crypto: iaa - Add compression mode management along with fixed mode
Hello Tom Zanussi,
The patch b190447e0fa3: "crypto: iaa - Add compression mode management along with fixed mode" from Dec 5, 2023 (linux-next), leads to the following Smatch static checker warning:
drivers/crypto/intel/iaa/iaa_crypto_main.c:532 init_device_compression_mode() error: not allocating enough for = 'device_mode->aecs_decomp_table' 5352 vs 1600
drivers/crypto/intel/iaa/iaa_crypto_main.c
510 static int init_device_compression_mode(struct iaa_device *iaa_device,
511 struct iaa_compression_mode *mode,
512 int idx, struct idxd_wq *wq)
513 {
514 size_t size = sizeof(struct aecs_comp_table_record) + IAA_AECS_ALIGN;
515 struct device *dev = &iaa_device->idxd->pdev->dev;
516 struct iaa_device_compression_mode *device_mode;
517 int ret = -ENOMEM;
518
519 device_mode = kzalloc(sizeof(*device_mode), GFP_KERNEL);
520 if (!device_mode)
521 return -ENOMEM;
522
523 device_mode->name = kstrdup(mode->name, GFP_KERNEL);
524 if (!device_mode->name)
525 goto free;
526
527 device_mode->aecs_comp_table = dma_alloc_coherent(dev, size,
528 &device_mode->aecs_comp_table_dma_addr, GFP_KERNEL);
529 if (!device_mode->aecs_comp_table)
530 goto free;
531
--> 532 device_mode->aecs_decomp_table = dma_alloc_coherent(dev, size,
^^^^^^comp and decomp sizes are different. So we should be allocating aecs_decomp_table_record + IAA_AECS_ALIGN here probably.
533 &device_mode->aecs_decomp_table_dma_addr, GFP_KERNEL);
534 if (!device_mode->aecs_decomp_table)
535 goto free;
536
537 /* Add Huffman table to aecs */
538 memset(device_mode->aecs_comp_table, 0, sizeof(*device_mode->aecs_comp_table));
539 memcpy(device_mode->aecs_comp_table->ll_sym, mode->ll_table, mode->ll_table_size);
540 memcpy(device_mode->aecs_comp_table->d_sym, mode->d_table, mode->d_table_size);
541 regards,
dan carpenter
Hi Dan,
Quoted 38 lines · Dan Carpenter
Hello Tom Zanussi,
The patch
b190447e0fa3: "crypto: iaa - Add compression mode management along with fixed mode" from Dec 5, 2023 (linux-next), leads to the following Smatch static checker warning:drivers/crypto/intel/iaa/iaa_crypto_main.c:532
init_device_compression_mode()
error: not allocating enough for = 'device_mode-Quoted 1 line
aecs_decomp_table' 5352 vs 1600
drivers/crypto/intel/iaa/iaa_crypto_main.c
510 static int init_device_compression_mode(structiaa_device*iaa_device,
511 structiaa_compression_mode*mode,
512 int idx, structidxd_wq*wq)
513 {
514size_tsize =sizeof(struct aecs_comp_table_record)+IAA_AECS_ALIGN;
515 struct device *dev =&iaa_device->idxd->pdev->dev; 516 structiaa_device_compression_mode*device_mode; 517 int ret =-ENOMEM;
518
519device_mode=kzalloc(sizeof(*device_mode), GFP_KERNEL);
520 if (!device_mode)
521 return-ENOMEM;
522
523device_mode->name=kstrdup(mode->name, GFP_KERNEL); 524 if (!device_mode->name)
525 goto free;
526
527device_mode->aecs_comp_table= dma_alloc_coherent(dev, size,
528&device_mode->aecs_comp_table_dma_addr,GFP_KERNEL);
Quoted 8 lines
529 if (!device_mode->aecs_comp_table)
530 goto free;
531
--> 532device_mode->aecs_decomp_table= dma_alloc_coherent(dev, size,
^^^^^^
comp and decomp sizes are different. So we should be allocatingaecs_decomp_table_record+IAA_AECS_ALIGNhere probably.
Thanks for pointing this out.
This is actually unused by the current code - it was part of the canned mode that was removed during review.
I'll submit a patch to remove it completely, and will fix it if/when we resubmit canned mode.
Thanks,
Tom
Quoted 17 lines
533
&device_mode->aecs_decomp_table_dma_addr,GFP_KERNEL); 534 if (!device_mode->aecs_decomp_table) 535 goto free;
536
537 /* Add Huffman table to aecs */
538memset(device_mode->aecs_comp_table, 0, sizeof(*device_mode->aecs_comp_table));
539 memcpy(device_mode->aecs_comp_table->ll_sym, mode-Quoted 1 line
ll_table,mode->ll_table_size);540 memcpy(device_mode->aecs_comp_table->d_sym, mode-
Quoted 1 line
d_table,mode->d_table_size);541
regards,
dan carpenter