Re: [PATCH] riscv: Convert uses of REG_FMT to %p

From: kbuild test robot
Date: Sat Jul 28 2018 - 12:26:25 EST

Next message: Dmitry Safonov: "Re: [PATCH 00/18] xfrm: Add compat layer"
Previous message: Srinath Mannam: "[PATCH v2] mmc: host: iproc: Add ACPI support to IPROC SDHCI"
In reply to: Joe Perches: "[PATCH] riscv: Convert uses of REG_FMT to %p"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Joe,

I love your patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.18-rc6 next-20180727]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url: https://github.com/0day-ci/linux/commits/Joe-Perches/riscv-Convert-uses-of-REG_FMT-to-p/20180728-104236
config: riscv-defconfig (attached as .config)
compiler: riscv64-linux-gcc (GCC) 8.1.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
GCC_VERSION=8.1.0 make.cross ARCH=riscv

All error/warnings (new ones prefixed by >>):

In file included from include/linux/kernel.h:14,
from include/asm-generic/bug.h:18,
from arch/riscv/include/asm/bug.h:75,
from include/linux/bug.h:5,
from include/linux/mmdebug.h:5,
from include/linux/mm.h:9,
from arch/riscv/mm/fault.c:23:
arch/riscv/mm/fault.c: In function 'do_page_fault':
>> arch/riscv/mm/fault.c:197:60: error: expected ')' before 'REG_FMT'
pr_alert("Unable to handle kernel %s at virtual address " REG_FMT "\n",
^~~~~~~
include/linux/printk.h:288:21: note: in definition of macro 'pr_fmt'
#define pr_fmt(fmt) fmt
^~~
arch/riscv/mm/fault.c:197:2: note: in expansion of macro 'pr_alert'
pr_alert("Unable to handle kernel %s at virtual address " REG_FMT "\n",
^~~~~~~~
In file included from include/linux/printk.h:7,
from include/linux/kernel.h:14,
from include/asm-generic/bug.h:18,
from arch/riscv/include/asm/bug.h:75,
from include/linux/bug.h:5,
from include/linux/mmdebug.h:5,
from include/linux/mm.h:9,
from arch/riscv/mm/fault.c:23:
>> include/linux/kern_levels.h:5:18: warning: format '%s' expects a matching 'char *' argument [-Wformat=]
#define KERN_SOH "\001" /* ASCII Start Of Header */
^~~~~~
include/linux/kern_levels.h:9:20: note: in expansion of macro 'KERN_SOH'
#define KERN_ALERT KERN_SOH "1" /* action must be taken immediately */
^~~~~~~~
include/linux/printk.h:300:9: note: in expansion of macro 'KERN_ALERT'
printk(KERN_ALERT pr_fmt(fmt), ##__VA_ARGS__)
^~~~~~~~~~
arch/riscv/mm/fault.c:197:2: note: in expansion of macro 'pr_alert'
pr_alert("Unable to handle kernel %s at virtual address " REG_FMT "\n",
^~~~~~~~
arch/riscv/mm/fault.c:197:37: note: format string is defined here
pr_alert("Unable to handle kernel %s at virtual address " REG_FMT "\n",
~^

vim +197 arch/riscv/mm/fault.c

07037db5 Palmer Dabbelt 2017-07-10 @23 #include <linux/mm.h>
07037db5 Palmer Dabbelt 2017-07-10 24 #include <linux/kernel.h>
07037db5 Palmer Dabbelt 2017-07-10 25 #include <linux/interrupt.h>
07037db5 Palmer Dabbelt 2017-07-10 26 #include <linux/perf_event.h>
07037db5 Palmer Dabbelt 2017-07-10 27 #include <linux/signal.h>
07037db5 Palmer Dabbelt 2017-07-10 28 #include <linux/uaccess.h>
07037db5 Palmer Dabbelt 2017-07-10 29
07037db5 Palmer Dabbelt 2017-07-10 30 #include <asm/pgalloc.h>
07037db5 Palmer Dabbelt 2017-07-10 31 #include <asm/ptrace.h>
07037db5 Palmer Dabbelt 2017-07-10 32
07037db5 Palmer Dabbelt 2017-07-10 33 /*
07037db5 Palmer Dabbelt 2017-07-10 34 * This routine handles page faults. It determines the address and the
07037db5 Palmer Dabbelt 2017-07-10 35 * problem, and then passes it off to one of the appropriate routines.
07037db5 Palmer Dabbelt 2017-07-10 36 */
07037db5 Palmer Dabbelt 2017-07-10 37 asmlinkage void do_page_fault(struct pt_regs *regs)
07037db5 Palmer Dabbelt 2017-07-10 38 {
07037db5 Palmer Dabbelt 2017-07-10 39 struct task_struct *tsk;
07037db5 Palmer Dabbelt 2017-07-10 40 struct vm_area_struct *vma;
07037db5 Palmer Dabbelt 2017-07-10 41 struct mm_struct *mm;
07037db5 Palmer Dabbelt 2017-07-10 42 unsigned long addr, cause;
07037db5 Palmer Dabbelt 2017-07-10 43 unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
07037db5 Palmer Dabbelt 2017-07-10 44 int fault, code = SEGV_MAPERR;
07037db5 Palmer Dabbelt 2017-07-10 45
07037db5 Palmer Dabbelt 2017-07-10 46 cause = regs->scause;
07037db5 Palmer Dabbelt 2017-07-10 47 addr = regs->sbadaddr;
07037db5 Palmer Dabbelt 2017-07-10 48
07037db5 Palmer Dabbelt 2017-07-10 49 tsk = current;
07037db5 Palmer Dabbelt 2017-07-10 50 mm = tsk->mm;
07037db5 Palmer Dabbelt 2017-07-10 51
07037db5 Palmer Dabbelt 2017-07-10 52 /*
07037db5 Palmer Dabbelt 2017-07-10 53 * Fault-in kernel-space virtual memory on-demand.
07037db5 Palmer Dabbelt 2017-07-10 54 * The 'reference' page table is init_mm.pgd.
07037db5 Palmer Dabbelt 2017-07-10 55 *
07037db5 Palmer Dabbelt 2017-07-10 56 * NOTE! We MUST NOT take any locks for this case. We may
07037db5 Palmer Dabbelt 2017-07-10 57 * be in an interrupt or a critical region, and should
07037db5 Palmer Dabbelt 2017-07-10 58 * only copy the information from the master page table,
07037db5 Palmer Dabbelt 2017-07-10 59 * nothing more.
07037db5 Palmer Dabbelt 2017-07-10 60 */
07037db5 Palmer Dabbelt 2017-07-10 61 if (unlikely((addr >= VMALLOC_START) && (addr <= VMALLOC_END)))
07037db5 Palmer Dabbelt 2017-07-10 62 goto vmalloc_fault;
07037db5 Palmer Dabbelt 2017-07-10 63
07037db5 Palmer Dabbelt 2017-07-10 64 /* Enable interrupts if they were enabled in the parent context. */
1125203c Christoph Hellwig 2018-01-04 65 if (likely(regs->sstatus & SR_SPIE))
07037db5 Palmer Dabbelt 2017-07-10 66 local_irq_enable();
07037db5 Palmer Dabbelt 2017-07-10 67
07037db5 Palmer Dabbelt 2017-07-10 68 /*
07037db5 Palmer Dabbelt 2017-07-10 69 * If we're in an interrupt, have no user context, or are running
07037db5 Palmer Dabbelt 2017-07-10 70 * in an atomic region, then we must not take the fault.
07037db5 Palmer Dabbelt 2017-07-10 71 */
07037db5 Palmer Dabbelt 2017-07-10 72 if (unlikely(faulthandler_disabled() || !mm))
07037db5 Palmer Dabbelt 2017-07-10 73 goto no_context;
07037db5 Palmer Dabbelt 2017-07-10 74
07037db5 Palmer Dabbelt 2017-07-10 75 if (user_mode(regs))
07037db5 Palmer Dabbelt 2017-07-10 76 flags |= FAULT_FLAG_USER;
07037db5 Palmer Dabbelt 2017-07-10 77
07037db5 Palmer Dabbelt 2017-07-10 78 perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, addr);
07037db5 Palmer Dabbelt 2017-07-10 79
07037db5 Palmer Dabbelt 2017-07-10 80 retry:
07037db5 Palmer Dabbelt 2017-07-10 81 down_read(&mm->mmap_sem);
07037db5 Palmer Dabbelt 2017-07-10 82 vma = find_vma(mm, addr);
07037db5 Palmer Dabbelt 2017-07-10 83 if (unlikely(!vma))
07037db5 Palmer Dabbelt 2017-07-10 84 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 85 if (likely(vma->vm_start <= addr))
07037db5 Palmer Dabbelt 2017-07-10 86 goto good_area;
07037db5 Palmer Dabbelt 2017-07-10 87 if (unlikely(!(vma->vm_flags & VM_GROWSDOWN)))
07037db5 Palmer Dabbelt 2017-07-10 88 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 89 if (unlikely(expand_stack(vma, addr)))
07037db5 Palmer Dabbelt 2017-07-10 90 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 91
07037db5 Palmer Dabbelt 2017-07-10 92 /*
07037db5 Palmer Dabbelt 2017-07-10 93 * Ok, we have a good vm_area for this memory access, so
07037db5 Palmer Dabbelt 2017-07-10 94 * we can handle it.
07037db5 Palmer Dabbelt 2017-07-10 95 */
07037db5 Palmer Dabbelt 2017-07-10 96 good_area:
07037db5 Palmer Dabbelt 2017-07-10 97 code = SEGV_ACCERR;
07037db5 Palmer Dabbelt 2017-07-10 98
07037db5 Palmer Dabbelt 2017-07-10 99 switch (cause) {
07037db5 Palmer Dabbelt 2017-07-10 100 case EXC_INST_PAGE_FAULT:
07037db5 Palmer Dabbelt 2017-07-10 101 if (!(vma->vm_flags & VM_EXEC))
07037db5 Palmer Dabbelt 2017-07-10 102 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 103 break;
07037db5 Palmer Dabbelt 2017-07-10 104 case EXC_LOAD_PAGE_FAULT:
07037db5 Palmer Dabbelt 2017-07-10 105 if (!(vma->vm_flags & VM_READ))
07037db5 Palmer Dabbelt 2017-07-10 106 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 107 break;
07037db5 Palmer Dabbelt 2017-07-10 108 case EXC_STORE_PAGE_FAULT:
07037db5 Palmer Dabbelt 2017-07-10 109 if (!(vma->vm_flags & VM_WRITE))
07037db5 Palmer Dabbelt 2017-07-10 110 goto bad_area;
07037db5 Palmer Dabbelt 2017-07-10 111 flags |= FAULT_FLAG_WRITE;
07037db5 Palmer Dabbelt 2017-07-10 112 break;
07037db5 Palmer Dabbelt 2017-07-10 113 default:
07037db5 Palmer Dabbelt 2017-07-10 114 panic("%s: unhandled cause %lu", __func__, cause);
07037db5 Palmer Dabbelt 2017-07-10 115 }
07037db5 Palmer Dabbelt 2017-07-10 116
07037db5 Palmer Dabbelt 2017-07-10 117 /*
07037db5 Palmer Dabbelt 2017-07-10 118 * If for any reason at all we could not handle the fault,
07037db5 Palmer Dabbelt 2017-07-10 119 * make sure we exit gracefully rather than endlessly redo
07037db5 Palmer Dabbelt 2017-07-10 120 * the fault.
07037db5 Palmer Dabbelt 2017-07-10 121 */
07037db5 Palmer Dabbelt 2017-07-10 122 fault = handle_mm_fault(vma, addr, flags);
07037db5 Palmer Dabbelt 2017-07-10 123
07037db5 Palmer Dabbelt 2017-07-10 124 /*
07037db5 Palmer Dabbelt 2017-07-10 125 * If we need to retry but a fatal signal is pending, handle the
07037db5 Palmer Dabbelt 2017-07-10 126 * signal first. We do not need to release the mmap_sem because it
07037db5 Palmer Dabbelt 2017-07-10 127 * would already be released in __lock_page_or_retry in mm/filemap.c.
07037db5 Palmer Dabbelt 2017-07-10 128 */
07037db5 Palmer Dabbelt 2017-07-10 129 if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(tsk))
07037db5 Palmer Dabbelt 2017-07-10 130 return;
07037db5 Palmer Dabbelt 2017-07-10 131
07037db5 Palmer Dabbelt 2017-07-10 132 if (unlikely(fault & VM_FAULT_ERROR)) {
07037db5 Palmer Dabbelt 2017-07-10 133 if (fault & VM_FAULT_OOM)
07037db5 Palmer Dabbelt 2017-07-10 134 goto out_of_memory;
07037db5 Palmer Dabbelt 2017-07-10 135 else if (fault & VM_FAULT_SIGBUS)
07037db5 Palmer Dabbelt 2017-07-10 136 goto do_sigbus;
07037db5 Palmer Dabbelt 2017-07-10 137 BUG();
07037db5 Palmer Dabbelt 2017-07-10 138 }
07037db5 Palmer Dabbelt 2017-07-10 139
07037db5 Palmer Dabbelt 2017-07-10 140 /*
07037db5 Palmer Dabbelt 2017-07-10 141 * Major/minor page fault accounting is only done on the
07037db5 Palmer Dabbelt 2017-07-10 142 * initial attempt. If we go through a retry, it is extremely
07037db5 Palmer Dabbelt 2017-07-10 143 * likely that the page will be found in page cache at that point.
07037db5 Palmer Dabbelt 2017-07-10 144 */
07037db5 Palmer Dabbelt 2017-07-10 145 if (flags & FAULT_FLAG_ALLOW_RETRY) {
07037db5 Palmer Dabbelt 2017-07-10 146 if (fault & VM_FAULT_MAJOR) {
07037db5 Palmer Dabbelt 2017-07-10 147 tsk->maj_flt++;
07037db5 Palmer Dabbelt 2017-07-10 148 perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MAJ,
07037db5 Palmer Dabbelt 2017-07-10 149 1, regs, addr);
07037db5 Palmer Dabbelt 2017-07-10 150 } else {
07037db5 Palmer Dabbelt 2017-07-10 151 tsk->min_flt++;
07037db5 Palmer Dabbelt 2017-07-10 152 perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS_MIN,
07037db5 Palmer Dabbelt 2017-07-10 153 1, regs, addr);
07037db5 Palmer Dabbelt 2017-07-10 154 }
07037db5 Palmer Dabbelt 2017-07-10 155 if (fault & VM_FAULT_RETRY) {
07037db5 Palmer Dabbelt 2017-07-10 156 /*
07037db5 Palmer Dabbelt 2017-07-10 157 * Clear FAULT_FLAG_ALLOW_RETRY to avoid any risk
07037db5 Palmer Dabbelt 2017-07-10 158 * of starvation.
07037db5 Palmer Dabbelt 2017-07-10 159 */
07037db5 Palmer Dabbelt 2017-07-10 160 flags &= ~(FAULT_FLAG_ALLOW_RETRY);
07037db5 Palmer Dabbelt 2017-07-10 161 flags |= FAULT_FLAG_TRIED;
07037db5 Palmer Dabbelt 2017-07-10 162
07037db5 Palmer Dabbelt 2017-07-10 163 /*
07037db5 Palmer Dabbelt 2017-07-10 164 * No need to up_read(&mm->mmap_sem) as we would
07037db5 Palmer Dabbelt 2017-07-10 165 * have already released it in __lock_page_or_retry
07037db5 Palmer Dabbelt 2017-07-10 166 * in mm/filemap.c.
07037db5 Palmer Dabbelt 2017-07-10 167 */
07037db5 Palmer Dabbelt 2017-07-10 168 goto retry;
07037db5 Palmer Dabbelt 2017-07-10 169 }
07037db5 Palmer Dabbelt 2017-07-10 170 }
07037db5 Palmer Dabbelt 2017-07-10 171
07037db5 Palmer Dabbelt 2017-07-10 172 up_read(&mm->mmap_sem);
07037db5 Palmer Dabbelt 2017-07-10 173 return;
07037db5 Palmer Dabbelt 2017-07-10 174
07037db5 Palmer Dabbelt 2017-07-10 175 /*
07037db5 Palmer Dabbelt 2017-07-10 176 * Something tried to access memory that isn't in our memory map.
07037db5 Palmer Dabbelt 2017-07-10 177 * Fix it, but check if it's kernel or user first.
07037db5 Palmer Dabbelt 2017-07-10 178 */
07037db5 Palmer Dabbelt 2017-07-10 179 bad_area:
07037db5 Palmer Dabbelt 2017-07-10 180 up_read(&mm->mmap_sem);
07037db5 Palmer Dabbelt 2017-07-10 181 /* User mode accesses just cause a SIGSEGV */
07037db5 Palmer Dabbelt 2017-07-10 182 if (user_mode(regs)) {
07037db5 Palmer Dabbelt 2017-07-10 183 do_trap(regs, SIGSEGV, code, addr, tsk);
07037db5 Palmer Dabbelt 2017-07-10 184 return;
07037db5 Palmer Dabbelt 2017-07-10 185 }
07037db5 Palmer Dabbelt 2017-07-10 186
07037db5 Palmer Dabbelt 2017-07-10 187 no_context:
07037db5 Palmer Dabbelt 2017-07-10 188 /* Are we prepared to handle this kernel fault? */
07037db5 Palmer Dabbelt 2017-07-10 189 if (fixup_exception(regs))
07037db5 Palmer Dabbelt 2017-07-10 190 return;
07037db5 Palmer Dabbelt 2017-07-10 191
07037db5 Palmer Dabbelt 2017-07-10 192 /*
07037db5 Palmer Dabbelt 2017-07-10 193 * Oops. The kernel tried to access some bad page. We'll have to
07037db5 Palmer Dabbelt 2017-07-10 194 * terminate things with extreme prejudice.
07037db5 Palmer Dabbelt 2017-07-10 195 */
07037db5 Palmer Dabbelt 2017-07-10 196 bust_spinlocks(1);
07037db5 Palmer Dabbelt 2017-07-10 @197 pr_alert("Unable to handle kernel %s at virtual address " REG_FMT "\n",

:::::: The code at line 197 was first introduced by commit
:::::: 07037db5d479f90377c998259a4f9a469c404edf RISC-V: Paging and MMU

:::::: TO: Palmer Dabbelt <palmer@xxxxxxxxxxx>
:::::: CC: Palmer Dabbelt <palmer@xxxxxxxxxxx>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

Attachment: .config.gz
Description: application/gzip

Next message: Dmitry Safonov: "Re: [PATCH 00/18] xfrm: Add compat layer"
Previous message: Srinath Mannam: "[PATCH v2] mmc: host: iproc: Add ACPI support to IPROC SDHCI"
In reply to: Joe Perches: "[PATCH] riscv: Convert uses of REG_FMT to %p"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]