressed/piggy.o LD arch/arm/boot/compressed/vmlinux OBJCOPY arch/arm/boot/zImage Kernel: arch/arm/boot/zImage is ready UIMAGE arch/arm/boot/uImage Image Name: Linux-5.1.0-rc6-next-20190423-00 Created: Tue May 7 07:40:23 2019 Image Type: ARM Linux Kernel Image (uncompressed) Data Size: 4671240 Bytes = 4561.76 kB = 4.45 MB Load Address: 80008000 Entry Point: 80008000 Kernel: arch/arm/boot/uImage is ready biju@be1yocto:/data/biju/linux-next$ /data/biju/smatch/ c2xml compat/ ctags Documentation/ graph obfuscate smatch_data/ sparse sparsei test-dissect test-lexing test-parsing validation/ cgcc compile cwchash/ example gvpr/ smatch smatch_scripts/ sparsec sparse-llvm test-inspect test-linearize test-unssa biju@be1yocto:/data/biju/linux-next$ /data/biju/smatch/smatch_scripts/kchecker drivers/phy/renesas/ CHECK scripts/mod/empty.c CALL scripts/checksyscalls.sh CALL scripts/atomic/check-atomics.sh CHECK drivers/phy/renesas/phy-rcar-gen2.c drivers/phy/renesas/phy-rcar-gen2.c:354 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:360 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:409 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:420 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' CHECK drivers/phy/renesas/phy-rcar-gen3-usb2.c drivers/phy/renesas/phy-rcar-gen3-usb2.c:602 rcar_gen3_phy_usb2_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen3-usb2.c:670 rcar_gen3_phy_usb2_probe() warn: passing zero to 'PTR_ERR' biju@be1yocto:/data/biju/linux-next$
regards,
Biju
Quoted 90 lines
-----Original Message-----
From: Dan Carpenter <[email protected]>
Sent: 03 May 2019 14:10
To: Biju Das <[email protected]>
Cc: [email protected]
Subject: [bug report] phy: renesas: phy-rcar-gen2: Add support for r8a77470Hello Biju Das,
The patch
b7187e001a10: "phy: renesas: phy-rcar-gen2: Add support for r8a77470" from Apr 10, 2019, leads to the following static checker warning:drivers/phy/renesas/phy-rcar-gen2.c:403 rcar_gen2_phy_probe() warn: array off by one? 'data->select_value[channel_num]'
drivers/phy/renesas/phy-rcar-gen2.c262 static const u32 pci_select_value[][PHYS_PER_CHANNEL] = { 263 [0] = { USBHS_UGCTRL2_USB0SEL_PCI,
USBHS_UGCTRL2_USB0SEL_HS_USB},
264 [2] = {USBHS_UGCTRL2_USB2SEL_PCI,USBHS_UGCTRL2_USB2SEL_USB30},This select array has three elements.
265 }; 266 267 static const u32 usb20_select_value[][PHYS_PER_CHANNEL] = { 268 { USBHS_UGCTRL2_USB0SEL_USB20,
USBHS_UGCTRL2_USB0SEL_HS_USB20},But this one only has two.
269 }; 270 271 static const struct rcar_gen2_phy_data rcar_gen2_usb_phy_data = { 272 .gen2_phy_ops = &rcar_gen2_phy_ops, 273 .select_value = pci_select_value, 274 }; 275 276 static const struct rcar_gen2_phy_data rz_g1c_usb_phy_data = { 277 .gen2_phy_ops = &rz_g1c_phy_ops, 278 .select_value = usb20_select_value, 279 };[ snip ]
382 for_each_child_of_node(dev->of_node, np) { 383 struct rcar_gen2_channel *channel = drv->channels + i; 384 u32 channel_num; 385 int error, n; 386 387 channel->of_node = np; 388 channel->drv = drv; 389 channel->selected_phy = -1; 390 391 error = of_property_read_u32(np, "reg", &channel_num); 392 if (error || channel_num > 2) { ^^^^^^^^^^^^^^^ The code seems to assume they allhave 3 elements
393 dev_err(dev, "Invalid \"reg\" property\n"); 394 return error; 395 } 396 channel->select_mask = select_mask[channel_num]; 397 398 for (n = 0; n < PHYS_PER_CHANNEL; n++) { 399 struct rcar_gen2_phy *phy = &channel->phys[n]; 400 401 phy->channel = channel; 402 phy->number = n; 403 phy->select_value = data->select_value[channel_num][n]; ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Smatchwarning.
404 405 phy->phy = devm_phy_create(dev, NULL, 406 data->gen2_phy_ops); 407 if (IS_ERR(phy->phy)) { 408 dev_err(dev, "Failed to create PHY\n"); 409 return PTR_ERR(phy->phy); 410 } 411 phy_set_drvdata(phy->phy, phy); 412 } 413 414 i++; 415 }regards,
dan carpenter
That check hasn't been released. (It marks every "if (foo > bar) " as off by one unless it can be proved as okay...
Quoted 5 lines · Biju Das
biju@be1yocto:/data/biju/linux-next$ /data/biju/smatch/smatch_scripts/kchecker drivers/phy/renesas/ CHECK scripts/mod/empty.c CALL scripts/checksyscalls.sh CALL scripts/atomic/check-atomics.sh CHECK drivers/phy/renesas/phy-rcar-gen2.c
Quoted 8 lines
drivers/phy/renesas/phy-rcar-gen2.c:354 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:360 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:409 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen2.c:420 rcar_gen2_phy_probe() warn: passing zero to 'PTR_ERR' CHECK drivers/phy/renesas/phy-rcar-gen3-usb2.c drivers/phy/renesas/phy-rcar-gen3-usb2.c:602 rcar_gen3_phy_usb2_probe() warn: passing zero to 'PTR_ERR' drivers/phy/renesas/phy-rcar-gen3-usb2.c:670 rcar_gen3_phy_usb2_probe() warn: passing zero to 'PTR_ERR' biju@be1yocto:/data/biju/linux-next$
Are you using the latest Smatch code? I'm trying to figure out how to reproduce these false and it feels like it should be really easy to do a fresh clone and clean kernel tree but I can't make it work.
I got them for a while, but I fixed it and I thought I had pushed all the fixes...
regards,
dan carpenter
Hi Dan Carpenter,
Thanks for the feedback.
Quoted 5 lines
Subject: Re: [bug report] phy: renesas: phy-rcar-gen2: Add support for r8a77470
That check hasn't been released. (It marks every "if (foo > bar) " as off by one unless it can be proved as okay...
OK. Good to know.
Quoted 19 lines · Biju Das
Quoted 1 line
biju@be1yocto:/data/biju/linux-next$
/data/biju/smatch/smatch_scripts/kchecker
drivers/phy/renesas/Quoted 5 lines
CHECK scripts/mod/empty.c CALL scripts/checksyscalls.sh CALL scripts/atomic/check-atomics.sh CHECK drivers/phy/renesas/phy-rcar-gen2.c
drivers/phy/renesas/phy-rcar-gen2.c:354rcar_gen2_phy_probe()warn:passing zero to '
PTR_ERR'Quoted 1 line
drivers/phy/renesas/phy-rcar-gen2.c:360rcar_gen2_phy_probe()warn:passing zero to '
PTR_ERR'Quoted 1 line
drivers/phy/renesas/phy-rcar-gen2.c:409rcar_gen2_phy_probe()warn:passing zero to '
PTR_ERR'Quoted 1 line
drivers/phy/renesas/phy-rcar-gen2.c:420rcar_gen2_phy_probe()warn:passing zero to '
PTR_ERR'Quoted 2 lines
CHECK
drivers/phy/renesas/phy-rcar-gen3-usb2.cdrivers/phy/renesas/phy-rcar-gen3-usb2.c:602
rcar_gen3_phy_usb2_probe()warn: passing zero to 'PTR_ERR'Quoted 1 line
drivers/phy/renesas/phy-rcar-gen3-usb2.c:670
Quoted 8 lines
rcar_gen3_phy_usb2_probe()warn: passing zero to 'PTR_ERR'Quoted 1 line
biju@be1yocto:/data/biju/linux-next$
Are you using the latest Smatch code? I'm trying to figure out how to reproduce these false and it feels like it should be really easy to do a fresh clone and clean kernel tree but I can't make it work.
I got them for a while, but I fixed it and I thought I had pushed all the fixes...
I am sure ,I have used the latest snapshot from master branch.
Regards,
Biju
Quoted 3 lines · Biju Das
Quoted 1 line
I got them for a while, but I fixed it and I thought I had pushed all the fixes...
I am sure ,I have used the latest snapshot from master branch.
Could you send me your .config?
I can't think how that would matter, but I'm really at a loss...
regards,
dan carpenter