| Message ID | 20221105095357.21199-1-yekai13@huawei.com |
|---|---|
| Headers | show |
| Series | crypto: hisilicon/qm - reconstruct and optimize the debugfs code | expand |
On Sat, Nov 05, 2022 at 09:53:54AM +0000, Kai Ye wrote: > > +static void dfx_regs_uninit(struct hisi_qm *qm, > + struct dfx_diff_registers *dregs, int reg_len) > +{ > + int i; > + > + /* Setting the pointer is NULL to prevent double free */ > + for (i = 0; i < reg_len; i++) { > + kfree(dregs[i].regs); > + dregs[i].regs = NULL; > + } > + kfree(dregs); > + dregs = NULL; > +} The line that I complained about is still here. Cheers,
On 2022/11/5 18:18, Herbert Xu wrote: > On Sat, Nov 05, 2022 at 09:53:54AM +0000, Kai Ye wrote: >> +static void dfx_regs_uninit(struct hisi_qm *qm, >> + struct dfx_diff_registers *dregs, int reg_len) >> +{ >> + int i; >> + >> + /* Setting the pointer is NULL to prevent double free */ >> + for (i = 0; i < reg_len; i++) { >> + kfree(dregs[i].regs); >> + dregs[i].regs = NULL; >> + } >> + kfree(dregs); >> + dregs = NULL; >> +} > The line that I complained about is still here. > > Cheers, ok thanks kai
On Mon, Nov 07, 2022 at 02:53:28PM +0800, yekai (A) wrote: > > > On 2022/11/5 18:18, Herbert Xu wrote: > > On Sat, Nov 05, 2022 at 09:53:54AM +0000, Kai Ye wrote: > >> +static void dfx_regs_uninit(struct hisi_qm *qm, > >> + struct dfx_diff_registers *dregs, int reg_len) > >> +{ > >> + int i; > >> + > >> + /* Setting the pointer is NULL to prevent double free */ > >> + for (i = 0; i < reg_len; i++) { > >> + kfree(dregs[i].regs); > >> + dregs[i].regs = NULL; > >> + } > >> + kfree(dregs); > >> + dregs = NULL; > >> +} > > The line that I complained about is still here. > > > > Cheers, > > ok Just to be clear, it's the last line "dregs = NULL" that I was referring to. It makes no sense to zero a variable that is on the stack. Cheers,
On 2022/11/7 16:47, Herbert Xu wrote: > On Mon, Nov 07, 2022 at 02:53:28PM +0800, yekai (A) wrote: >> >> On 2022/11/5 18:18, Herbert Xu wrote: >>> On Sat, Nov 05, 2022 at 09:53:54AM +0000, Kai Ye wrote: >>>> +static void dfx_regs_uninit(struct hisi_qm *qm, >>>> + struct dfx_diff_registers *dregs, int reg_len) >>>> +{ >>>> + int i; >>>> + >>>> + /* Setting the pointer is NULL to prevent double free */ >>>> + for (i = 0; i < reg_len; i++) { >>>> + kfree(dregs[i].regs); >>>> + dregs[i].regs = NULL; >>>> + } >>>> + kfree(dregs); >>>> + dregs = NULL; >>>> +} >>> The line that I complained about is still here. >>> >>> Cheers, >> ok > Just to be clear, it's the last line "dregs = NULL" that I was > referring to. It makes no sense to zero a variable that is on > the stack. > > Cheers, I got it. I will fix it and resend the patch set in the next version. Thanks Kai
Split a debugfs.c from qm to reduce the qm code complexity. And optimize the process of the command dumping. Some optimizations on command dumping in debugfs.c file code. changes v1->v2: - delete a redundant line. Kai Ye (4): crypto: hisilicon/qm - modify the process of regs dfx crypto: hisilicon/qm - split a debugfs.c from qm crypto: hisilicon/qm - the command dump process is modified crypto: hisilicon/debugfs - delete redundant null assignment operations drivers/crypto/hisilicon/Makefile | 2 +- drivers/crypto/hisilicon/debugfs.c | 1172 ++++++++++++++++++ drivers/crypto/hisilicon/hpre/hpre_main.c | 6 +- drivers/crypto/hisilicon/qm.c | 1360 ++------------------- drivers/crypto/hisilicon/qm_common.h | 87 ++ drivers/crypto/hisilicon/sec2/sec_main.c | 6 +- drivers/crypto/hisilicon/zip/zip_main.c | 6 +- include/linux/hisi_acc_qm.h | 8 +- 8 files changed, 1381 insertions(+), 1266 deletions(-) create mode 100644 drivers/crypto/hisilicon/debugfs.c create mode 100644 drivers/crypto/hisilicon/qm_common.h